Loading...
HomeMy WebLinkAboutMINUTES - 03281995 - H.6 H. 6 THE BOARD OR SUPERVISORS OF CONTRA COSTA COUNTY, CALIFORNIA Adopted this Order on March 28, 1995 by the following vote: AYES: Supervisors Rogers, Smith, DeSaulnier, Torlakson, Bishop NOES: None ABSENT: None ABSTAIN: None ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ SUBJECT: Proposed Ordinance Prohibiting Advertising of Tobacco Products and Alcoholic Beverages on Billboards Near Schools Following introductory comments of Supervisor Jim Rogers, the Chair opened the hearing on the proposed Ordinance that would prohibit advertising tobacco products and alcoholic beverages on billboards near schools. The following persons spoke in support of the proposed Ordinance: Peter Burnett, Substance Abuse Advisory Board, 890 Hamilton Drive, Pleasant Hill; Lisa Korpus, 2444 Cherry Hills Drive, Lafayette; Juliette Linzer, American Lung Association of Contra Costa/Solano Counties, 105 Astrid Drive, Pleasant Hill; Julie Freestone, Contra Costa County Tobacco Prevention Program, 826 32nd Avenue, Richmond; Galen Ellis, Contra Costa County Health Services Department 597 Center Avenue, Suite 325, Martinez; Joyce White, 648 South 18th Street, Richmond; Delene Bliss, American Cancer Society, 2413 Pine Street, Martinez; Joel E. White, MD, FACR, American Cancer Society, 318 Ridgestone Court, Walnut Creek; Kaye Rosso, American Cancer Society, 34 Ashbrook, Moraga; David Jenne, 433 LeJean Way, Walnut Creek; Joann Pavlinel, 6243 Highland Avenue, Richmond; and Howard Korpus, 2444 Cherry Hills Drive, Lafayette.. The following persons spoke in opposition to the proposed Ordinance: Donna Lucar, Anheuser-Busch, 1029 F Street, Suite 400, Sacramento; George Broder, Patrick Media Group, 1601 Maritime Street, Oakland; Ronald W. Beals, Gannett Outdoor Co. , 1311 Terminal Street, West Sacramento; and P. Cameron DeVore, Anheuser-Busch, 2600 Century Square, 1501 - 4th Avenue, Seattle, Washington 98101. All persons desiring to speak were heard, the hearing was closed. j u At the conclusion of the Board's discussion on this matter, the Board REFERRED the proposed Ordinance to the Internal Operations Committee for further review. The Board REQUESTED County Counsel to incorporate the information presented today and to prepare appropriate findings for presentation to the Internal Operations Committee. The Board REQUESTED staff to explore the issues presented at the hearing this day including an analysis of the law and costs associated therewith, review of the issues by the Substance Abuse Commission, industry representatives, and the City/County Relations Commission. I hereby certify that this is a true and correct copy of an action taken and entered on the minutes of the Board of Supervisors on the date shown. ATTESTED: yZy 99, /9'r a PHIL BATCHELOR,Clerk of the Board of Supervisors and County Administrator 6r —�Deputy cc: Internal Operations Committee Substance Abuse Commission Advertising Representatives City/County Relations Commission County Counsel County Administrator Health Services Director Community Development Director .......—.................... —..—..............�...e +avvv�Iesct['v16f.'.6G•fifiC i'ViiL•:Y.rf.ii.:1f'C'r.'K tf.XSt.H'.Gw.r.L'w t:.t:r u:e. ...�..._,"•v.. ,..... _ Contra TO: BOARD OF SUPERVISORS " Costa n. Count FROM: HARVEY E. BRAGDON y DIRECTOR OF COMMUNITY DEVELOPMENT �siT DATE: March 28, 1995 ra �oori t i SUBJECT: Ordinance Prohibiting Advertising of Tobacco Products and Alcoholic Beverages on Billboards Near Schools. SPECIFIC REQUEST(S) OR RECOMMENDATIONS) & BACKGROUND AND JUSTIFICATION RECOMMENDATIONS 1. Determine that the proposed Ordinance prohibiting advertising of tobacco products and alcohol beverages on billboards near schools is exempt from California Environmental Quality Act under Section 150612(b) (3) of the State CEQA Guidelines. J 2. Approve the Draft Ordinance prohihiting.-advertising-of tobacco products and alcohol beverages on billboards near schools. 3. Introduce the Ordinance, waive reading and set a date for adoption. FISCAL IMPACT There will be unknown costs associated with the enforcement of this Ordinance. BACKGROUND/REASONS FOR RECOMMENDATIONS On January 24, 1995, the Board of Supervisors directed County ` Counsel to prepare a Draft Ordinance prohibiting advertisement of alcoholic beverages or tobacco products on billboards within 2,000 feet of schools. On February 21, 1995, the County Planning Commission held a fully noticed public .hearing on the proposed Ordinance, took testimony and considered all of the information presented relative to the proposed Ordinance. Upon the close of the public hearing, the Commission discussed the matter and determined that the Board should approve the Draft:Ordinance. The Commission also recommended that the Board expand its efforts to encourage other jurisdictions to adopt similar ordinances, including mobile I CONTINUED ON ATTACHMENT: % YES SIGNATIIRis _ RECOMMENDATION OF COUNTY ADMINISTRATOR RECOMMENDATION OF BOARD COMMI EE _ APPROVE OTHER SIGNATURE(S) : ACTION OF BOARD ON APPROVED AS RECOMMENDED OTHER VOTE OF SU VISORS I HEREBY CERTIFY TTHIS IS A UNANIMOUS (ABS TRUE AND CORREC COPY OF AN AYES: ES: ACTION T N AND ENTERED ON THE ABSENT: ABS MIND OF THE BOARD OF ERVISORS ON THE DATE SHOWN. Contact:Dennis M. Barry 646-2091 ATTESTED cc: Community Development De ment PHIL BATCHELOR, CLERK OF County Counsel 7. OARD OF SUPERVISORS AND COU MINISTRATOR BY , DE DMB/ Page Two advertising, such as advertising on buses. The County Planning Commission further recommended that the Board of Supervisors direct staff to investigate the feasibility of modifying the Ordinance to increase the distance from schools and reduce the size of the signs regulated as much as possible, but indicated that adoption of the Ordinance should not be delayed for this purpose. I Resolution No.6-1995 RESOLUTION OF THE PLANNING COMMISSION OF THE COUNTY OF CONTRA COSTA, STATE OF CALIFORNIA, INCORPORATING FINDINGS AND RECOMMENDATIONS ON THE PROPOSED ORDINANCE PROHIBITING ADVERTISING OF TOBACCO PRODUCTS AND ALCOHOLIC BEVERAGES ON BILLBOARDS NEAR SCHOOLS ADDING ARTICLE #88-6-11 TO THE COUNTY ORDINANCE CODE. WHEREAS, on January 24, 1995, the Board of Supervisors directed County Counsel to prepare an ordinance prohibiting advertisement of alcoholic beverages and tobacco products on billboards within 2,000-ft., of schools; and WHEREAS, Government Code section 6585 requires that such an ordinance be presented to the County Planning Commission for study and report to the Board of Supervisors; and WHEREAS, the draft ordinance enumerates the purpose and findings, definition, exemption, prohibition of advertising of same, removal of existing advertisements, abatement of illegal advertisements as well as severability; and WHEREAS, upon receipt of referral, the Community Development Department staff consulted with various cities within the County which already have such an ordinance in effect; and WHEREAS, the proposed ordinance us exempt from the provisions of the California Environmental Quality Act pursuant to CEQA Guideline Section 15061 (b) (3); therefore, no environmental study was made; and WHEREAS, after having completed a draft that would cover County objectives, a public hearing was noticed and scheduled before the County Planning Commission for Tuesday, February 21 , 1995, whereat all persons interested therein might appear and be heard; and WHEREAS, seven(7) persons appeared at said public hearing speaking strongly in favor of said proposed ordinance; and WHEREAS, no one appeared to speak in opposition to the draft; and WHEREAS, the County Planning Commission having fully reviewed, considered and evaluated all the testimony and evidence submitted in this matter; and NOW, THEREFORE, BE IT RESOLVED that the County Planning Commission recommends to the Board of Supervisors of the County of Contra Costa, State of California, that this draft ordinance, which prohibits advertising of tobacco products and alcoholic beverages on billboards near schools be APPROVED; and Resolution No.6-1995 BE IT FURTHER RESOLVED that the County Planning Commission recommends that the Board of Supervisors expand its efforts to encourage other jurisdictions within the County to adopt similar ordinances including mobile advertising such as advertising signs on busses; and BE IT FURTHER RESOLVED that the County Planning Commission recommends that the Board of Supervisors direct staff to investigate and report on the feasibility of modifying the provisions of the ordinance to increase the distance from schools and reduce the size of signs regulated to the greatest extent practicable; and BE IT FURTHER RESOLVED that all written and graphic material developed for and pertaining to these proceedings are made a part of the record; and BE IT FURTHER RESOLVED that the Chairman and Secretary of this Commission will sign and attest the certified copy of this resolution and deliver the same to the Board of Supervisors all in accordance with the Government Code of the State of California. The instruction by the Planning Commission to prepare this resolution was given by motion of the Commission on Tuesday, February 21 , 1995, by the following vote: AYES: Commissioners - Braxton, Clark, Gaddis, Hanecak, Straus, Wong, Terrell. NOES: Commissioners - None. ABSENT: Commissioners - None. ABSTAIN: Commissioners - None. I, Marvin J. Terrell, Chairman of the Planning Commission of the County of Contra Costa, State of California, hereby certify that the foregoing was duly called and held in accordance with the law on Tuesday, March 7, 1995, and that this resolution was duly and regularly passed and adopted by the following vote of the Commission: AYES: Commissioners - CLARK, BRAXTON, WONG, STRAUS, GADDIS, HANECAK, TERRELL. NOES: Commissioners - NONE. Resolution No. 6-1995 ABSENT: Commissioners - NONE. ABSTAIN: Commissioners - NONE. V Chairman of the Planning Commission, Contra Costa County, State of California. AContra he Planning Commission, County, State of California. CONTRA COSTA COUNTY COMMUNITY DEVELOPMENT DEPARTMENT TO: County Planning Commission DATE: February 13, 1995 FROM: Harvey E. Bragdon, Director -4* Ll SUBJECT: Bil'lbotad Ordinance Regulating Alcohol and Tobacco Advertising This is a County-initiated addition to the Zoning Ordinance intended to regulate the location of billboards which advertise tobacco products and alcoholic beverages. The Board of Supervisors is seeking the input of the County Planning Commission. Staff recommends that the Commission take public testimony and then provide the Board of Supervisors with their recommendations regarding the proposed ordinance modifications. MF/aa LTRVIII/Bill.MF rn, COUNTY COUNSEL'S OFFICE ✓,�j��`a CDS CONTRA COSTA COUNTY f Tq MARTINEZ, CALIFORNIA 2 tp lf Date: February 2, 1995 'C�`►�T��p T To: Harvey E. Bragdon, Dir. of Community Development Attn: Mary Fleming, Assistant Director From: Victor J. Westman, County Counsel By: Silvano B. Marchesi, Ass't County Counsel 4D Re: Ordinance Regulating Advertisement of Tobacco Products and Alcoholic Beverages on Billboards On 24 January 1995, the Board of Supervisors directed this office to prepare an ordinance prohibiting advertisement of alcoholic beverages and tobacco products on billboards within 2,000 feet of schools. We have previously furnished you with a copy of the first draft of such an ordinance. Attached is Draft No. 2, which reflects the comments of members of the Board. In our view, Government Code sections 65850 and 65853 probably require that such an ordinance be presented to the planning commission before consideration by the Board. It has been determined that we should do so directly, rather than wait for a referral by the Board. Therefore, we request that your department present the attached draft ordinance to the County Planning Commission as soon as possible for study and report to the Board of Supervisors, in accordance with the provisions of Government Code section 65853 et seq. Please keep the Substance Abuse Advisory Board, the Youth Commission, and all other interested parties advised on hearing dates. Attachment cc: Board of Supervisors Phil Batchelor, County Administrator Val Alexeeff, Director, GMEDA Jeanne Maglio, Chief Clerk of the Board of Supervisors Substance Abuse Advisory Board Youth Commission, c/o Supervisor DeSaulnier's Office a i DRAFT NO. 2 ORDINANCE NO. 95- ORDINANCE PROHIBITING ADVERTISING OF TOBACCO PRODUCTS AND ALCOHOLIC BEVERAGES ON BILLBOARDS NEAR SCHOOLS The Contra Costa County Board of Supervisors ordains as follows (omitting the parenthetical- footnotes from the official text of the enacted or amended provisions of the County Ordinance Code): SECTION I. SUMMARY. This ordinance adds Article 88-6.11 to the County Ordinance Code to prohibit outdoor advertising of tobacco products and alcoholic beverages in the unincorporated area of the county on billboards near schools. SECTION II. Article 88-6.11 is added to the County Ordinance Code to read: Article 88-6.11 Billboards Advertising Tobacco Products and Alcoholic Beverages 88-6.1102 Purpose and Findings. The board of supervisors finds the following: (1) The county has a special responsibility to protect minors. (2) Most people develop their patterns of tobacco and alcohol use before becoming adults, at a time when they are impressionable and often incapable of making informed decisions. (3) The use,by minors of tobacco products and alcoholic beverages is dangerous because tobacco and alcohol can serve as "gateway drugs" to crack cocaine, marijuana, heroin, and other illegal drugs. (4) The county has strongly supported classroom education concerning the dangers of the use of all illegal drugs, but this message is undercut if there are advertisements near schools that encourage the use of tobacco products and the consumption of alcoholic beverages. (5) Many school districts in the county have endorsed restricting the advertising of tobacco products and alcoholic beverages near schools in order to avoid sending mixed messages to their students. (6) The City of Richmond's Anti-Drug Task Force and the Brookside Hospital Board of Directors, among others, strongly endorse a restriction on the advertising of tobacco products and alcoholic beverages near schools as a means to curb the illegal use of drugs by minors. (7) Accordingly, this board determines that the health, safety, and general welfare of the minors of the county could be benefitted by the regulation of advertising of tobacco products and alcoholic beverages on billboards near schools. 88-6.1104 Definition. "Billboard" means any outdoor advertising sign (as defined in this code) which is at least forty square feet in area. 88-6.1106 Exemption. This article shall not apply to a billboard that is an accessory sign, as defined in this code. 88-6.1108 Prohibition of Advertising of Tobacco Products and Alcoholic Beverages. , No person, firm, *corporation, PartnershiP� or other organization shall permit any nonaccessory billboard advertising a tobacco product or an alcoholic beverage to be placed or located within 2,000 feet of any public or private elementary or secondary school within the unincorporated area of the county. 88-6.1110 Removal of Existing Advertisements of Tobacco Products and Alcoholic Beverages. Any advertisement on a billboard that advertises a tobacco product or an alcoholic beverage existing on the effective date of this article and that violates section 88- 6.1108 shall be removed within 90 days after said effective date; provided that, any person, firm, corporation, partnership or other organization having a valid contract for the placement of such advertisement shall be permitted to retain such advertisement until the contract is terminated or expires, so long as said contract was entered into and in effect as of January 1, 1995, and is not renewed or extended beyond its initial term. 88-6.1112 Abatement of Illegal Advertisements. Whenever the Community Development Director, or his or her designee, determines that a violation of section 88- 6.1108 exists, the Director may initiate the abatement procedures specified in the Uniform Public Nuisance Abatement Procedure (article 14-6.4 of this code), or otherwise specified in this code. SECTION III. SEVERABILITY. If any section, subsection, subdivision, paragraph, sentence, clause, or phrase of this ordinance is for any reason held to be unconstitutional or invalid, such a decision shall not affect the validity of the remaining portions of this ordinance. The board of supervisors hereby declares that it would have passed the remainder of this ordinance irrespective of the unconstitutionality or invalidity of any section, subsection, subdivision, paragraph, sentence, clause, or phrase. ORDINANCE NO. 95- 2 q 71 t SECTION IV. EFFECTIVE DATE: This ordinance becomes effective 30 days after passage, and within 15 days of passage shall be published once with the names of the supervisors voting for and against it in the a newspaper published in this County. PASSED ON by the following vote: AYES: NOES: ABSENT: ABSTAIN: ATTEST: PHIL BATCHELOR, Clerk of the Board of Supervisors Board Chair and County Administrator By [SEAL] Deputy SBM ORDINANCE NO. 95- 3 ,gs COUNTY COUNSEL'S OFFICE ilee,0,',�, 0 CONTRA COSTA COUNTY MARTINEZ, CALIFORNIA Date: January 31, 1995 To: Board of S ervisors From: Victor J. Westman, County Counsel By: Silvano B. Marchesi, Ass't Countyounse C Re: Ordinance Regulating Advertisement of Tobacco Products and Alcoholic Beverages on Billboards Enclosed is a draft of the Tobacco-Alcohol ordinance, and we provide the following comments. 1. We have tried to keep the ordinance as close as possible to the Richmond initiative measure. a) In § 88-6.1106 (p. 2), we have conformed the terms to existing terms in the County Ordinance Code (e.g., "outdoor advertising sign," "accessory sign b) In § 88-6.1108 (p.2), we have specified that the ordinance applies to the unincorporated area of the County. c) We have omitted the section relating to "Public Nuisance; Injunction," since the County Code already includes a provision with the same effect (Co. Ord. Code, § 14-6.204). d) In § 88-6.1110 (p. 2), we have left the "grandfather" date blank. It is not clear to us how the date in the initiative measure (1 January 1990) was selected, so we have left it to the discretion of the Board. e) In § 88-6.1112 (p. 2), we have shortened the abatement provision to an incorporation of the existing Uniform Public Nuisance Abatement Procedure in the County Code. This approach also eliminates the need for a section on a Hearing by the Board of Supervisors. f) We have eliminated the section allowing amendment upon a 415 vote of the Board. Board of Supervisors January 31, 1995 Page 2 g) We also have prepared a red-lined draft of the ordinance. If any Board member would like to view that draft to compare the proposed ordinance with the Richmond initiative measure, please advise. 2. We have-discussed this proposed-ordinance briefly with Community Development Department staff. As far as we know, that department has not had any recent occasion to conduct any inventory of existing "outdoor advertising signs" in the unincorporated area near schools or that advertise tobacco products or alcoholic beverages. Thus, we are not aware of the number of any existing signs that would be covered by this ordinance. 3. Section 88-6.1110 (p. 2) appears to require the removal of the prohibited advertisements within 90 days after the effective date of the ordinance. We interpret this provision to require the removal only of the advertisement, and not the destruction of the billboard itself. If this is not the intent of the Board, please advise, so that we can clarify the language of this section. (See also, paragraph 4, below.) Also, we are unclear how this provision is to be reconciled with the grandfather clause that follows it. We assume that an existing prohibited advertisement can remain as long as it is covered by a contract that was in existence on the grandfather date and that has not been extended. 4. A number of cases have dealt with the issue of amortization periods, usually in connection with an ordinance that requires the removal of a prohibited sign. We have not seen any such amortization period (for removal of a sign) shorter than 20 months, and usually it extends over a period of five years or more. 5. Regarding the second recommendation in the Board Order of 24 January 1995, we will begin work on a separate ordinance to ban, prospectively, new billboards ("non-accessory signs") in the unincorporated area of the County. We will obtain comments and assistance from Community Development Department staff and present a draft for Board consideration in the near future. 6. The State Planning and Zoning Act provides that a zoning ordinance that modifies a regulation pertaining to signs and billboards must be heard by the planning commission before it is considered by the Board of Supervisors. (Gov. Code, §§ 65853, 65850) While it is not completely clear whether these provisions apply to a regulation of the content of commercial billboards, as in this case, we recommend that the proposed ordinance be referred to the County Planning Commission for hearing in accordance with Government Code section 65853. i r r Board of Supervisors January 31, 1995 Page 3 6. By copy of this memo, we request the Clerk of the Board to schedule the proposed ordinance on the Board's agenda of 7 February 1995 for referral to the County Planning Commission . The Substance Abuse Advisory Board and the Youth Commission are advised of this item (unless rescheduled by the Board) pursuant to Recommendation (3) of the Board's Order of 24 January 1995. Attachments cc: Phil Batchelor, County Administrator Val Alexeeff, Director, GMEDA �rvey E. Bragdon, Director, Community Development Department Jeanne Maglio, Chief Clerk of the Board of Supervisors i Substance Abuse Advisory Board, c/o Chuck Deutschman, Health Services Dept. Youth Commission, c/o Supervisor DeSaulnier's Office i r - j CLEAN DRAFT NO. 1 4 i ORDINANCE NO. 95- ORDINANCE PROHIBITING ADVERTISING OF TOBACCO PRODUCTS AND ALCOHOLIC BEVERAGES ON BILLBOARDS NEAR SCHOOLS The Contra Costa County Board'of Supervisors ordains as-follows (omitting the parenthetical footnotes from the official text of the enacted or amended provisions of the County Ordinance Code): SECTION I. SUMMARY. This ordinance adds Article 88-6.11 to the County Ordinance Code to prohibit outdoor advertising of tobacco products and alcoholic beverages in the unincorporated area of the county on billboards near schools. SECTION II. Article 88-6.11 is added to the County Ordinance Code to read: Article 88-6.11 Billboards Advertising Tobacco Products and Alcoholic Beverages 88-6.1102 Purpose and Findings. The board of supervisors finds the following: (1) The county has a special responsibility to protect minors. (2) " Most people develop their patterns of tobacco and alcohol use before becoming adults, at a time when they are impressionable and often incapable of making informed decisions. (3) The use by minors of tobacco products and alcoholic beverages is dangerous because'tobacco and alcohol can serve as "gateway drugs" to crack cocaine, marijuana, heroin, and other illegal drugs. (4) The county has strongly supported classroom education concerning the dangers of the use of all illegal drugs, but this message is undercut if there are advertisements near schools that encourage the use of tobacco products and the consumption of alcoholic beverages. (5) Many school districts in the county have endorsed restricting the advertising of tobacco products and alcoholic beverages near schools in order to avoid sending mixed messages to their students. (6) The City of Richmond's Anti-Drug Task Force and the Brookside Hospital Board of Directors, among others, strongly endorse a restriction on the advertising of tobacco products and alcoholic beverages near schools as a means to curb the illegal use of drugs by minors. (7) Accordingly, this board determines that the health, safety, and general welfare of the minors of the county could be benefitted by the regulation of advertising of tobacco products and alcoholic beverages on billboards near schools. 88-6.1104 Definition. 'Billboard" means any outdoor advertising sign (as defined in this code) which is at least forty square feet in area. 88-6.1106 Exemption. This article shall not apply to a billboard that is an accessory sign, as defined in this code. 88-6.1108 Prohibition of Advertising of Tobacco Products and Alcoholic Beverages: No person, firm, corporation, partnership, or other organization shall permit any nonaccessory billboard advertising a tobacco-product or an alcoholic beverage to be placed or located within 2,000 feet of any public or private elementary or secondary school within the unincorporated area of the county. 88-6.1110 Removal of Existing Advertisements of Tobacco Products and Alcoholic Beverages. Any advertisement on a billboard that advertises a tobacco product or an alcoholic beverage existing on the effective date of this article and that violates section 88- 6.1108 shall be removed within 90 days after said effective date; provided that, any person, firm, corporation, partnership or other organization having a valid contract for the placement of such advertisement shall be permitted to retain such advertisement until the contract is terminated or expires, so long as said contract was entered into and in effect as of and is not renewed or extended beyond its initial term. 88-6.1112 Abatement of Illegal Advertisements. Whenever the Community Development Director, or his or her designee, determines that a violation of section 88- 6.1108 exists, the Director may initiate the abatement procedures specified in the Uniform Public Nuisance Abatement Procedure (article 14-6.4 of this code), or otherwise specified in this code. SECTION III. SEVERABILITY. If any section, subsection, subdivision, paragraph, sentence, clause, or phrase of this ordinance is for any reason held to be unconstitutional or invalid, such a decision shall not affect the validity of the remaining portions of this ordinance. The board of supervisors hereby declares that it would have passed the remainder of this ordinance irrespective of the unconstitutionality or invalidity of any section, subsection, subdivision, paragraph, sentence, clause, or phrase. SECTION IV. COMPENSATION. Current law does not require the county to pay compensation to any owner or advertiser affected by this ordinance. However, should the current law be changed or interpreted to require that compensation must be paid due to any provision.of this ordinance, then such provision shall not be enforced. In the event that some provisions become unenforceable due to such chancre or interpretation of the law to ORDINANCE NO. 95- 2 require that compensation be paid, the other provisions that do not require under such change or interpretation that compensation be paid shall continue to be effective. SECTION V. NO EFFECT ON LEGAL OBLIGATIONS. Nothing in this ordinance shall alter or diminish any legal obligation otherwise required in common law or by statute or regulation. This ordinance shall be withoutprejudice to the enactment of ordinances to provide for additional regulation of billboards and outdoor advertising signs. SECTION VI. EFFECTIVE DATE. This ordinance becomes effective 30 days after passage, and within 15 days of passage shall be published once with the names of the super-visors voting for and against it the a newspaper published in this County. PASSED ON by the following vote: AYES: NOES: ABSENT: ABSTAIN: ATTEST: PHIL BATCHELOR, Clerk of the Board of Supervisors Board Chair and County Administrator By [SEAL] Deputy SBM (1/31/95) ORDINANCE NO. 95- 3 BOARD OF SUPERVISORS , CONTRA COSTA COUNTY , CALIFORNIA AFFIDAVIT OF MAILING In the Matter of Hearing on Proposed County Ordinance Amendment prohibiting advertising of tobacco products and alcoholic beverages on billboards within 2 ,000 feet of schools. I declare under penalty of perjury that I am now, and at all times herein mentioned have been , a citizen of the United States , over age 18; and that today I deposited in the United States Postal Service in Martinez , California , postage fully prepaid , a certified copy of Notice of Hearing to the following : to the attached list I declare under penalty of perjury that the foregoing is true and correct. Dated March 13 , 1995 at Martinez, California. a Ann Cervelli, -1)'eputy -Clerk 3 Steve Shinh e 3annett. Outdoor* Co. L69'5 Eattshore Highway 3erke,ley, CA 94710 Joyce White 548 So. 15th St. - 2ichmond, CA94804 .ass Caulfield 1 Francisca Drive loraga, CA 94566 Rich Hildebrand 1695 Eastshore Highway Berkeley, CA 94710 Nancy Andrews 138 Fountainhead Ct. Martinez , CA 94553 isa Korpus 444 Cherry Hills Drive afayette, CA 94549 oel E. White MD,FACR 15 La Casa Via #102 alnut Creek , CA 94598 udith Van Brocklin 7 'Santa Barbara Road leasant Hill , CA 94523 ancy Charles 374 Shoreview Ct. ay Point, CA 94565 ilie Freestone 26 32nd Street ichmond, CA 94804 Dave Jenne 433 Lejean Way Walnut Creek, CA 94596 PROOF OF PUBLICATION (2015.5 C.C.P. ............ STATE OF CALIFORNIA RECEIVED County of Contraa Costa !am a citizen of the United States and a resident of the County aforesaid; I am over the age of eighteen years,and not a party to or interested in the above-entitled matter. D �� � I am the Principal Legal Clerk of the Contra Costa Times,a newspaper of APR general circulation,printed and published at 2640 Shadeiands Drive in the City of Walnut Creek,County of Contra Costa,94596. And which newspaper has been adjudged a newspaper of general {i` ,1(BOARQ OF SUPERVISORS circulation by the Superior Court of the County of Contra Costa,State of California,under the date of October 22,1934.Case Number 19764. COM 6 COSTA CO. The notice, of which the annexed is a printed copy(set in type not smaller than nonpareil),has been published in each regular and entire issue of said newspaper and not in any supplement thereof on the following dates,to-wit: ............ ., .gypo../5' ................................I.............. ...... all in the year of 19_5sl — 1 certify(or declare)under penalty of perjury that the foregoing is true and correct. Executed at Walnut Creek,California. On this day of.r /Llr19. .� nature Lesher Communications,Inc. Contra Costa Times P.O.Box 4147 Walnut Creek,CA 94596 (510)935-2525 Proof of Publication of: (attached is a copy of the legal advertisement that published) NOTICE OF A PUBLIC HEARING BEFORE THE CONTRA COSTA COUNTY BOARD OF SUPERVI- SORS COUNTY WIDE AREA NOTICE is hereby given that on TUESDAY, MARCH 28, 1995? AT 2 P.M. IN ROOM 107 of the County Administration Building, corner of Pine and Esco- bar Streets, Martinez, California, the Contra Costa County Board of Supervisors will hold a public hearing to consider the following matter: Recommendation of the Contra Costa County Plan- ning Commission on a proposed County ordinance amendment prohibiting advertising of tobacco products and alcoholic beverages on billboards within 2,000 feet of schools. If you challenge this matter in Court, you may be limited to raising only those issues you or someone else raised at the public hearing described in this notice, or in written correspondence delivered to the County at, or prior to, the public hearing. Date: March 10, 1995 PHIL BATCHELOR Clerk of the Board of Supervisors a,n,d County Administrator By Ann Cerveiii Deputy Clerk Legal GCT 5422 Publish March 15, 1995 I Ti NOTICE OF A PUBLIC HEARING BEFORE THE CONTRA COSTA COUNTY BOARD OF SUPERVISORS COUNTY WIDE AREA NOTICE is hereby given that on Tuesday, March 28 , 1995 , at 2 P.M. in Room 107 of the County Administration Building, corner of Pine and Escobar Streets, Martinez, California, the Contra Costa County Board of Supervisors will hold a public hearing to consider the following matter: Recommendation of the Contra Costa County Planning Commission on a proposed County ordinance amendment prohibiting advertising of tobacco products and alcoholic beverages on billboards within 2, 000 feet of schools . If you challenge this matter in Court, you may be limited to raising only those issues you or someone else raised at the public hearing described in this notice, or in written correspondence delivered to the County at, or prior to, the public hearing. Date : March 10, 1995 PHIL BATCHELOR, Clerk of the Board of Supervisors and County Administrator t gy CL Ann ervelli, Deputy Clerk Substance Abuse Advisory Julie Freestone.; NOTICE OF A PUBLIC HEARING BEFORE THE CONTRA COSTA COUNTY BOARD OF SUPERVISORS COUNTY WIDE AREA NOTICE is hereby given that on Tuesday, March 28 , 1995, at 2 P.M. in Room 107 of the County Administration Building, corner of Pine and Escobar Streets, Martinez, California, the Contra Costa County Board of Supervisors will hold a public hearing to consider the following matter: Recommendation of the Contra Costa County Planning Commission on a proposed County-initiated addition to the Zoning Ordinance intended to further regulate cabarets . If you challenge this matter in Court, you may be limited to raising only those issues you or someone else raised at the public hearing described in this notice, or in written correspondence delivered to the County at, or prior to, the public hearing. Date : March 10, 1995 PHIL BATCHELOR, Clerk of the Board of Supervisors and County Administrator B Ann Cervelli, Deputy Clerk ;rDi I I zz mcg O - - 2C/) o� cn o o� m . ` C) ; i m C0 a : z� co C7 O co CN1 ; ! C 2 70 CD i # n Z m F- (p 0 (�j C m z gi u 11 w m o PS- m y � CD 0) �� _� _ � mi � C: I�fes} Q F m O o hU)=I- CD � 1 I [CD �Z �. (D ' ' m m CA ` `�' cn H ( tr tr I"t3 H m n o j °rot a c o (D Eg (D � (D sl ,h t� = z to S? Z3 N (DD C3 C m Un R' m m " Z o R� Fi fi O L'' �m C Sn mO O m D < s; N # tS Iv rt v U) -o m w _n C;a M c CD to w i� FJ €D �d r s l C D 50 D = vi rt k to m !- rt I ' m DZ co ' rn ; ��N� m cm cD uC ED O O �R O €-€ W oo m 0 O w-< v u) sZ 1 f] my m , wwC)rn Cl) 13' F W m 1 _ 5 fh 3 O D tS t0h i 1�-9 mn n caV0N z w rh O 0 w<m co N- 0 sy it Q „ � vc O o z m� w000 c rn !]� I O 0 n o m D W it t' n 11 -Cl)� i D .o O m C) N N E fl A� £j C: ZZ C I d Z rt j m ` < m .. Fi€ x €� ¢ C 0 H i v O TJ I to jar f r43 �J 9 w # 0 �Utt bo fD r fJ 1 O (D P� F_ w O {J m � er fz k� � K >✓ i Z tU P €l� ( y O ( En c t LTJ D H m j 0 £1, y O r ` ; m .. 00 0' CO-r, � { d Fl- rt �- I iN n U rt OI , io O 0 1 i ct N O m n ct I o p , N• p c 1 Z nc -I D ui O n N3 =+ cn x G)m cc) L m 1. r z z0 Cl)� �z3 rt p m (D ( ft 1 m Q1 fJl € W n 2 C D m vm rt cn s i _< O- 0 o _u O D ( € m v� M0 � m c H z ! O :0 m0 �D mco a o 0 p (D _u z m rj D td O ; 0 7r m 0 m ' m l d H ' iv tDt� r o -xc m I # C % N 0 t-j w m rn IDv 0 Z )o I I rt 1� # ' CO � O { me d I ly (D (� 1 I n W 0Fl 7� tr O tTJ Z I W (D c € Op O fZ �� CU m 3 1 z� D € Km }- LTI m 0 �O s i w z . t 1 s t CD m 1-, I u, ! < oWE. Orid's �� zat r aaMaib tza se 'tar. x t e Y;CY sped3' l�lbilhon axa `S+ m n x y> 90and�nearly, $22rbxilhon.b���� ,� �4 � � �. • y .You areqhe e y � attfled t�3 BRUaRt X21 ;995 at 7 30 •rt` ••xx c4, rM'"AG L'4r.t Rc - #n=family planpvng, repracluc , r d rn In cro[rt �VIcB ie N i{!IstMa B i�li i` �'t65 t PIn �=Street . altt$and MAIDS previthori china '" f r lbein `A artlnez 'a1IfO a o'lt l l COrntl1iSSi0n °wiN hone are�beur��urged to :�} � y� ►. i �g•�, , ute"oin tlydatbillron b��7'' r ' GOt1S�{Ia (/��SED fit Q 1 11011' 1C1IDeQ:aS�.fC?�C?WS .^..'•t.,,„y J.y, i ''eMi a " .`V nda 7 2lbxilxon��YMul,0 � tg p �r .s s-'So artice amen rnent{prohibiting 1 �t e�r;stn ib cc °cid ct andolevea "o'riifibciar swithir, . rease in spendingonpopula y - i geS4 r d ot:: arly si`���s `Gu�cSeptemhers�yy# _` ':2Q(} et Of Ch'OO�Srld v�i °ri et7r im@rtC�afion StO the .:Baal d` of �05 =T r „*• x 'ty.+ Nytk x ;,}::#! ,d i ,�`> •Y'"�k� . rt• �'+`� .; �- *Fra' a ti nCe are nOttencoUTa �< n Sti . 'rviSOCS. � .,,,3> ti:� �E F 7k_.r�" t ✓ t t 7� er, vi _pe {,x }' et xa i �*is 3 ' are,hav not beezi �ny new , t ra='?"rt?iu.6 s' a 3i a u:#7fa ,rN: q�',,, menu rtc1 m theabence of . � ial ones,the$I7 laillzon g�" � If yo khat engeathe.,pro] ct;ln,co , ,y u may �e,itmlted t`o raising oniy;.#Bose:. 3 : "'a uaiti0.i wt,.?- w'.'n,.aa.'v` .* r'{y: .:tlitt.>r"t•tY.b"kf 'r`9.. �.;. a- x, ;i'fia^..tJ.Aw,�,...s r,. r;._i� "4 ay.,,�tztra, n#?zi,w ax;t:f ^,}x ?..,:: .r. - iore'•e121.'ve tha3n eve 4 said �issues -ou°porfsom oris else aratsed; at-::t ie Aublic`hearin described .in this r.'. $,., r^:"k! .:•#. r �, F yk?"yF.r t,(j!,_N,,.a;t'.,,:• r' '; .k,. 32y�.w'J3% �;kx,',-f.•. yt �75te+'a ...,., -'r`,V+; ;^v7'i.a finding,a,membe rof the J S� r „ notices or In wri#ten Orrd ond-0 c delle ererd t the=Coon at;-,"or. rior`.to o��,4rd h i directoof x �� tY ;. P t:2.}�'.:r.i5. x 7t x # -y.. {'i .i4tvMr oaattlieFockefel t� jtlle public hearingt s t�� JJ .;r: f f_ <,v, S '{� y 4 ..:5 3r}: •E 7.�" a. w xY,isry nt•A,YE�A�.k :' {�i: iaation�� `lYew�Yore' 4, ' ITIS .1\ R r ^�a.`� ivr�... s k°a" ,; £s xr ..>r* ) �vember,donor nations met 5 t p, yY, a r "6'r"' =:. .y rye, �, 4• :s r,: `, oto:disciisss'hes-Cairo, lata' �` For#urtller,d2taiis,e Cohtact t l ntrd ti StajCOU'nty l-OmmUnity ra "tA�t '; ,l?;;„ .'F�f� .,E`. `''`;1["C 'ak'rv .. xis ;.tye +,ky,°;:. gyp, �+i'. '•y..., .„ ,.... .sfsr ;'r,: t 4, C90. ';comTllitte�d"itself Q{ ail >QeVelo int'¢D.e/M�'artmen " fUPt�eYStreetMartne�j Cairfo'rnl 4ly;5 ..1 "E_ A $ f lX ) k '-"EF"„ ",t,.. f tc:� .i r.4. j La j i I,-i•,i! y�x j .. L.,.._,r r -! [ AIN , •;���4 .� � mg, EN�II� t 51;0 f46�20311 ntriesraitrf .Il' y;M `xw �cSw � *.�: r. rK�'xMt`4:23;:' w+rssbu�37� t�1 ►rs,are going throust, ougtl C�p esri" lllh2l 'til 1 g1i1 R s x f. j. : r lh ibtitte'22IOT,e,tie VCR, y ion eSSlOnal SitE �idY��"�`1t1�'�n,�,�i� Janis Mo � LgCi ai }tc1�SAVM rV Bndin pr.oa�'gram ; • � e s a s8tn a e cx a to r a R. ca n t az i r k 3 t aid zSau ::�Sneir "br T yr & y"aw Y `F F x j�,�3 eq r�{ �yy�%'w 7 s tit B Kj tt s t Es...,,r `•,a. tn+:✓;r t.4f ,k.,Jnx��r. '3t`^ +h., :j73^ :.S,y •^l S'tS.^`,.0 .b.a{ �"Ii�,`.'.�.$ f......'£+It`tY'fF, 9 a�, ''t;t#.� r,.�$�^ 'b ...,z3Fn =.I',i� _ P•;N 4St ir�stratoratttie U^SJf1�°cn ' s c�i < 3 c` thE'r8h u t�Jtlfted �tn ['.FC �':. [[yy y, .<, _mat ;TE {:"",YCt_.,i t ?dt,':..;zi:A S�� :�S N+'�z%�1'._. Y.1V'�N.. yn:..,:' y; f,ti..�.,� br naa Cy� ,� • Yw.3 .tYlw'"jRS ,#n1..:'!.',t 5 r r ..at. ,.�a.J�ucvclut�,�►� ,� _., k inT HOo 412 Cl7 Mr Br,.len A"dt tfirtl tratlon Surldln X651. Fine ;Street ',t: cJh P b x -f y , IM, ° : !4�3 '� r a�rw .�A c �"� Ioomieiiforeca ollow�s ti arttnda aIifomi h8 C nt }'.P�annln COml711SSIOiI' Wlli .,v .! a,.:a.t,f tS"dF'w. r{`` 7..rru# 'x4Mry.....F; "'M 1 "'A,v:t ,:JR:.F» i ,Ri >6'r :.h iressive pledges�beforei s�* onsiCier a�PRCiSEQ E?RD�(tAIE aICatlonxde$crlbed'" s foilows: #Ferenc8 1r eK €; iy 13 7 is , d # f:s a S : C �e S7l;to �, a �'tt S #v�� s w r r xt t x dark? rx 1 r s qct N.- .y1+.1 rr' t ,iS{�;�'� i"•t #viiis ri r}? "`t.ut#i q;F,'£l�,nt f.`! H °fir'},#, ptti4 'Ye`'':} y{'YA'Y`r =#�,t rated States increased i fi i Tl tS;iS� ' '�COUnty#Itlitlated' adC�Ittonr,to theZoning ,O`rdinance `intended to ]_aid ffb0$448,millio tfi fI t .,t:r€ a :rr #+rr+:.� 11e 4,te11fltM"I",. D i 4 is,ta .-: �t. ° t a #. furt.er regUiate,cabaets# On January�sl 5 t i ' illipnln 1995 `' an 0� Y f rkr75 i99lt she Board of Supervisors.;': •�' d §^ t`#k° a b't y -:;S a -1r.J a„n. ,m rE .# fy a 1d �{ :ea $a :;;illi©nfove evenk directedhe t,�Community as Deionpm'ert Depirtmelntthe MPlanning J r � ,:sk•-±sa r, ,..`; t ` '� 3 k„=,,..... £tg. -c '..” qt sa;."�:.. 'f �Sr h'k`iBY.;:"Ia 1. '{"" $1�.,# 'Aa.° 1..x. t,:. Afli...,�t.. q 1 p3opulation�a AIDS; „ �T i ornm�s'slon, Lythe :;.Re/�"�ional ,Plannlrlryy ,C ommisslon:.s�z and.�the`�4St sniff s� + r W,c'wth?t;`r ,R, .` rs t�. .tt„+4cN y::+. ^J^yyf^ t:t+. x.5,.+ ,.:'$'^.: +ta 't:p."X`.st .rtx,^Pqt-Y"j, r _-..' 'wC FG'f "":at 'W yr.L. ...'ti4: x�y,.,:•J r .` v..,.;:. , rr� # t wDepartmentto review'and comment ujon the (jrd Deed Cabaret:,ordinance:# haSSaidl" bb0$t iu.. ' Y;x 4 sz .: 5 wm�f § iadSf �r f .+,.....t„3n,frra aa'..a:ltr E ,..1, t,z; zr> }µ _ • MTheBard askedhe Re ionalPlannin remissions�tOF o'entrahz >redttzan pe entand � £J �,: nn . •, a ^„meg' C,$ 7, ^, a .their ' ariattg{` �o'” recommen�c ations`#:t`o� ,tfl Cn Pil r Commissions fi4 The sBoard (' 11 peau Union £f 1£t.=i w,,:k"X.�x n`k ks thf 34�sy ytrpo ...**c°' £-t„.h...,x3: to rk"rr~w d1 !'ahy S.atT ,� - " .5�.����a'y� "k'--:. •f:3 .{;'•'a' ski c asesas ill tth ^ rsedA„fig�n re w z t Uest2Ci the OU #y Plannin ,A Issi rt b� t motly and a >. � .� , r�. a� siveIbic esti M: #¢� < . rovide thBoar i ti :I , '� _ _ .0 . . nll;take,mzichmare, u p.. p . . d i h . ,corpen d tOt #a ra�ldeso elatltude fora. . :td. y.:,�^ren..}++,' 3'kk ,1.!r, ,'f�6c' A aE..„t..� .:.k`a.'9,s”."crrkna.; -z,' .n tSi:'S='v.a *1X,..;�*w4" s moreforbFranceanda� 0- 1J �' h ` .. ,xm� finfreq ent c rltablpy ori other rlor pro# ,a tirrltie ,in #his aria 7The Board 12'5'times Itiit m ":w i # a3 t,Y=flta ,la*aY3•s .�• ".�c'z . fit a rr .- <a - ;:,, , , > ,. � . ftrtier requested #hatthe ,#Dun.' ;Pl�finlri �Cmmisslon consder•:annual ro proJechons by,�Papuia ��f.:YVv � t 'k�. "� w+a x�,reviev of all such A : Internahonalxt '<:r7a y.,y„1 r•'"'" ,`� ives,g Some:`activistsi U L g '; n�r§t. �v 2F} 'C!, Y #?t ,yp rA1� y.-���A.. t{:;i 0 efUl' t rti w. a q a 4t $3 €,s°l t� , p„ � tt Forftaitherl?detaits, contact°the:Co,tea Costa_�ouny�V';Communit\/ eta S.S.',r"et 9tw°e, =.,ki y {,� J.: ri $xdit�€�i5'gs° f. .;5,�,.s:? E,+ 7r.; 'A ,.;3°r v 'aJ. 'J t J l effign Venaat the.;:CoiI' �'l t: a t =Y3 _ ta” s " p �r •„ "x# ,Development.�;�Departrnent; 65'i PlneStreet,:Martinez, California, glven�th wrYd very b ;7 w: f L # at 5*10)"646-20911 , a ,x ;y ,r a ? u it £ ; i, �{or,phone DENNIS M BAiYC,, . Y es,Wve. Rely gaals,anax ; , r3 ,, , , �i ..& r '0. ita�z..,, .r. �-ssi fY.. ,trategies to reach those fta t gyp' s r 'r. ; k i z of vtc �J7t Joesph`�van�Ateridonk,a, � Br �d r ' Harvey;E rag on, Dr ,ector rr '+ i a '' s !p c '.� � t -( i {17A r ty'(t'# 1 z cutive kdirector at 'the t,# fr k it11�t „ z, y F• ContraGosta:Caun#y iontFizna9Y4.'> 4k ! 5 {i1 { 247 l �� �f 7 fi A� 7 Yi YJr 5 J`},'+�"` 'Y •` �°'ayt .l l Y < � {'4. t t ?,• : up to�the'governrnents f `' ;� ri k Commu�unity=Development Department a> k h m- t s� _£} ( S�SE`r7� i Id to come u with` a' s t ' oq tsYJTA y # ?•} - J � L � <` ..,..t-,_,{ a,itTy f` t ; .. 5?r� 7 'y ''y'j ;�T st �t a �}t t* - •. '. r tte�snatfr # xrr T ti > , u C�ant .. dA'cN 3 h.�{Ti! � NY3p t° i4�xl '7't� Y .^'F" �y 'f ,r f•s ?�' , 4 IBJ '#d b¢ ?7gy�t 1 . "♦,�,`' #;'�:. 'T� 31f1r,. ��it£`}Jrat`'�It��� � LWmErr oumom I March 21, 1995 Victor J. Westman �,�ar County Counsel --� M'EZ,caz�F Contra Costa County County Administration Building 651 Pine Street, 9th Floor Martinez, CA 94553-0116 Re: Proposed Outdoor Advertising Copy Prohibitions Dear Mr. Westman: Gannett Outdoor and other outdoor advertising companies and advertisers are all perplexed why the Board is rushing to adopt an ordinance that will force extensive and expensive litigation with no information or background on whether the ordinance really addresses a problem. The proposed ordinance is a direct prohibition of speech content. While some reasonable restrictions on commercial speech can be allowed, they must meet the strict four part test of the Central Hudson case [Central Hudson v. Public Service Comm`n, 447 U. S. 557, 100 S. Ct. 2343, 65 L.Ed.2d 341 (1980] and the numerous cases that have followed (cases that almost uniformly strike down the proposed restriction, e. g., Metromedia (now Patrick) and Pacific Outdoor(now Gannett) vs. the City of San Diego [453 US 490, 101 S. Ct. 2882, 69 L.Ed.2d 800 (1981)]. To determine the validity of a governmental restriction on commercial speech, a four part test was developed by the Supreme Court in Central Hudson: the legislation on a (1) legal activity will only be upheld if (2) it seeks to implement a substantial governmental interest, (3) directly advances that interest, and (4) reaches no farther than necessary to accomplish that interest (Central Hudson at 447 U. S. 563-566, 100 S. Ct. 2350-2351). This ordinance does not come close to meeting these tests: the County does not even know how many signs will be effected and how many have had tobacco advertising -- indeed, County administrators have asked the sign companies if they have this information. No court will defend an ordinance which directly restricts First Amendment rights where the County does not even know what it will effect. Of course, the County will have to take a national lead in defending a "finding" that billboard advertisements for legal, adult products are a direct cause of minors using illegal drugs and substances, such as heroin. The County has presented no information, surveys, or data to support this sweeping conclusion. GANNETT OUTDOOR • LEGAL COUNSEL'S OFFICE " l� 1311 TERMINAL STREET, WEST SACRAMENTO, CA 95691 (916)372-8252 FAX(916)372-8601 Victor J. Westman March 21, 1995; p. 2 This ordinance would be a governmental action interfering with a citizen's constitutional rights: thus, like the San Diego case, any legal action brought to challenge this ordinance would be brought under the federal Civil Rights Act, 42 USC 1981, 1983. This means that not only would the affected companies be entitled to recover any actual damages, but their legal fees as well. The City of San Diego paid the two outdoor companies almost $500,000 in 1982 -- a similar action today would be appreciably more expensive. Of course, this is on top of the County's own legal expenses and any damages the companies are awarded. To our knowledge, all of the outdoor companies in Contra Costa County comply with the Code of Practices of the Outdoor Advertising Association of America, and do not allow any tobacco or liquor advertisement within 500 feet of schools. We have never been contacted by anyone in Contra Costa County that this has not been an adequate policy. However, Gannett Outdoor would like to avoid litigation. The direct economic effect of this ordinance in Contra Costa County is unknown: the county has not told us which (or how many) signs might be effected. However, we cannot allow the heart of our business --the right of advertisers to use our signs -- to be illegally limited through legislation, even if the direct effect of this ordinance was not substantial. Sometimes, the ability to advertise in certain areas is key to an advertisers whole campaign -- the damages to Gannett could be much more widespread than the loss of just these signs. As we demonstrated a few years ago in Richmond, the outdoor industry and its advertisers are prepared to mount a serious defense of our rights, even when the direct effect alone may not justify that expenditure in that one instance. We believe we would be joined in any litigation by numerous advertisers, advertising associations, and perhaps free speech advocates. Rather than commence an adversary proceeding, we would like to work with the County. While we cannot agree signs within 2000 feet (just under four tenths of a mile) of a school create any problem, perhaps the easiest method would be to relocate these signs. We would propose the County issue a new permit for each sign in the target area, and the companies will move their existing signs to other commercial or industrially zoned areas in the County. This should be a "win- win:" the County eliminates any chance of tobacco/liquor advertising within 2000 feet of the schools, while the companies keep the same number of signs. Practically, this will eliminate the on-going problem of patrol and enforcement the County would otherwise have to undertake to enforce this ordinance. We would Victor J. Westman March 21, 1995; p. 3 be willing to discuss any reasonable restrictions on the relocated signs, and/or any "formulas" for relocation. We would ask the County to postpone action on the proposed ordinance to consider this alternative proposal. We would like to work with you to avoid litigation. However please do not take our willingness to discuss alternatives as any weakening of our resolve to do whatever is legally necessary to defend our constitutionally protected rights. Thank you for your attention to this matter. Veryyours, R na W. Beals Vice President/Legal Counsel cc: Phil Batchelor, Administrator and Clerk of the Board MAR-27-1995 16:52 FROM FINOLE LIBRARY TO 6461059 F'.01 3241 Colusa Street Pinole, CA 94664 March 27 , 1995 Via Facsimile Transmission_ to; 546-1059 Contra Costa County Board of Supervisors 651 Pine Street Martinez, CA Dear Members of the Board , I understand that a proposed ordinance to prohibit alcohol and tobacco advertising billboards near schools will be on your agenda tomorrow, March 28. I strongly urge you to enact this ordingpce. The tobacco and alcohol industries together spend billions of dollars each year to promote their products. A significant portion of their advertising is designed to attract young users. This kind of advertising has no place near our schools. I hope you will stand up to the tobacco and alcohol industries on this issue and enact this measure. I hope that the cities in Contra Costa County then will follow suit by enacting similar ordinances . Thank you very much for your consideration of my comments. very truly yours , Gerald H. Schlintz TOTAL P.01 10 r MAR - 28cr 5 TUE 1 5 24 SL4F� u _ DE Soo I h i Br P i0 1 © rrr W p. LU a U Q � COUNTYWIDE YOUTH COMMISSION j 8c3 2301 Stanwell Drive Concord, CA 94520 LLI (510) 646-5763 U TO: Board of Supervisors FROM: Countywide Youth Commission DATE; March 7, 1995 SUBJECT: Recommendation on Proposed County Billboard Ordinance On February 22, 1995, the Countywide Youth Commission unanimously voted to recommend to the Board of Supervisors adoption of the proposed ordinance to ban billboards advertising alcohol and tobacco products within 2,000 feet of schools. The Youth Commissioners also had for the Supervisors' consideration a few additions to the ordinance as stated In the following summary taken from the minutes of the February 22 meeting. Commissioner Long stated that he attended the Public Hearing on the issue on Tuesday, February 21 . He suggested that the ordinance should include store fronts near schools to also be required to remove tobacco and alcohol advertising. Commissioner Hansen agreed that there is need for the ordinance but mentioned that peer pressure is the main reason youth are involved with tobacco and alcohol. She cited that her town (Danville) doesn't have any billboards and she still knows kids that drink and smoke, Commissioner Kefer agreed with the proposed ordinance and disagreed with Hansen, stating that any step towards helping youth avoid tobacco or alcohol use is a good step, She also noted that advertising is a form of peer pressure. Commissioner Zusman stated that peer pressure is a problem, but there are many programs In the county that can help youth resist peer pressure. The DARE program was given as an example. It was asked how many billboards the ordinance would effect. Supervisor DeSaulnier said since it would apply to only the unincorporated areas of the county, only four to six billboards would be changed by the ordinance. MAR - 28 - 95 TUE 15 : 2S SUP -,.- . DeSCL " I r� i er P 02 'o- Billboard rBillboard Ordinance March 7, 1995 Page Two often spread to the cities and that it was their responsibility to bring it to their respective councils. Commissioner Long suggested that these billboards be replaced by positive messages. Commissioner Hansen questioned the amount of money it would take to change these billboards and asked if the money could be used for programs in the schools as an alternative, Supervisor Torlakson stated that the ordinance would be implemented over time, leaving the companies who own the billboards time to find new buyers for their advertisement space so there would be no cost to the county. He also said that there are many tobacco and alcohol prevention programs currently in piece. Commissioner Bennett said that billboards near schools serve no purpose because people in school cannot legally use their products, Commissioner Smith then motioned to recommend to the Board of Supervisors adoption of the ordinance. It was seconded and unanimously approved. lvlb DATE: April 27, 1995 TO: Jeanne Maglio, Chief Clerk of the Board FROM: Christine Wampler, Secretary RE: Dr. Joel White's Original Research Documents/H. 6 3-28-95 On Wednesday, April 26, 1995, I spoke with Dr. White regarding the return of his original research documents which were submitted to our office on 3-28-95 for Hearing Item H. 6. Dr. White stated that he requested the return of his documents when they were submitted, however, they had not been returned. After researching our files, it was determined that Dr. White's documents were in the possession of County Counsel. County Counsel was asked to return the documents to our office as soon as possible. When the documents were returned I telephoned Dr. White, and per his instructions, I mailed the original documents to his office at the following address: Dr. Joel White 115 La Casa Via Suite 102 Walnut Creek, CA 94598 On Thursday, April 27, 1995, the package was picked-up by a County courier for delivery to the Concord Post Office. r J� Is Th eek - = fie = a Push-_`,�mcja a Q 0_*__-­ hildhood is a time for imagination, but last fall 600,000 fifth-graders got more fantasy than was good for them in a g respected children's publication. Defying mountains of � � � ,?1 0� � ,���ia� �a ���1�@�'U�ul� scientific evidence, eek 'Reader left you audience la , ,U'r� NUU lfJ , with the impression that ere;is some qu -on c er �0 0 M1 1lV 4 1:1�J � 0f uuomN� ill , America would be better oSvthout cigarettes �� . .T. r . 0� mQ W , It is not as if the Render actually promoted '" Its r o sins were more tle' iE-no less eye-popping in a now � increasingly awake#"the risks of tobacco. Their Oct 1994 article "Do Cigarettes Have a Future," distribu classrooms aro the country, interjects"balance"where plaritiyely, the art�rle claims, "mane tobacco growers exists and contr ersy where none is remotely warranted �x „ iiowtfmdemsel�esthout a avay to earn money. Typical of t resentation is a chart thatpure s to �j6 `several +usan c�bacct�growers may well sound describe the costs d.benefits of a smoke-free nati a xke a lot6# td3mied farixirs town l I year old,but the figure concept that is itself a red herring, since neither Com deservesstfrio�iext ren for adults. As the the Occupa al`Safety and Health Administr initis proposed a t►onwide smoking ban. Incredibly, m ' �voitds r 'El oral lies arc ervice a3nts +ut ui a`March 8, 1994, IW ort, Not al ©b ass om khe„-o cco'business is trivial PYA are expended possible costs of a smoke-frees ty than a national, er cti and>onl Haar anally a re Tonal the benefits. S ch as: "Nearly 47,000 people who " k for z c' ," g g _ saner Yaf�cp�ieeiniel3ott s pare t�f tobacco growing, y aaetioniiig and maiiufati ing lobs aie concentrated in just three states Nt� Cart3hria, 'Ua a and Kentuckl' Only in r.. f ICentueky ties the ren of total state employment in §` tc�ba4-reagdos tireereent (barely). In North U Q 1 i trarcilina he�pe� is ij h#1�+below three percent, M c�a � aoaQ � � z� glue nit`evesite gait c ohs are threatened anywhere r _ y�vha# tleultts, mtt, dcts a responsible look at ° kSi ala ase;In est potential job tobacco companies would lose their jobs. . . vie devastation to health tobacco growers would have no one to sell thBaked hibacco? Sens#s" o riot, "after all, merely Farmers would not be able to earn the $3 billion that eco .abelieve" re3s akhnktvekenkr�kmg'and a shortened life crops bring in each year. . Taxes on cigarettes haver "t ias=eieelatio m housands of studies. nearh S 12 billion that governments have used For such things as _ s "' ose5 aEhe W ��Reacler article Mmes,t� quantifying the health care and school funding" resulting St tci society><s fts statement that "some studies The benefits ledger, meanwhile, is couched in lame language suggest bi asf da5llai= " are'spent on smoking-related more appropriate to a discussion of cigarettes 35 years ago. " Mane scientists believe there is a link between smoking and a max, shortened life span,"the Reader says, so"a nation of nonsmokers � v� would probably live longer lives." "Manu"scientists?"Probably"? 'S i 0 WIQQ 0 G " "' C Instead of presenting an unflinching look at the dangers of i � �1�@Q A, U0 e Qj 0 0 �� a0 smoking, the 11'eeklr Reader offers a subtle brief for smokers' A0A O���ii m IN I, M ub IN& y VM rights and tobacco employment. Even the cover sports a color JJ�IJ, � �,,� 1lWls UV lJ �1��y, photograph of angry tobacco workers protesting high cigarette t l ��TU�,&N to Ujfl taxes and smoking bans, while the inside text repeatedly 1 i r � describes a decline in smoking as a threat to jobs and business. Priorities•Vol 6 No 4• 1994 13 illnesses. Billions, indeed. Try S50 billion per year. That is the misunderstanding. . . I would take issue %vith anv treatment of estimate made in July 1994 by the Centers for Disease Control my work that suggests that it supports the view that there are and Prevention (CDC) after the most extensive survey of its not significant external costs associated with smoking, or that kind, involving 35,000 people. major increases in tobacco taxes are unjustified." Six months earlier, an article in the March 31, 1994, Nei• In an interview for this article, Manning expanded on this England Journal of Medicine recounted in copious detail the still thesis. He noted that his studies, while accurate so far as they growing evidence of tobacco's toll. It includes one-fifth of go, did not include the medical ts-of lo« birth weight babies deaths due to cardiovascular disease, 85% of all lung cance�sr , or passive smog, which In ie�a Nould be "substantial." 30% of all cancers and tens of thousands ;f-Aeadis jfrom yeti�f 4-fie cigarette es'did cover external costs, pulmonary illnesses such as bronchi_ d emphysemahNo Manning added, higher taxe oWd still be justified on other such data appear in the Week1 rtr meed, fbetacle onnds,'including the deter ;eue'ta teens who might discusses smokers' ri t�pa€tfta�ceshe �tl;erwse take up the habit. tobacco industry �my mind," said Manning, "people don t appreciate the smoking. r> k schen they make the decision to smok weir teens." So deficien ._ the article xhat tt pmmpted Z�omas.Brandt, a spokesmR wealthi3Q tUUU`J(g ; Washington,tc wonder aloui whether nts slaztt might have J ,,, �, o somethin to d©with the fact that e Ye ly leder as ovcrned by 0 0, J%W , �1 �1 At n K-III C( muncabonsfa tout of Kohlberg ICravns Roberts & �1 0�% ��o � �1� Co., th argent sharehalder of SJR Nabisco-' randt's spicion C� 1 0 � I' (l 0 , A 0 '' may we�be outiandLsh Reader executives certauily deny n " 01. e GI 031% —but fact emains that ri€ocusi�g on smokers'nghts lost 7b i 52 jobs andNe�`f iiias�'rsf ligher taxes; the Acle does mimic l% a �Q dp ilWQ 'Ju lbUUo recent to s-flf>the ctgarette{lobl�y to tts battles aroundi.,the country a st=srnoku>g restrictions and ltzglr tobtacco tares Here we arrive at the nub of the pr em with the Week]}- Among the dustry's most frequent elanns, for example, is Reader's approach. It was directed, afte l; at.'kids about to embark on the eery years when 80%to 90 of future smokers take up t the habit. According to th QC's Morbidity and .. _ _ anahiy Weekly Report of August 19, 4,"The average age at " vluch smokers try their first ciga tie=is 14 1/2 years and 70°l��f-sxi�c�kersne regular smokers by age x $ ears" Iedna 'essaesilayba substantial role in their y e 0 behavior-too since: "The three most commonly purchased 0�r� l -�` 0 , brands among adolescent smokers were the three most heavily 11 0 „n �J advertised brands in 1993." Significantly, among the three was 1 Camel, a brand virtually ignored by children until the 11no introduction of Old Joe cartoon advertisements in 1988. Nor �{.a 0 was that the first time kids responded to cigarette makers' blandishments. The tobacco industry's decision to target female hold be �,� smokers in 1967 resulted in a clear increase in the addiction of �!Ri ` ' . girls 11 through 17 years old, according to an article last w February in the Journal of the American Medical Association. It has been estimated that 3,000 children start smoking that tobacco taxes alr moking's "external costs" every day. Given time, no doubt, they will accumulate on their once pension, Social Security and other"savings" from shorter oven all the arguments they will ever need regarding smokers' lives are factored into the equation. rights. What they really could use novo,before their addiction is As it happens, there is considerable irony to such an fixed, is a greater appreciation of the risk they have embraced. argument. The principle author of the studies on which the That was the opportunity fumbled in the Weekly Reader. claim is based has repudiated the exploitation of his research. Willard G. Manning, Ph.D., of the Institute for Health Services Research at the School of Public Health, the University of Minnesota, wrote the director of the Congressional Research Service in February 1994 alerting him to the fact that, "my VINCENT CARROLL IS EDITOR OF THE EDITORIAL PAGE AT DENVER'S work on this issue has been the subject of a great deal of ROCKY MOU Tti.v NTEIVS. 14 Priorities•Vol 6 No 4• 1994 HOOKED ON TOBACCO�. THE TEEN EPIDE-M- -C The ver the past three de- age smoking rate--even though way to fight the nation's leading pre- r cades, the number of nearly $150-million of the money ventable cause of death, cigarette smokers in the U.S.has went directly into school programs, EvogbWs"ng it industry "° slowly but steadily de- and much of the general campaign needs kids chned.It's been a true public-health was targeted at adolescents. "I was 14, at summer sleep-away success story: For millions of ex- The voices of medical reason are camp,with all these girls my age who to smoke. smokers,the drive for self-preservation up against forces that can overcome were from New York City.They all It's getting has proved stronger than the addic- teen-agers' fear of disease and smoked and I thought they looked g tive power of nicotine and the per- death—at least long enough for the so cool. Smoking was one of my plenty. But suasive power of the tobacco indus- teens to become addicted to nicotine. first rebellious acts, you know? All parents— try. But now that progressive trend The tobacco industry has waged these really cool girls were doing it," p may come to a halt. The drop in a relentless campaign to recruit That's how it started for a subur- and the rest the nation's smoking rate is leveling smokers; in spite of being banned ban Connecticut teen-ager well call of us—can adult and for a chilling reason: As from the airwaves, cigarettes are Allison M. {because she believes her adult smokers quit,there's a steady advertised more heavily than any parents don't know she smokes).At fight backsupply of teen-agers stepping up to product except cars. first it was an occasional thing. By Second Of replace them. Despite the industry's protests senior year she was up to a pack Every day, 3000 American teen- that it has no such intentions, the a day. Now an 18-year-old college tW0 partS agers reach adulthood as confirmed campaign has clearly snared kids, freshman and finally buying her cigarette smokers. That's roughly not only with the infamous Joe Marlboro Lights legally, she knows equal to the number of adults who Camel,but also with a host of other she's addicted, although she never give up smoking or die from the dis- print advertising,merchandising stra- thought that would happen."I don't eases it causes. Most teens started tegies, and even direct-mail cam- think I could quit right now," she when they were legally too young to paigns. Add in the power of peer says. "I don't want to have to go smoke; kids under 18 smoke an pressure and the willingness of store without a cigarette.I get so irritable estimated 17 billion of the 500 billion owners to sell cigarettes to minors, if I can't have one when I want to." cigarettes sold each year in the U.S. and it's no wonder that so many Allison's story is typical, says the Teens are the primary source of teens are hooked. new smokers; after they turn 20, There is much at stake here.A sig- R almost no one starts. nificant drop in teen-age smoking "Since the first Surgeon General's would cut deep into the tobacco T public health industry's main source of new cus- p 'n 964 he report t 1 p ttY movement has been very successful tomers. But from a public-health in convincing adults not to start standpoint,keeping kids away from Cool? smoking," says John Pierce, an epi- cigarettes is the single most effective Joe Camel and demiologist at the University of California, San Diego. "But we've other ad images ; {see rl ht} had very little impact on kids." also late Today, about one in three high smoking with school seniors say they smoked in being stylish, the past month, and about one in , i sexy,popular, five smoke daily.That's particularly Independent, alarming because,unlike the teens of and thin. 30 or 40 years ago,these new smok- ers know full well how dangerous _ r cigarettes are;they've been hearing "� e for years that tobacco is addictive and deadly. d Antitobacco and public-health ad- vocates have tried to adapt their messages to teens, using school programs and aggressive ad ,r campaigns. But they have little to show for their efforts. In California, for instance, a $600- million antismoking campaign t cut the overall smoking rate by an impressive 30 percent.Yet it couldn't make a dent in the teen- 142 ; r, University of California's John Pierce. lar among kids because"they're the When smokers start "The overriding thing is the image of strongest and smoothest.") Nearly every smoker began as a teen-ager. being cool. They all think they're Personal independence. One For a group of people in their 30s who had not going to get addicted, that they of the main developmental tasks of been daily smokers at some point--or still can stop—but they can't." adolescence is to assert indepen- were--this graph shows their age when they Of course,teens are growing up in dence from one's parents by Ion a culture that's long been perme- structing one's own identity. Marl- first Smoked. ated with attractive images of smok- boyo has successfully exploited that ers and cigarettes. Hollywood does need for years with its cowboy alone 30% its share, but it's cigarette makers on the range.A Virginia Slims cam- who have so thoroughly wallpapered paign last spring said the cigarette our world that their ads and logos was"as free-spirited as you." are inescapable. The images they Weight control.This theme plays 20% emphasize--in magazines popular into many teen-age girls'preoccupa- with teen-agers, on the billboards tion with thin figures. Mislys are they see at sporting events and on "slim and sassy." A Virginia Slims the way to school, on the sides of ad says,"If I ran the world,calories 10% , buses,and in the doorways of conve- wouldn't count."The models in the t nience stores—resonate perfectly advertisements are extremely thin, with the psychological and social wearing fashionable clothes. Even needs of adolescents: the cigarettes themselves are extra- 0% Social acceptance.Cigarette ads slender.Fashion and celebrity maga- 0.11 12.13 14-15 16-17 18-19 20-24 25.29 30-39 play to the craving for popularity. zines that help promote an ultra- Source:U.S.Centers for Disease Control and Prevention. Newport ads invariably feature confi- slim beauty ideal carry plenty of cig- dent-looking young couples having arette advertising but rarely speak fun together,often with an erotic sub- out editorially about the dangers of garbage is stripped away,successful text Camel ads call Joe a "smooth smoking. cigarette advertising involves show- character'; with his saxophone, his An advertising executive who had ing the kind of people most people panache,his Ray-Bans,and his trendy worked on the Marlboro account would like to be, doing the things wardrobe, he looks it (At a high was quoted in the 1994 edition of the most people would like to do, and school we visited,one boy who appar- annual U.S.Surgeon General's report, smoking up a storm. I don't know ently had absorbed the ads'message which for the first time focused any way of doing this that doesn't volunteered that Camels were popu- on teen-age smoking."When all the tempt young people to smoke." T-shirts and tattoos Ads are just a small part of the tobacco industry's campaign. Cigar- ette makers spend nearly half their l marketing money on so-called"value added"promotions like coupons and �� premiums. Some you can pick up If I ll e ` 11 right at the store with your cigar 4h� WO � , t ettes. With two packs you get a 1 YY 1OTflashlight or can holder. With four } r t � � � grr you get a baseball cap.With five you ccel1 ,.x. get a T-shirt. Wo'T i n f Cigarette makers also rely heavily 4l! tt on catalog promotions. Customers Vtttt � � amass bar-codes or certificates-- tt ': "Camel Cash" is the best-known example—to trade for merchandise. HEN Marlboro has sent vans around to convenience-store parking lots so customers can redeem their"Marl- Li baro miles"on the spot.Many prod ucts in the catalogs would appeal to adolescents. A Virginia Slims fw 33 SIMPLY "V-wear" catalog last vear fea- � ; !: x tared temporary tattoos.A"Camel 3� . Cash" catalog included a suede 1. baseball jacket, a beach towel, and a charm bracelet. Though the catalog order forms u require customers to state that they're over 21,these products do find their way into the hands of minors.John Slade,a physician at the University of 143 Marlboro no �6ba it. Medicine and Dentistry of New Jer• percent of 3-year-olds,and 91 percent likely to change their minds many sey who studies tobacco marketing, of 6-year-olds, could match the Joe times before adolescence. conducted a nationwide phone sur- Camel cartoon character with theto POrf110dOd ? vey of children age 12 to 17. He cigarette it was promoting. At the found that 11 percent owned at least same time,Joseph DiFranza,a phy- If you ask teen-agers, they will one promotional item. sician at the University of Mass- insist that they could never,ever be Mail-order promotions have en- achusetts who researches tobacco influenced by an ad. "I personally abled tobacco companies to build promotion, found that nearly 98 don't think advertising had that enormous direct-mail lists of smok- percent of high-school students rec- much bearing on me," Allison M. ers. Slade's poll found that 7.6 ognized the character, versus 72 says."I can't believe that seeing Joe percent of teen-agers had received percent of adults. (We saw this for Camel on a billboard will make a cigarette companies'mail addressed ourselves when we visited a 10th- 15-year-old think, 'Oh, Joe Camel directly to them. Extrapolating to grade health class and showed stu- smokes cigarettes so I should,too.'" the entire U.S. population this age, dents a Joe Camel advertisement. But Slade says this attitude is to be Slade estimates that there are 1.6 Only one of them said he had no idea expected. "Advertising at its best million teen-agers' names on the who the character was—a boy who leaves impressions and influences companies'mailing lists. had immigrated from Sri Lanka two people without their noticing it,"he weeks earlier.) says. "It doesn't surprise me that's Who's the target? R.J. Reynolds denied targeting what people say. But look at what Cigarette marketers have a di- children with the Joe Camel charac- they do." lemma.Their industry code says they ter and hired Richard Mizersky, a In fact,there is abundant evidence must aim their sales pitch at adults University of Florida marketing re- that cigarette advertising effectively —but market research shows that searcher, to repeat the Fischer reaches kids before they're even nearly all smokers start smoking, study. Like Fischer, he found high out of middle school, shaping their and become loyal to a specific brand, recognition rates.But he also asked perceptions and behavior. Several before adulthood. the children whether they liked studies have found that adolescents The companies insist their mar- cigarettes, and found that "there's consistently overestimate the num- keting efforts don't intentionally tar- clearly no link between recognition ber of people who smoke—thinking get teen-agers. The $5-billion they and liking." While 41 percent of 3- that smoking is more socially ac- spend every year on advertising and year-olds said they liked cigarettes, ceptable than it is—and that young- promotion, they say, is intended to less than 4 percent of the 6-year- stern exposed to the most advertising promote their brands among adults olds said so. overestimate the number the most. who already smoke. Other researchers,however,point When kids begin to experiment with But the tobacco companies'intent out that the response to a straight- smoking, they experiment with the hardly matters. There's abundant forward. question like Mizersky's most heavily advertised brands. evidence that, whether or not the may not predict behavior. Even if That certainly appears to have companies plan it, children of all young children tell a researcher been the case with Camels. Here's ages are paying close attention to they don't like cigarettes, they are what's known. In 1986, the brand's their messages. Nothing captures this issue as S,�R:j 9w13 tn�.��;1"3uvz, �.: well as R.J. Reynolds's Joe Camel. �. - ►--``�. In 1991, University of Georgia re searcher Paul Fischer found that 30 ` Ad dollars and teen smokers " Cigarette makers say their ads are aimed at r ` ` x >► adults, not kids. But the more money spent t t on print and outdoor advertising for a partic ular brand, the more popular it is among �. ,; r underage smokers. Adult smoking prefer ences don't follow spending quite as closely. t; Cigarette Major-market ad Market share Market share *, brand costs,In millions ages 12.18 U.S.overall ' VI Marlboro $75.6 60% 24% 1 ' Camel 42.9 13 4 Newport 34,5 13 5 ' �5 �b oromotl0nal goodles ' Kool 20.5 1 3 s i ,Companies spend much of their - _ Winston 17.6 1 7 k�advertfsin mone on Clbtht and °=.- ear boxed for sale aloe with a Sources:Maxwell Consumer Report,Competitive Media ti y r xr� Reporting,U.S.Centers for Disease Control and Prevention, tgaretle paCliS All figures are for 1993. "r F 144 CONSUMER REPORTS MARCH 1995 market share among 17-to-24-year- Everywhere olds was less than 3 percent. In Cigarettes are one of the 1988,Joe Came]was introduced.By most common 1989,the market share among under- products on billboards, age smokers was 8 percent.By 1993, which help wallpaper the share among underage smokers the Outdoors with had risen to 13 percent. tobacco images. Similar effects have been seen —_ - before.John Pierce,of the University of California, has used decades' - _ worth of Federal health-interview data to reconstruct the change in cigarette brand shares over time, all but universal.The next steps in- shown, tend to share certain risk and found they consistently correlate elude believing the ads promote factors: low socioeconomic status, with ad campaigns. For,example, smoking's benefits,having a favorite poor school achievement,excessive the percentage of teen-age girls who ad, having a favorite brand, and rebelliousness and risk-taking, low started smoking regularly rose either owning a cigarette promo- self-esteem, dropping out of school, sharply in the late 1960s with the tional item or being willing to wear and not planning on college. introduction of Virginia Slims, the one. Kids who score highest on the None of this appears to apply to first cigarette brand designed and scale are four times as likely as African-American teen-agers, how- marketed specifically for women. the lowest-scoring kids to say they ever. In a cultural shift that has ' During the same period,the rate for might try a cigarette if one of their caught public-health authorities by boys didn't change. best friends offered it. surprise, the smoking rate among Many studies have established that black teens, which once matched i Peer influence having a circle of smoking friends that of white teens, has dropped What's happening, Pierce thinks, powerfully predisposes children to steadily since 1976. The percentage is that "advertising makes people experiment with smoking.Based on of black high-school seniors who susceptible" to experimenting with interviews with some 5000 Cali- smoke daily is now just 4 percent, smoking."But once they're suscepti- fornians aged 12 to 17,Pierce found compared with 23 percent among ble, advertising doesn't make them that teen-agers who had best friends whites. Health researchers are furi- want to experiment. Exposure to of both sexes who smoked were 13 ously seeking an explanation. "We peers does." times as likely to have smoked want to bottle it,so we can sustain it Based on polling of California within the past month as youngsters for black teens and pass it along to schoolchildren, Pierce and his col- without smoking friends. By con- white teens,"says Michael Eriksen, leagues have developed a"smoking trast,having a smoking family mem- director of the U.S. Centers for Di- susceptibility" scale that captures ber didn't even double the likeh- sease Control and Prevention's Office this process. The first rung on the hood of a youngster's smoking. on Smoking and Health, scale is awareness of ads—which is Fortunately,four out of five people make it through high school without Easy to buy } i'� becoming daily smokers.Those who Teens primed by ads and sur- a{,u _� � do start smoking, research has rounded by friends who smoke just h� T ,ls'denim Jacket,from the: �f +Aarlboro catalog, requires buy x Where new smokers come from 3200'.C,igarettes= a pack a._..;. The smoking rate among adults, having dropped for years, has leveled day foe nearly.: off now. The adult smokers who quit or die are being replaced by a half a • r ka steady supply of young people who began smoking in their teens. Y ear. Percentage of high-school seniors who said they had smoked in the past 30 days. 40% 35% -- � Ei 30% 25% ' 1976 1980 1985 1990 1993 Percentage of adults who said they were regular smokers. � F 40% 14 v� C r,L.��� 1976 1980 1985 1990 1993 - Sources:U.S.Centers for Disease Control and Prevention;'Monitoring the Future,'University o/Michigan. CONSUMER REPORTS MARCH 1995 145 need a ready source of cigarettes stores.They were able to buy cigar- Amendment grounds would be likely, to allow them to progress to outright ettes at 137 stores, including six of but many legal scholars believe the addiction. No problem. Although the seven displaying the sticker. Supreme Court would uphold such a every state outlaws the We of to- restriction. What: bacco products to anyone under 18, The tobacco industry argues that study after study has documented According to the latest data,half of teen smoking rates haven't changed that those laws go almost com- all lifelong smokers will die prema- much in countries that have enacted pletely unenforced.Some stores are turely—an average of eight years such bans. But the most compre- careful not to sell to underage teens, early.For that reason alone,halting hensive review ever done on the j but not enough;"I can count on one the teen-age smoking epidemic must subject,by New Zealand health offi- ' hand the number of times I've been be a public-health priority.But how? cials in 1989, found that teen-age carded for a cigarette,"says Allison. Last year, an expert panel of the smoking rates declined faster in The youngest smokers often say National Academy of Sciences ad- countries with stricter ad bans. they buy their cigarettes from vend- dressed the question and recom- Advertising bans would be more ing machines. mended three major strategies:Ban- effective, too, if marketers didn't Under public pressure, cigarette ning nearly all cigarette advertising find so many ways to evade them. makers and the Tobacco Institute and promotion,raising cigarette taxes Canada banned ads,but left one loop- have attempted to address the prob- to make smoking less affordable, hole: Companies could still sponsor lem in several ways; for example, and enforcing the laws against sell- cultural and sporting events. The they give tobacco retailers stickers to ing to teens. companies do, and make sure that post by the cash register reminding Advertising. Ideally,we'd like to every poster, advertisement, bill- teen-agers that they can't legally see Congress ban all cigarette ad- board, or program contains a large buy cigarettes.But the stickers don't vertising and also all promotional image of the brand's logo and name. appear to make clerks less willing uses of cigarette-brand logos. Many In European countries, companies to sell to teens. Joseph DiFranza, tobacco-control activists have sought barred from advertising their cigar- the University of Massachusetts re- such a ban,for years, and it often ettes simply advertise clothing, searcher, took five underage teen- gains majority support in public- matches, or cigarette lighters that agers to 156 central Massachusetts opinion polls. A challenge on First bear their logo. CHERRY FAVORED TOBACCO BITS ANOTHER MARKETING TRIUMPH Most kids have tried smoking tobacco.But few of them elementary school.Few girls have tried it. sucked on it—until after the "smokeless tobacco" The way snuff is packaged and formulated helps new industry launched an aggressive, 20-year campaign users to gradually take up the habit and advance to prod- f to attract new customers.The promotional effort—the ucts with stronger tastes and higher nicotine absorption, U.S. Surgeon General said in a report last year— the Surgeon General said.U.S.Tobacco's product line succeeded in greatly increasing the product's use includes Skoal Bandits,mint-flavored tobacco wrapped among the young. in teabag-like pouches that keep the fragments from Most smokeless tobacco is snuff, which users tuck floating around in the mouth.Then comes Skoal Long next to their gums. In 1970, snuff was an old man's Cut,a loose,coarse-cut product somewhat stronger in habit. Federal surveys found that 3.4 percent of men taste and nicotine, available in such flavors as winter- over 65 habitually"dipped,"compared with fewer than 1 green and cherry.The company's best-selling product is percent of young men aged 18 to 24.Then ads and pro- the fine-cut,nicotine-laden Copenhagen. motions from U.S. Tobacco—which dominates the Addiction is common. The nicotine readily enters snuff market—started associating the product with the bloodstream through the mouth's mucous mem- rodeos, rock stars and monster trucks. Like cigarette branes, and users who try to quit suffer withdrawal manufacturers,snuff makers vehemently deny market- symptoms just as severe as a smoker's. ing their products to underage users. But some ads Bathing your mouth daily in an alkaline tobacco appeared to "target male adolescents," the Surgeon solution has health consequences far more immediate General said,"by providing explicit instructions for use than those from smoking cigarettes.The most serious (sometimes delivered by well-known professional ath- is oral cancer. Studies have found precancerous letes) and by suggesting that the product could be patches of tissue in the mouths of about half of current used without adult detection." teen-age users;with continued tobacco use,about one in By 1991, use of snuff had risen more than 10 times 20 such lesions will become cancerous within five among young men—to 7.5 percent—while actually years.Smokeless tobacco is also terrible for the gums: declining by about one-third among male senior citizens. They become inflamed and recede. It's illegal to sell snuff to anyone under 18. But About one-third of teen-age users among teen-aged boys, its use is now : develop that condition. For these nearly as common as that of cigarettes. — Y "`` reasons, in 1986 the Surgeon By their last year in high school, half General's report concluded that have tried it and one in five are current there was"no safe use"of smoke- users.Many users say they first tried it in less tobacco. 146 CONSUMER REPORTS MARCH 1995 Not cigarette ads begin as early as the fourth or fifth That would be grade, when children may first be- c 1 illegal in Belgium, come susceptible to the images in where these ads cigarette ads. Urge them to identify were published. But seductive images. a legal loophole 0 Make your feelings clear. Children who understand the depth permits these ads of their parents'opposition to it are for Camel matches less likely to smoke. r ' and Marlboro 0 Give them a reality check. lighters. Point out—perhaps while walking past office workers smoking in doors and alleys—that, despite the ads, Given the present political reali- machines. We think the machines the vast majority of adults do not ties in Washington, a national ban should be banned, as they already smoke and no longer even tolerate on cigarette advertising is unlikely. are in some communities. the practice in public. • But states and communities can and Taxes. In 1982, Canada boosted 0 Emphasize health. Kids are should take their own steps to re- cigarette taxes to among the highest notoriously unconcerned about get- duce teen-age smoking,through both in the world. Teen-age smokers ting sick. Tell them anyway: Teen- law enforcement and taxation. turned out to be unwilling or unable age smokers have weaker lungs, Restricting sales. In 1992, Con- to pay $4 or more for a pack of cough more, and suffer worse upper- gress enacted the Synar Amend- cigarettes; the youth smoking rate respiratory infections. Young ath- ment, which requires states to put dropped from 40 percent in 1981 to letes don't perform as well if they teeth into their laws against selling just 16 percent 10 years later. "For smoke.And the more years a person cigarettes to minors--or lose much- the teen market, price sensitivity is smokes, the greater is the risk of needed Federal money for drug and extremely high," says David Mair, lung cancer in middle age. alcohol treatment. However, the associate director of the Canadian 0 Emphasize addiction. Nico- amendment itself has gone unen- Council on Smoking and Health, tine is so addictive that some ex- forced for nearly three years, await- "For every 10 percent increase in perts compare it to heroin.And,once ing clearance of the regulations price,there's a 13 percent reduction hooked, kids find it just as hard to by the Office of Management and in consumption." kick the habit as adults do.Trouble Budget. In the meantime, any real Unfortunately,Canada also learned is, there's no way to predict which progress on enforcing sales laws that when you lower the price,teen kids will become addicted. So it's has occurred at the local level. smoking goes up. Under political best not even to experiment. In communities where police pressure from the tobacco industry, El Help them say no.This tech- rigorously enforce the law, sales Canadian lawmakers rolled back the nique is used in many formal sub- and teen smoking drop. In Solano tax in February 1994 and prices stance-abuse prevention courses in County, California, intensive en- dropped again. By the end of last school but can easily be adapted at forcement of sales laws reduced the year,teen-age smoking rates had al- home.As best you can,play the part percentage of stores selling to ready climbed back up to 19 percent of an admired friend or acquain- minors from 72 percent to 21 percent. NOW to help a teen tance trying to get your teen-ager to One obvious problem is the extra try a cigarette.Help your child work work this can impose on police.To Adolescents start smoking for out ways to turn down the offer. get around that, some communities many intermingled reasons. Some 0 Don't smoke. If you are a have licensed tobacco retailers in factors that seem to fortify children smoker and are unable or unwilling the same way liquor stores are against tobacco experimentation— to quit,at least explain to your chil- licensed.7lie licensing fees pay for self-esteem, academic achievement, dren that you are in the grip of a fear- enforcement.License laws provide a skills for dealing with peer influ- some addiction—and hide your strong incentive for retailers to card ence, and a close parent-child rela- cigarettes. Smoke less in front of young-looking customers; if they're tionship—are built up slowly from your children and make their rooms not careful, they could lose their toddlerhood on; they can't be pro- smoke-free zones. license and thus lose the lucrative vided overnight when a child reaches 0 Impose consequences. If, in adult trade as well. the high-risk age of 12 or 13.Still,if spite of your efforts, you find your In the year and a half after Wood- you're looking for ideas, here are child experimenting with cigarettes, ridge, Ill., began licensing cigarette several steps you can try to help do not treat it as a minor"lids-will-be- retailers, surveys found the propor- your children shun cigarettes: kids"infraction.Treat it as what it is: bon of seventh- and eighth-graders 0 Talk. Children whose parents an act that puts your child at very who reported experimenting with don't talk to them regularly are at high risk of developing a life-threat- cigarettes fell from 46 percent to 23 greater risk for experimenting with ening addiction. Impose whatever percent,and the proportion of daily cigarettes. Make a point of dis- sanctions your family uses for a major smokers fell from 16 percent to 5 cussing your children's lives and misdeed—and don't back down. ■ percent.Over that same period,corn- feelings. Make sure you know their pliance checks found that the pro- friends .(and the friends' parents). Reprints of this report, and January's portion of stores willing to sell to mi- That will help you find out whether report on secondhand smoke, are nors fell from 70 percent to 3 percent any of the friends is trying out smok- available. For pricing information, But restricting store sales will be ing, so you can talk about it with write: CU/Reprints, 101 Truman nearly meaningless if teens are still your own child. Ave., Yonkers, N.Y. 10703-1057. Or able to buy cigarettes from vending 0 Help them decode ads.Ideally, call:914 378-2448. CONSUMER REPORTS MARCH 1995 147 From the Centers for Disease Control and Prevention Leads From the Morbidity and Mortality Weekly Report Atlanta, Ga Attitudes Toward Smoking Policies in Eight States United States, 1993 MMWR. 1994;43:786-789 ring smoking in fast-food restaurants cigarettes$1 per pack;however,many 2 tables omitted (range:42.5%-63.0%)and at indoor sport- (47.9%to 66.1%)believed that such an LEGISLATION regulating smoking has ing events (55.4%-66.9%) than in sit- increase would be unfair to cigarette i at least two functions: to protect non- down restaurants(39.5%-50.6%)and in- smokers. Belief in the effectiveness of ! smokers from the adverse health effects door malls(33.4%-56.5%).Overall,smok- teenage access restrictions was high of environmental tobacco smoke and to ers were less likely than nonsmokers to among both smokers (41.8% to 79.3%) prevent young persons from smoking) support banning smoking in the differ- and nonsmokers (60.2%to 88.4%). To characterize public attitudes toward ent locations. Reported by the folio-Adng BRFSS coordinators: D such legislation,the National Cancer In- Hargrove-Roberson, MSW, Louisiana; J Jackson- stitute(NCI)and the American Cancer Preventing Teenagers from Smoking Thompson, PhD, Missouri; G rnm-ta r, MS, Nam Society used the Behavioral Risk Fac- JerSCv;E Capwe1rP M-, thio;N Hann,MPH,.Oklay Respondents were given a list of five Noma; M Lane, MPH, South Carolina; R Diam-oTd' for Surveillance System(BRFSS)to sur- strategies that might prevent teenag- 'IGI tTexas;K Holm,M as m on.Surveillance vProgram, National Cancer' nstnu e, ational Insti- ey persons in eight states* July- ers from smoking and asked whether tutes of Health.Div of Chronic Disease Control and August 1993 as part of the American they believed the strategies were not at Community Intervention, Office of Surveillance and Stop Smoking Intervention Study for all effective,somewhat effective,or verAnalysis,and Office on Smoking and Health,National Cancer Prevention' This report sum- Center for Chronic Disease Prevention and Health p effective.Each of the strategies was be- Promotion,cnc. marizes the survey findings. lieved to be effective(i.e.,somewhat or BRFSS provides state-specific esti- very)by most respondents:in particu- CDC Editorial Note: The findings in mates of the prevalence of selected risk lar,65.3%-77.8%ofrespondents believed this report are consistent with previous behaviors to be used for planning,imple- that banning all smoking inside and out- studies that have documented public sup- menting, and evaluating public health side school property would be an effec- port for regulating tobacco use in public programs.Each month,state health de- tive strategy.Most respondents(79.1%- places' For example, in 1987, 72% of partments use survey sampling and ran- 89.6%)favored a ban on smoking inside adults in seven Minnesota communities dom-digit-dialing techniques'to conduct school buildings that applies to students, favored prohibiting smoking in public telephone interviews with adults aged visitors, and teachers; 66.2%-85.1% of buildings.'In 1989,findings from a sur- 18 years.During July-August 1993,a respondents favored a ban on the use of vey conducted for the NCI Community total of 20 questions were added to any tobacco product (including ciga- Intervention Trial for Smoking Cessa- BRFSS in the eight states to assess sup- rettes, cigars, pipes, and chewing to- tion (COMMIT)'indicated that among port for policies related to cigarette bacco)at school-sponsored events(e.g., persons in 10 communities, 62%-100% smoking.' To estimate the state popu- football games and field trips). supported restricting or banning smok- lation prevalences,'data were weighted Banning all ci arette advertisingwas ing in selected locations. Most favored to the age-,race-,and sex-specific popu- const ere o e an effective strate restricting smoking in five locations lation counts from the most current cen- smo ng among Teenagers by_ (bars, restaurants, bowling alleys, pri- sus(or intereensal estimate)in each state ��0 c o respon en s. n a ition, vate worksites,and government build- and for the respondents probability of 49.8%-66.5% of respondents believed ings) and banning it in three other lo- selection. SUDAANs was used to cal- that tobacco advertising influences per- cations(indoor sports arenas,hospitals, culate the 95%confidence intervals for sons to buy tobacco products.The pro- and doctors' offices). the prevalence estimates.For this study, portion of respondents who supported a These findings also confirm increas- sample sizes ranged from 252 to 431 per ban on advertising tobacco products at ing support for banning smoking in res- state; state-specific response rates for sports stadiums and arenas ranged from taurants s For example,16.2%to 32.3% completed interviews ranged from 63.6% 67.7%to 78.2% and the r ort' n who of respondents in the COMMIT study' to 93.3%.Current smokers were defined supported a ban on advertising tobacco favored banning smoking in restaurants, as persons who had smoked at least 100 prq(aiuc s on ; t,anrds rnnaed from compared with 39.5%to 63.0%of BRFSS cigarettes and who reported being a 62.6%to 77.2%. respondents. In addition, the BRFSS smoker at the time of the interview. i�rtions of respondents be- findings distinguish between fast-food Environmental Tobacco Smoke lieved in the effectiveness of selected and sit-down restaurants. Support for measures to limit teenagers access to banning smoking in fast-food restaurants Respondents were given a list of pub- tobacco products,including stronger en- was stronger than support for banning lic locations and asked whether,for each forcement of laws prohibiting the sale of smoking in sit-down restaurants,possi- setting,smoking should be allowed in all cigarettes to minors (77.1% to 85.5%), bly because of the perception that fast- areas(do not restrict),allowed in some banning all cigarette vending machines food restaurants tend to cater to and be areas (restrict), or not allowed at all (69.3%to 79.3%and increasing the price frequented by children and adolescents' (ban). Public opinion about whether to of a pack of cigarettes(55.4%to 67.7%). Previous studies2 have documented restrict or ban smoking varied across Most respondents (54.1% to 68.8%) fa- high levels of support for measures to settings: support was greater for ban- vored increasing the tax on a pack of prevent teenagers from smoking.7•10 The JAMA,February 15, 1995-Vol 273, No. 7 From the CDC 531 BRFSS findings indicate widespread be- of Medicine,2 which include the need to tions involved with youths; (3)restrict lief in the effectiveness of such mea- (1)adopt and enforce tobacco-free poli the advertising and promotion oftobacco sures and suggest broad support for ban- cies in all public locations, especially products;and(4)increase the excise tax ning the use of any tobacco product at those that cater to and are frequented on cigarettes. school-sponsored events. Finally, the by children and youths;(2)adopt tobacco- References 10 available. BRFSS findings indicate support for rec- free policies that apply to persons at- *Louisiana, Missouri, New Jersey, Ohio, Oklahoma, ommendations issued by the Institute tending events sponsored by organiza- south Carolina,Texas,and Washington. Jimson Weed Poisoning Texas, New York, and California, 1994 MMWR. 199544:41-44 the patient was hallucinating and had reported that fires in the Los Angeles 1 table omitted fully dilated pupils,dry mouth,and de- area may have promoted regrowth of INGESTION of Jimson weed (Datura creased bowel sounds.He became pro- Jimson weed in defoliated areas. stramonium),which contains the anti- gressively agitated and was sedated with Reported by:DM Perrotta,PhD,Bur of Epidemiology, cholinergics atropine and scopolamine, intravenous diazepam and alprazolam. Environmental Epidemiologist,Texas Dept of Health; can cause serious illness or death. Spo- Hallucinations continued for 36 hours. LN Nickey,MD,El Paso City-County Health and En- radicincidents ofintentionalmisuse have On October 11 he was dischar ed for Pharmental District, El Paso. M Raid, T HospitCaraccal, e g PharmD,HC Mofenson,MD,Winthrop Univ Hospital, been reported throughout the United psychiatric counseling.He had a history Long Island Regional Poison Control Center, New States,and clusters of poisonings have of chronic substance abuse. York;C Waters,Injury Control Program,D Morse, MD,State Epidemiologist,New York State Dept of occurred among adolescents unaware of During October 8-November 15,a re- Health.AM Osorio,MD,S Hoshiko,MPH,Div of En- its potential adverse effects. This re- gional poison-control center was contacted vironmental and Occupational Disease Control, GW Rutherford,III,MD,State Epidemiologist,California port describes incidents of Jimson weed about this case and for information about State Dept of Health Svcs.Div of Environmental Haz- poisoning that occurred in Texas, New 13 other identified cases of Jimson weed ards and Health Effects,National Center for Environ- York, and California during June-No- intoxication.The mean age of the 14 pa- mental Health,CDC. vember 1994. tients was 16.8 years(range:14-21 years), and eight were male.In the five incidents CDC Editorial Note: D. stramonium Texas for which quantity of Jimson weed expo- grows throughout the United States and, On June 19, 1994, the El Paso City- sure was reported,ingestion ranged from historically,was used by American Indi- County Health and Environmental Dis- 30 to 50 seeds per person. ans for medicinal and religious purposes. trict was notified of two male adolescents Manifestations included visual hallu- All parts of the Jimson weed plant are (aged 16 and 17 years)who had died from cinations (12 persons), mydriasis (10), poisonous,containing the alkaloids atro- D.stranwnium intoxication.On June 18, tachycardia(six),dry mouth(five),agi- pine,hyoscyamine,and scopolamine.Jim- the decedents and two other male ado- tation(four),nausea and vomiting(four), son weed—also known as thorn apple, lescents had consumed tea brewed from incoherence (three), disorientation angel's trumpet, and Jamestown weed a mixture of roots from a Jimson weed (three), auditory hallucinations (two), (because the first record of physical symp- plant and alcoholic beverages, then fell combativeness (two), decreased bowel toms following ingestion occurred in asleep on the ground in the desert.Fam- sounds(two),slurred speech(two),uri- Jamestown,Virginia,in 16761�—is amem- ily and police found the decedents the nary retention(one),and hypertension ber of the nightshade family.The toxicity following afternoon.The other two ado- (one). Four patients were treated and of Jimson weed varies by year,between lescents reported .drinking only small released from E Ds, six were hospital- plants,and among different leaves on the amounts of the tea:one experienced hal- ized, three were admitted to an inten- same plant.Although all parts of the plant lucinations; the other had no signs or sive-care unit (ICU), and one refused are toxic,the highest concentrations of symptoms.Neither was treated,nor were medical care.Five of these patients were -anticholinergic occur in the seeds(equiva- biologic specimens collected. Screening treated with activated charcoal,one was lent to 0.1 mg of atropine per seed).The of a toxicologic postmortem blood sample administered gastric lavage, and none estimated lethal doses of atropine and from one decedent detected atropine(55 received physostigmine. scopolamine in adults are >_10 mg and ng/mL)and a blood alcohol concentration >24 mg,respectively.12 (BAC)of 0.03 g/dL(in Texas,intoxication California Symptoms of Jimson weed toxicity is defined as a BAC-0.1 g/dL).Analysis On October 22, 1994, two male and usually occur within 30-60 minutes after of the tea identified atropine,ethanol,and four female adolescents(aged 15-17 years) ingestion and may continue for 24-48 scopolamine. with a history of drinking Jimson weed hours because the alkaloids delay gas- tea were transported to an ED.Two per- trointestinal motility.Ingestion of Jim- New York sons were discharged from the ED;four son weed manifests as classic atropine On the morning of October 9,1994,an were admitted to the ICU because of poisoning.Initial manifestations include 18-year-old man from Long Island was symptoms that included headache, fa- dry mucous membranes,thirst,difficulty brought to an emergency department tigue, disorientation, fixed or sluggish swallowing and speaking,blurred vision, (ED)by his mother after she found him dilated pupils, tachycardia (heart rates and photophobia, and may be followed in his bedroom unclothed and halluci- >120 beats per minute), and hallucina- by hyperthermia, confusion, agitation, nating. Reports from friends indicated tions. These four patients were moni- combative behavior,hallucinations typi- he had ingested 50 Jimson weed seeds tored with electrocardiograms, treated cally involving insects, urinary reten- and had used controlled substances(i.e., with physostigmine and activated char- tion,seizures,and coma.3 Treatment con- cocaine,"ecstasy,"and marijuana) at a coal,and discharged on October 23.The sists of supportive care, gastrointesti- party the previous night.On evaluation, Los Angeles County Forestry Division nal decontamination(i.e.,emesis and/or 532 JAMA,February 15, 1995—Vol 273,No.7 From the CDC original Contributions Trends in Cigarette Smoking in the United States The Changing Influence of Gender and Race Michael C.Fiore,MD,MPH;Thomas E.Novotny,MD;John P.Pierce,PhD;Evridiki J.Hatziandreu,MD,DrPH; Kantilal M.Patel,PhD;Ronald M.Davis,MD Trends in the prevalence, initiation, and cessation of cigarette smoking are tired predominantly by men.`Smoking reported for the US population using weighted and age-standardized data from prevalence among adult men increased seven National Health Interview Surveys(1974 to 1985).The decline in preva- rapidly, reaching more than 50% by lence was linear, with the prevalence for men decreasing at 0.91 percentage 1955.'Smoking rates among men were points per year to 33.5%in 1985 and the prevalence for women decreasing at already declining by the time the Sur- 0.33 percentage points per year to 27.6%in 1985.For whites the rate of decline geon General's first report on smoking and health was released in 1964. (percentage points per year) was 0.57, to 29.4% in 1985, and for blacks the In contrast,smoking rates for women decline was 0.67, to 35.6% in 1985. Smoking cessation increased among all remained at low levels during the early gender-race groups from 1974 to 1985, with the yearly rate of increase (in part of this century and began rising percentage points per year)about equivalent for blacks(0.75)and whites(0.77), with the gradual disappearance of the while it was higher in women (0.90) than in men (0.67). Smoking initiation social restrictions that previously limit- decreased among young men (—1.03), while it remained about the same in ed smoking by women.In 1964,smoking young women (+0.11). Initiation decreased at a more rapid rate in blacks rates for women had not yet reached (—1.02) than in whites (—0.35). We conclude that smoking prevalence is their highest levels."'Warner and Murt' decreasing across all race-gender groups,although at a slower rate for women have argued that an antismoking cam- than men, and that differences in initiation, more than cessation, are primarily paign may have a different impact on smoking prevalence among men and responsible for the converging of smoking prevalence rates among men and women because of these historic differ- women. ences in Smoking rates between the (JAMA 1989;261:49-55) sexes. Although more is known about changes in smoking behavior among whites than among blacks, smoking prevalence has been reported to be CIGARETTE smoking has been identi- reducing the prevalence of smoking.Al- higher among blacks than among whites fied as a major public health problem though these efforts have been some- foreach National Health Interview Sur- and the single most preventable cause of what successful, as indicated by the vey(NHIS)year since 1965.$In a recent premature death in the United States.' steady decline in overall smoking preva- report using data from the 1985 NHIS," During the last 25 years,a considerable fence,' the rate of decline has been blacks and whites were shown to differ public health effort has been directed at variable among different sociodemo- in two aspects of smoking behavior: graphically defined subgroups in our blacks appeared to quit at a lower rate population. In particular,differences in than whites,and black smokers smoked From the Office on Smoking and Health,Center for gender,race,and educational level have fewer cigarettes per day than white chronic Disease Prevention and Health Promotion, been noted." smokers.After controlling for socioeco- Centers for Disease Control,Atlanta.or Fiore is now with the Department of medicine, Center for Health Smoking patterns in the United nomic status, the investigators did not Sciences,University of Wisconsin,Madison. States have differed between men and find a statistically significant difference Reprint requests to Chief,Epidemiology Branch,Of- fice on Smoking and Health,5600 Fishers Lane,Rock- women.During the first half of the 20th in the prevalence of ever smoking be- ville.MD 20657(Dr Pierce). century, smoking was a behavior prac- tween these races. JAMA,Jan 6, 1989—Vol 261,No. 1 Trends in Cigarette Smoking—Fiore et al 49 Table 1.—Sample Sizes of National Health Interview Survey Tobacco Use Supplement by Race and Sex No.(%)of Subjects' Survey Year White Men Black Men White Women Black Women Total 1974 8929 (40.5) 792 (3.6) 11151 (50.6) 1180 (5.4) 22052 (100.1)t 1976 8551 (40.8) 832 (4.0) 10421 (49.7) 1174 (5.6) 20978 (100.1)t 1978 4366 (41.3) 360 (3.4) 5267 (49.8) 578 (5.5) 10571 (100) 1979 8911 (40.8) 868 (4.0) 10800 (49.5) 1253 (5.7) 21832 (100) 1980 3894 (40.8) 374 (3.9) 4690 (49.1) 595 (6.2) 9553 (100) 1983 8182 (39.0) 751 (3.6) 10768 (51.4) 1262 (6.0) 20963 (100) 1985 11423 (36.8) 1654 (5.3) 15133 (48.7) 2872 (9.2) 31082 (100) *Data are for subjects 20 years of age and older. tPercentages do not add to 100%because of rounding. In this report we analyze trends in changed for the 1985 NHIS.B Changes and between blacks and whites. For smoking prevalence in the United included oversampling for blacks so that analysis by race, Native Americans, States as a function of gender and race. estimates for this group would be more Hispanics, Alaskan Natives, Asians, Because two major factors,smoking ini- reliable, reducing the number of PSUs and Pacific Islanders were excluded be- tiation and smoking cessation, deter- from 376 to 201,and selecting two PSUs cause of insufficient sample sizes for mine smoking prevalence, trends in per stratum instead of one. trend analyses. these measurements are also examined The supplemental survey on tobacco to better understand the observed use included self-respondents only. Statistical Analysis changes in smoking prevalence. The When a selected respondent could not All estimates reported were weight- analysis uses data from the NHIS con- be interviewed in person, the survey ed to reflect the US population and to ducted by the National Center for was administered by telephone. The adjust for sampling designs. Further- Health Statistics (NCHS). The large sample was then stratified by age, sex, more, they were age-standardized to sample size, high response rates, and and racial distribution of the US popula- the 1985 age distribution for appropri- consistent methods and design of these tion for the survey year and weighted to ate comparisons between the survey national, population-based, cross-sec- reflect the individual probability of se- years. Standard errors for prevalence tional surveys make them a unique re- lection as well as other survey design and quit ratio estimates were adjusted source for describing smoking behavior features!Because the youngest age in- for design effect using variance curves in the United States. cluded in the NHIS varied across sur- provided by the NCHS.'Current smok- vey years, we restricted our study to ing prevalence,quit ratios,and smoking METHODS respondents aged 20 years and older. prevalence among 20- to 24-year-olds were calculated for each .NHIS year Sample Design questionnaire from 1974 through 1985 and then strati- The NCHS, through the NHIS, has Ever smokers were defined as per- fied by sex and race. A line of best fit collected health information since 1964 sons who reported that they had using least-squares regression was con- from a probability sample of the civilian, smoked at least 100 cigarettes in their structed for each of these measure- noninstitutionalized population of the lifetime. They were further identified ments.'Ninety-five percent confidence United States. In cross-sectional sur- as current smokers or former smokers intervals for values predicted by the line veys in 1974, 1976, 1978, 1979, 1980, based on responses to the question"Do of best fit were calculated.The Rz statis- 1983,and 1985,the NCHS administered you smoke now?" The rate of quitting tic was calculated for each regression. a supplemental questionnaire on tobac- was measured by the quit ratio,which is These coefficients assess the linearity co use to a subsample of the population the proportion of ever smokers who are and fit of the observed data about the surveyed. Only NHIS data since 1974 former smokers.'°This rate includes all regression lines.' When the slopes of were analyzed because of the similarity ever-smokers who were not smoking at the regression lines were close to 0,the t of the survey design and the exclusion of the time of the survey; ie, both long- test rather than the R' statistic was proxy data since that time. Detailed term and short-term quitters. Accord- used to assess linearity. The signifi- descriptions of the NHIS design and ingly, it is not an index of success in Cance of observed differences in the procedures have been reported pre- maintaining abstinence. Like preva- rates of change (slopes) between the viously." lence measures, quit ratios are point races and sexes was tested using the For these surveys,the United States estimates based on cross-sectional,pop- gaussian z statistic.' was divided into 1924 geographically ulation-based surveys. Smoking initia- defined areas called primary sampling tion is more difficult to measure using RESULTS units(PSUs). The PSUs were grouped cross-sectional studies of adults such as Sample Size and Response Rates into 376 strata (combinations of PSUs the NHIS.However,smoking initiation with similar characteristics), and one is known to occur mainly during the teen Sample sizes for the NHIS supple- PSU was then randomly selected from years;most people who become regular mental questionnaire on tobacco use are each stratum.Finally,a small cluster of smokers start by age 20 years."There- listed in Table 1. For each survey year, households was randomly selected for fore, the prevalence of smoking among sample sizes for the total population interview from each of the selected the youngest age group in the NHIS (aged 20 years and older) by race and PSUs. Members of these households (20-to 24-year-olds)was used as an indi- sex are provided. For both races, the were contacted by mail and then visited cator of smoking initiation. All results proportion of women is consistently by a representative of the US Bureau of were stratified by sex and race to assess greater than the proportion of men in the Census. The sample design was differences between men and women each year. However, the magnitude of 50 JAMA,Jan 6,1989—Vol 261,No.1 Trends in Cigarette Smoking—Fiore et al Table 2.-Trends in Adult Cigarette Smoking in United States by Race and Sex Based on Linear Regression Analysis' Unear Model Estimatea,%t Measurement 1974 1985 Slope,Mean t SEM P* W Prevalence Adults(overall) 36.7 30.4 -0.58 t 0.05 .001 0.97 Men 43.4 33.5 -0.91±0.06 .001 0.98 Women 31.2 27.6 -0.33=0.06 .002 0.88 Whites 36.1 29.4 -0.57±0.05 <.001 0.97 Blacks 44.0 35.4 -0.67±O.15 .007 0.80 White men 42.3 32.1 -0.87±0.08 <001 0.96 Black men 54.6 40.6 -1.15±0.13 <.001 0.95 White women 30.9 26.9 -0.32±0.05 .001 0.91 Black women 36.3 31.5 -0.26±0.20 NS § Ault ration Men 38.5 45.8 0.67±0.10 .001 0.90 Women 29.9 39.8 0.90±0.13 .001 0.91 Whites 38.0 47.1 0.77±0.12 <.001 0.89 Blacks 21.8 32.9 0.75±0.14 <.001 0.85 White men 41.0 49.0 0.67±0.12 .002 0.87 Black men 23.1 35.9 1.04±0.10 <.001 0.96 White women 32.1 43.3 0.95±0.15 .002 0.88 Black women 20.2 29.7 0.46±0.23 NS § Initiationll Men 44.8 33.4 -1.03±0.27 .013 0.75 Women 33.3 34.6 0.11±0.27 NS § Whites 47.1 32.3 -0.35±0.19 .044 0.40 Blacks 38.6 27.9 -1.02±0.48 .096 0.48 'White men 42.6 31.6 -0.93±0.25 .014 0.74 Black men 61.3 27.5 -2.24±0.55 .010 0.77 White women 35.2 33.1 0.16±0.25 NS § Black women 33.8 32.6 -0.10±0.46 NS § -Data are weighted and age-standardized for subjects aged 20 years and older and are from National Health Interview Surveys conducted in 1974,1976,1978 through 1980,1983,and 1985. tObserved values from National Health Interview Surveys are shown in Figs t through 5. *Values were calculated using t statistic with the null hypothesis that the slope is not significantly different from 0;NS Indicates not significant(P>A). §Slope was riot significantly different from 0. IlProportion of ever smokers who are former smokers. gPrevalence among 20-to 24-year-olds. this bias is constant over the study peri- The smoking prevalence was higher prevalence than women.The rate of de- od. The mean response rate for the to- for blacks than whites in each survey cline for white men(-0.87 percentage bacco use supplement to the NHIS was year from 1974 to 1985. The estimated points per year)was significantly great- 89.8%, with a range of 88.5%to 91.1% prevalence of smoking among blacks de- er than the rate of decline for white (NCHS,unpublished data,1988). creased from 44.0%in 1974 to 35.4%in women (-0.32 percentage points per Smoking Prevalence 1985 at a rate of -0.67 percentage year, P<.05) (Table 2). The rate of de- points per year (Table 2). Among cline for black men(-1.15 percentage The overall estimated prevalence of whites, smoking prevalence decreased points per year) was also significantly smoking among adults in the United from 36.1%in 1974 to 29.4%in 1985 at a greater than the rate of decline for black States has decreased steadily from rate of -0.57 percentage points per women (-0.26 percentage points per 36.7%in 1974 to 30.4%in 1985,declining year (Table 2). The regression lines fit year,P<.05)(Table 2). at a rate of 0.58 percentage points per the observed data very well for both Smoking Cessation year(Table 2).This decline has occurred blacks(R2=0.80)and whites(RE=0.97) among both men and women, with a (Fig 2).Although the actual prevalence The quit ratio(defined as the propor- steeper rate of decline for men(Fig 1). estimates were higher for blacks than tion of ever smokers who are former The estimated prevalence of smoking whites in each survey year,the rates of smokers) increased in both men and for men decreased from 43.4%in 1974 to decline (slopes) from 1974 to 1985 did women between 1974 and 1985.The es- 33.5%in 1985 at a mean rate of -0.91 not differ between the races(P=.255). timated quit ratio for men increased percentage points per year (Table 2). The rate of decline in current smoking from 38.5%in 1974 to 45.8%in 1985 at a Among women,smoking prevalence de- prevalence was higher for black men mean rate of change of +0.67 percent- creased from 31.296 in 1974 to 27.6%in (-1.15 percentage points per year) age points per year(Table 2). The quit 1985 at a rate of -0.33 percentage than for white men(-0.87 percentage ratio for women increased from 29.9%in points per year (Table 2). The regres- points per year,P=.03),but the decline 1974 to 39.8%in 1985 at a mean rate of sion lines fit the observed data very well for black women (-0.26 percentage change of+0.90 percentage points per for both men (R'=0.98) and women points per year) was not statistically year (Table 2). This higher rate of W=0.88) (Fig 1). From 1974 to 1985, different from that for white women change per year in quit ratios among the rate of decline (slope) for women (-0.32 percentage points per year, women compared with men approached was significantly less than that for men P=.38)(Fig 3). For both races in each but did not reach statistical significance (P=.001). survey year,men had a higher smoking (P=.08). JAMA,Jan 6,1989-Vol 261,No.1 Trends in Cigarette Smoking-Fiore et al 51 The quit ratios between 1974 and 1985 increased from 21.8% to 32.9% for 50 blacks and from 38.0% to 47.1% for whites(Table 2).Although these values were higher for whites than for blacks in each survey year, the quit ratios for both races increased linearly from 1974 40 to 1985. The rates of change (slopes) were significantly greater than 0 and again did not differ between the races o (P=.436)(Table 2). 6 30 - -----�-- -�r------- The rate of change in quit ratios was m -�- significantly higher for black men than for white men(+1.04 vs+0.67 percent- m age points per year,P=.01)but did not n statistically differ between black wom- g' 20 en and white women (+0.46 vs +0.95 E percentage points per year, P=.31) Men women (Table 2 and Fig 4). The quit ratios for Observed values • ■ men were greater than those for women 10- Regression Line among both races and increased for each S2 lope 0.98 0.88 group between 1974 and 1985;the slopes were significantly different from 0 for all groups but black women. Among 0 whites,the rate of change in quit ratios 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 was higher for women than for men;this difference approached but did not reach Year statistical significance (P=.07) (Table 2). However, among blacks, the rate of change for men was significantly great- Fig t.-Weighted and age-standardized smoking prevalence among men and women aged 20 years and er than that for women(P=.01)(Table older from 1974 to 1985 in National Health Interview surveys in United States. 2). Smoking Initiation Between 1974 and 1985,smoking ini- tiation(as measured by the prevalence of current smokers aged 20 to 24 years) decreased markedly in men(from 44.8% to 33.4%) while remaining essentially Fig 2.-Weighted and age-standardized smoking prevalence among blacks and whites aged 20 years and unchanged in women (ffom 33.4% to older from 1974 to 1985 in National Health Interview Surveys in United States. 34.6%). However, smaller sample sizes for this age group resulted in estimates 50 with more variability(Fig 5). For men aged 20 to 24 years, the observed prevalence is well fitted by the weight- ed least-squares regression line (RZ=0.75), with a mean rate of decline 40 of 1.03 percentage points per year(Ta- ble 2). For women ageri'20 to 24 years, i------I the slope of the regression line was not significantly different from 0, making 30- the 0 the Rz computation inappropriate; a m one-sample t statistic showed no statis- 5 tically significant differences between rL IP the seven observed values from 1974 to 1985(P=.97).Thus,from 1974 to 1985, c 20 smoking initiation remained at about E Blacks Whites 34%among young women while declin- Observed values ■ • ing for young men at a rate of about one Regression Line ---- percentage point per year to about 33% 10 Slope -0.67 -0.57 in 1985. R1 0.80 0.97 Regarding racial differences, smok- ing initiation decreased substantially from 1974 to 1985 for both whites and 0 blacks (-0.35 vs -1.02 percentage 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 points per year) (Table 2). These rates Year of change between blacks and whites approached but did not reach statistical 52 JAMA,Jan 6,1989-Vol 261,No.1 Trends in Cigarette Smoking-Fiore et al 1 60 60 5 50 40 -� 6 40 30 30 ----- -- --_--------- ----- C rn c 220 Black White o Men Men Black White tE Observed Values E ■ e Women Women Regression Line — ---- Observed Values ■ • 10 Slope 0.95 0.96 10 Regression Line — ---- Slope -0.26 -0.32 R' NS 0.91 01 0 Nq �0�6 ,�0� ,019 �0N9 ,000 � � 0Ib p ,009 �g1A 19�619�019�N--;A0 19P9 400 119�X902 40919� -'0b9 Year Year Fig 3.—Left, Weighted and age-standardized smoking prevalence among black and white men aged 20 years and older from 1974 to 1985 in National Health Interview Surveys in United States.Right,Weighted and age-standardized smoking prevalence among black and white women aged 20 years and older from 1974 to 1985 in National Health Interview Surveys in United States.NS indicates not significant. Fig 4.—Left.Weighted and age-standardized quit ratios(former smokerstever smokers)among black and white men aged 20 years and older from 1974 to 1985 in National Health Interview Surveys in United States.Right,Weighted and age-standardized quit ratios among black and white women aged 20 years and older from 1974 to 1985 in National Health Interview Surveys in United States.NS indicates not significant. 0.60 0,60- 0.50- ` 0.50 0.40 0.40- cc .40 cc 0.30 cc 0.30 O t7 0.20 Black White 0.20 Black White Men Men Women Women Observed Values ■ • Observed Values ■ • 0.10 Regression Line — ---- 010 Regression Line ---- Slope 1.04 0.67 Slope 0.46 0.95 R, 0.96 0.87 R' NS 0.88 0 0 �9�b �0�0,016,9� 10�� ,019�g00 1°�0,�°�219��`� X900 .�9�6,�9�h,9�6,9.(l n9Ab,900 1001,902,909100 1945 Year Year JAMA,Jan 6,1989—Vol 261,No.1 Trendsin Cigarette Smoking—Fioreet a[ 53 a s0 0 80 v v 0 0 it 60 y 60- 4 N pp .2 N N rT E 40 0 __ 40 Q — — Q — — -- ------- ----- 8 U m_ Black White m Black White 20 Men Men 20 Women Women Observed Values ■ • Observed Values ■ • Regression Line - -- Regression Line rn Slope -2.24 -0.93 c Slope -0.10 0.16 c R1 0.77 0.74 o R° NS NS EE 0 Cn 0 �g1 a,g�6,g16,g1� ,g�g,gg0,gg�,g$Z,gg�,ggA,gad 1gf A,g�5,g�6 1g(t 1g�0,g.(g�ggo �gg�egg2 1ggg Year Year Fig 5.—Lett,Weighted and age-standardized smoking initiation(smoking prevalence among 20-to 24-year olds) among black and white men from 1974 to 1985 in National Health Interview Surveys in United States.Right,Weighted and age-standardized smoking initiation among black and white women from 1974 to 1985 in National Health Interview Surveys in United States.NS indicates not significant. significance (P=.10). The rate of Moreover, these gender differences tained consistent quitting activity(mea- change for black men was significantly were observed for both blacks and sured here using the quit ratio)during higher than that for white men(P=.02) whites.If these trends continue into the the last ten years. Overall, a higher (Table 2 and Fig 5).Among women,the 1990s,smoking rates for men and wom- proportion of men than women have slopes of the regression lines for both en will converge, changing the historic quit smoking; however, the rate of races were very close to 0(Table 2 and observation that men smoke at a higher change per year in quitting is higher for Fig 5);a one-sample t test indicated that rate than women."As a result of this women than for men(although this dif- the slopes for black and white women finding,one can also predict a change in ference was only of borderline statisti- from 1974 to 1985 were not significantly the differential mortality patterns be- cal significance in this study). These different from 0(P<.05). tween the sexes.The large difference in data do not support the hypothesis that mortality from smoking-related dis- men and women differ in their rates of COMMENT eases such as lung cancer between wom- quitting over time.18 These data do not, lb monitor progress in reducing en and men18 may, after a latency peri- however, address the relative success smoking in the United States,valid and od, gradually decrease. In fact, lung of men and women in maintaining smok- reliable estimates of smoking preva- cancer mortality rates have already be- ing abstinence. lence in the population are needed. To gun to decline among men but are still Smoking initiation rates,unlike quit- estimate the population prevalence at increasing among women." ting activity, have differed markedly any given point requires a large, Recently reported preliminary data between men and women during the last representative, cross-sectional survey. from the 1987 NNIS,with a population- ten years. Smoking prevalence among The NHIS,with large sample sizes and based sample size of more than 40 000 20-to 24-year-olds has fallen rapidly in high response rates, provide the most persons, support the linear projections young men.In contrast,this indicator of representative and reliable estimates of based on trends from 1974 to 1985 de- smoking initiation has remained con- health behavior in the United States. scribed in this report. The preliminary stant in women. As a result, young Additionally, self-reported smoking 1987 NHIS estimates for adult(aged 20 women have been smoking at a higher status has been demonstrated to be a years and older)prevalence were 31.7% rate than young men since about 1984. valid measure of smoking behavior us- for men and 26.8%for women(NCHS, These findings are similar to those of the ing biochemical markers in large com- unpublished data, 1988). These values annual survey of drug use among high munity surveys.18-15 are almost identical to the prevalence school seniors sponsored by the Nation- From 1974 to 1985,the prevalence of values of 31.7%for men and 26.9%for al Institute on Drug Abuse,19 which smoking among men decreased at a con- women predicted by our linear model have shown higher rates of smoking sistent rate of slightly less than one per- for 1987(Table 2). prevalence among young women than tentage point per year.During the same Two factors determine the preva- young men in each year since 1977. period, the prevalence of smoking lence of smoking: quitting activity and These data suggest that differences among women also fell, but at a rate smoking initiation. These data indicate in initiation rates rather than in cessa- only one third of that observed for men. that both men and women have main- tion rates are mostly responsible for the 54 JAMA,Jan 6,1989—Vol 261,No.1 Trends in Cigarette Smoking—Fiore at al i converging rates of smoking prevalence also decrease as a result. However, we data, 1988). These values were similar among men and women. The results found that the rate of change in preva- to the prevalence values of 34.1% for suggest that the public health message lence (as measured by the change in blacks and 28.3% for whites predicted to quit smoking has been effective in percentage points per year) for blacks by our linear model for 1987(Table 2). reaching increased numbers of men and and whites was essentially equal be- . National antismoking programs and women who smoke. However, the mes- tween 1974 and 1985. The parallel de- policies directed toward high-risk popu- sage to avoid starting smoking has been clines in smoking prevalence result lations have begun to address the dis- less successful in reaching young wom- from both the increase in the quit ratio parity between smoking prevalence en than in reaching young men. Th and the decrease in the rate of initiation; among blacks and whites.' Programs achieve the public health goal of a neither of these rates of change differ that consider the special cultural needs smoke-free society,young women must significantly by race. of blacks may be more successful in in- become a major focus of prevention ac- Although blacks had a higher current fluencing smoking behavior than those tivities, ideally during childhood and smoking prevalence for each year of the that do nota The need for these pro- adolescence. study than did whites, this disparity grams becomes even more apparent Blacks in the United States smoked at may diminish or eventually disappear. when one considers the aggressive ciga- higher'rates than whites in every sur- Our analysis found that among black rette advertising and marketing cam- vey year between 1974 and 1985. This men,smoking prevalence and initiation paigns that are targeting minorities." difference in absolute smoking preva- decreased at a faster rate and the quit The data presented in this report can lence may reflect a differential impact of ratio increased at a faster rate than assist the public health community in antismoking information on interven- among white men. identifying target populations, such as tions in the more distant past. Thus, As with gender,our race-specific lin- blacks and young women, for concen- there is a need to proceed with interven- ear projections for smoking prevalence trated smoking intervention efforts. tions among blacks so that the absolute in 1987 are approximately equal to the Additionally,this information can assist differences in current smoking can be actual 1987 NHIS estimates. The pre- in establishing challenging yet realistic eliminated. Racial differences in smok- liminary 1987 NHIS adult prevalence goals for smoking prevention among ing-related illnesses,such as cardiovas- estimates were 34.0% for blacks and subpopulations in the United States. cular disease' and lung cancer,21 may 28.8% for whites (NCHS, unpublished References 1. The Health Consequences of Smoking:Nicotine Dept of Health and Human Services publication cessation.Women Health 1986;11:237-251. Addiction:A Report of the Surgeon General, US (PHS)5834-A2.Hyattsville,Md,National Center 19. Johnston LD,O'Malley PM,Bachman JG:Na- Dept of Health and Human Services publication for Health Statistics,1958. tional Trends in Drug Use and Related Fhctors (PHS)88-8406. Rockville,Md,Office on Smoking 10. Pierce JP,Aldrich RN,Hanratty S,et al:Up- Among American High School Students and and Health,1988, take and quitting smoking trends in Australia 1974- Young Adults,1975-1986,US Dept of Health and 2. Cigarette smoking in the United States,1986. 1984.Prev Med 1987;16:252-260. Human Services publication(ADM)87-1535.Rock- MMWR 1987;36:581-585. 11. McGinnis JM,Shopland D,Brown C:Tobacco ville,Md,National Institute on Drug Abuse,1987. 3. Cigarette smoking among blacks and other mi- and health trends in smoking and smokeless tobac- 20. Kraus JF, Borhani NO,Franti CE:Socioeco- nority populations.MMWR 1987;36:404407. co consumption in the United States. Annu Rev nomic status,ethnicity,and risk of coronary heart 4. Novotny TE,Warner KE, Kendrick JS, et al: Public Health 1987;8:441-467. disease.AmJEpidemiol1980;111:407-414. Smoking by blacks and whites:Socioeconomic and 12. Kleinbaum DG,Kupper LL:Applied Regres- 21. Devesa SS, Diamond EL:Socioeconomic and demographic differences. Am J Public Health sion Analysis and Other Multivariate Metho& racial differences in lung cancer incidence.Am J 1988;78:1187-1189. Boston,Duxbury Press,1978. Epidemiol 1983;118:818-831. 5. The Health Consequences of Smoking for Wom- 13. Fortmann SP, Rogers T, Vranizan K, et al: 22. Smoking,7bbacco,and Cancer Program 1985 en:A Report of the Surgeon General,US Dept of Indirect measures of cigarette use:Expired-air car- Report, US Dept of Health and Human Services Health and Human Services.Rockville,Md,Office bon monoxide versus plasma thiocyanate. Prev publication(PHS)86-2687.National Cancer Insti- on Smoking and Health,1980. Med 1984;13:127-135. tute,Bethesda,Md,1986. 6. Hammond EC,Garfinkel L:Smoking habits of 14. Petitti DB,Friedman GD,Kahn W:Accuracy 23. Carillo JE:A rationale for effective smoking men and women.JNCI 1961;27:419-442. of information on smoking habits provided on self- prevention and cessation interventions in minority 7. Warner KE,Murt HA:Impact of the antismok- administered research questionnaires.Am J Pub- communities, in Proceedings, National Advisory ing campaign on smoking prevalence: A cohort tic Health 1981;71:308-311. Meeting ofthe Interagency Committee on Smoking analysis.J Public Health Survey 1982;3:374-390. 15. Pierce JP,Aldrich RN,Hanratty S,et al:Up- and Health,US Dept of Health and Human Ser- 8. The National Health Interview Survey Design, take and quitting smoking trends in Australia, vices publication(HHS/PHS/CDC)87-8403,1987. 1973-84, and Procedures, 1975-88, US Dept of 1974-1984.Prev Med 1987;16:252-260. 24. Davis RM:Current trends in cigarette adver- Health and Human Services publication(PHS)85- 16. Cancer Fhcts and Figures:1988. New York, tising and marketing.N Engl J Med 1987;316:725- 1320.Hyattsville,Md,National Center for Health The American Cancer Society,1988. 732. Statistics,1985. 17. Decrease in lung cancer incidence among 25. Cummings KM,Giovino G,Mendicino AJ:Cig- 9. The Statistical Design of the Health.Household- males:United States,1973-1983.MMWR 1986;35: arette advertising and racial differences in ciga- Interview Survey by Staff of the U.S. National 495-496,501. rette brand preference. Public Health Rep 1987; Health Survey and the Bureau of the Census,US 18. Orlandi MA: Gender differences in smoking 102:698-701. JAMA,Jan 6,1989-Vol 261,No.1 Trends in Cigarette Smoking-Fiore et al 55 Trends in Cigarette Smoking in the United States Educational Differences Are Increasing John R Pierce,PhD;Michael C.Fiore,MD,MPH;Thomas E.Novotny,MD; Evridiki J.Hatziandreu,MD,DrPH;Ronald M.Davis,MD National trends in smoking prevalence by educational category from 1974 by the National Institute on Drug through 1985 show that education has replaced gender as the major sociode- Abuse.'Each year since 1976, a large, mographic predictor of smoking status. Smoking prevalence has declined representative sample of high school se- across all educational groups but the decline has occurred five times faster mors in the United States have com- among the higher educated compared with the less educated. From 1974 to pleted questionnaires that include 1985,smoking prevalence among persons with less than a high school diploma smoking-related questions. Because declined to 342% (0.19 percentage points per year) whereas prevalence for these surveys do not include adoles- cents who left school before their senior persons with four years or more of college education declined to 18.4%(0.91 year,the sample underrepresents those percentage points per year). Smoking cessation activity increased across all who will make up the least-educated educational groups, but the rate of increase among the higher educated was group in the community. Results indi- twice that of lower-educated groups. Initiation of smoking among more-edu- cate that the more education a student cated men decreased rapidly to 15%in 1985 but leveled off by 1987.Until 1985, plans to complete, the less likely that less-educated young females were the only group in which smoking initiation student is to smoke.Further,the differ- was increasing.However,in 1987 a sudden and large decline in initiation among ence in smoking prevalence between less-educated females occurred. The apparent recent changes in initiation students who are planning further edu- cationpatterns by educational level suggest that the converging of smoking preva- and those who are not has in- patterns between the genders may not continue.The large and widening educa- creased Inthisve tide, ' tional a in smoking suggests that health promotion priorities need to be In this article, i examine whether gap 9 99 P P there has been a differential change in reassessed. smoking prevalence across educational (JAMA 1989Z1:56-60) categories in the United States since 1974. 7b elucidate the determinants of any such differences, we examine the IN CROSS-SECTIONAL surveys both gest that public health interventions independent impact of cessation and ini- in the United States and elsewhere, during the last 25 years have been more tiation on smoking prevalence. persons with less education consistently salient for the higher educated. This METHODS have reported higher smoking rates would indicate the need for a reassess- than those with more education." ment of public health approaches to re- As in an accompanying article,'this There has been some debate as to ducing smoking prevalence. study used smoking data collected by whether this association is causal.' If As indicated in a companion article,' the National Center for Health Statis- education does have an independent ef- changes in smoking prevalence over tics through the National Health Inter- fect on smoking, increasing the educa- time are a function of the proportion of view Surveys (NHIS) for the years tion level in the community would be the smoking population who have quit 1974, 1976, 1978, 1979, 1980, 1983, and one effective way of reducing smoking and the proportion of the adolescent 1985. Provisional estimates from the prevalence. Indeed, some of the re- population who have started smoking 1987 NHIS are reported but not in- ported decline in smoking in the United regularly during preceding years. The eluded in analyses. A detailed descrip- States may simply be a result of an in- well-known difference in smoking prev- tion of the NHIS design and procedures crease in the mean level of education in alence between the genders has been has been reported previously," and a the community. Examining trends in shown to be uninfluenced by race.' briefer description is contained in an smoking prevalence by educational lev- Moreover, this difference in prevalence accompanying article.' The mean re- el allows the testing of this hypothesis. has been narrowing since 1974,and this sponse rate for the tobacco use supple- Further,such trend analysis will indi- narrowing has been attributed primari- ment to the NHIS was 89.8%, with a cate whether the gap in smoking preva- ly to a higher rate of smoking initiation range of 88.5%to 91.1%(National Cen- lence between the educational levels is among young women, regardless of ter for Health Statistics, unpublished increasing. Such a finding would sug- race.' Of considerable importance is data,1988). whether this higher initiation rate among women is independent of educa- Questionnaire tion.If initiation was higher in sex-spe- As in the accompanying article,'indi- From the Office on Smoking and Health,Center for cific educational categories, this would viduals were divided into ever smokers Chronic Disease Prevention and Health Promotion, provide a clear indication of where pub- (people Centers for Disease Control,Atlanta.or Fiore is now p p (pe p e who had smoked at least 100 with the Department of Medicine, Center for Health lic health action needs to be improved. cigarettes in their lifetimes)and never Sciences,University of Wisconsin,Madison. The impact of education on smoking smokers(people who had never smoked Reprint request t Epidemiology Branch,Office can MD on Smoking and Health,,56D0 Fishers Lane,Rockville, initiationbe inferred from the annu- or who had smoked fewer than 100 ci ga- 20857(Dr Pierce). al study of high school seniors sponsored rettes).Ever smokers were divided into 56 JAMA,Jan 6,1989—Vol 261,No.1 US Smoking Trends—Pierce et al " 1 •_ current smokers and former smokers. Table 1.—Sample Sizes and Percentage Distribution, National Health Interview Survey, Tobacco Use • The quit ratio was defined as the pro- Supplement,United States,1974 to 1985* portion of ever smokers who are former No.(%)of subjects smokers.'° Smoking prevalence in the youngest age group (20 through 24 Survey <High School High School College Graduate Year Education Graduate Some College or More Total years)is used as the proxy variable for 1974 6294 (3e) 7634 (36) 3186 (14) 2747 (12) 22061 (100) indicating smoking initiation. 1976 7177 (34) 7682 (37) 3151 (15) 3022 (14) 21032 (100) The highest level of education com- 1978 3243(31) 4016 (38) 1763 (17) 1575 (15) 10597 (100) pleted was solicited from all respon- 1979 3627 (32) 4638 (39) 1936 (16) 1992 (13) 11995 (100) dents and was reported as the number 1960 2809 (29) 3647 (38) 1640 (17) 1592 (16) 9925 (100) of years of schooling completed.Zb ana- 1963 5704 (27) 6162 (36) 3729 (1e) 3692 (17) 21287 (100) lyze smoking prevalence and quitting 1985 8137 (26) 11733 (37) W77 (19) 5745 (18) 31692 (100) activity,we categorized respondents by educational level:those with less than a *Includes adults aged 20 years and older. high school diploma (fewer than 12 years of schooling completed), high Table 2.—Sample Size and Percentage Distribution in Youngest Age Group (20 to 24 Years), National school graduates (exactly 12 years of Health Interview Survey,Tobacco Use Supplement,United States,1974 to 1985 schooling completed), those with some college(13 to 15 years of schooling com- No.(%>of subjects pleted), and those with bachelor's de- survey High School Graduate At Least Some grees or greater (16 or more years of Year or Less college Totef schooling completed). When we ana- 1974 1678 (61) 1088 (39) 2766 lyzed smoking initiation (using preva- 1976 1624 (61) 1048 (39) 2672 lence for ages 20 through 24 years as a 1978 827 (61) 527 (39) 1354 proxy), we dichotomized educational 1979 1692 (60) 1113 (40) 2605 level into those with at least some col- 1980 818 (64) 465 (36) 1263 lege education (13 or more years) and 1963 1565 (62) 957 (38) 2522 those with a high school education or 1985 1867 (55) 1555 (45) 3422 :less(12 or fewer years). This was done so as not to confuse educational status with age. It was considered that by age 20 years most people who proceed with formal education beyond high school ac- 50 tually have started college. Statistical Analysis All prevalence estimates were weigh- 40 ted to reflect the US population and age standardized to the 1985 age distribu- tion.Standard errors for prevalence es- o 0 timates were adjusted for sampling de- 6 30 EI sign by using variance curves provided a o by the National Center for Health Sta- 9! tistics.For each survey year we plotted a a 111 smoking prevalence stratified by educa- o, tion:We used least-squares regression Y 20 to construct a line of best fit to the obser- E° vations and calculated 95% confidence intervals about values predicted by the High School<High School some College line of best fit. Student's t tests wereGraduate Graduate College Graduate+ used to determine if the slopes were 10 Observed Values • o a ■ significantly different than zero.The R2 Regression Line statistic was calculated for each regres- Slope R2 -0.30 -0.19 -0.78 -0.91 sion to give an estimate of the extent 79 .85 .94 .92 that the linear model fitted the data. 0- 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 RESULTS Year Sample Size and Response Rates Sample sizes for educational subcate- Fig 1.—Weighted and age-standardized observed smoking prevalence and regression lines,by educational gories for each year ranged from ap- category,among those aged 20 years and older, 1974 to 1985,United States.Slope reflects change in proximately 10 000 to approximately ands Nona Centers for ints per year. ioStatistics. Interview Surveys in 1974,1976,1978;1979,1980,1983, 30 000(Table 1). The proportion of the population with less than a high school diploma decreased from a high of 38%in converse trend was seen in the propor- the population in 1974 and 37%in 1985. 1974 to a low of26%in 1985.This change tion of the population who reported hav- Thus, the proportion of the population in the sample reflects the national trend ing had some college education or hav- that is more highly educated has been in recent years for more young people to ing graduated from college. These two increasing at approximately I percent- complete high school." An equivalent groups together accounted for 26% of age point per year. JAMA,Jan 6,1989—Vol 261,No.1 US Smoking Trends—Pierce et al 57 The youngest age group (20- to 24- Table 3.-Trends in Cigarette Smoking With Linear Regression Analysis,National Health Interview Survey, year-olds) did not show a similar in- United States,1974 to 1985 crease in the proportion of highly edu- Linear Model cated persons during the survey period F..etima�,%§ (Table 2). Thus, the trend observed for 1974 AM slope (of slope) Pt FP the total population probably reflects Prevalence the facts that older people are not as Less than high school graduate 36.3 34.2 -0.19 OM .003 0.85 highly educated as middle-aged or High school graduate 37.7 34.4 -0.30 0.07 .008 0.79 younger persons and that older people Some college 36.6 28.0 -0.78 0.09 <.001 0.94 have a much higher death rate. College graduate+ 28.5 18.4 -0.91 0.13 .001 0.92 Cessquit ation [former/ever amoker])Smoking Prevalence Less an h scho 37.2 43.7 +0.59 0.14 .008 0.79 Smoking prevalence declined across High school graduate 34.1 40.1 +0.54 0.09 .002 0.88 all educational levels from 1974 to 1985, Some college 36.4 43.8 +0.67 0.19 .017 0.71 and persons who were more educated College graduate+ 45.2 57.1 +1.08 0.22 .005 0.83 smoked at a lower rate each year than initiation(prevalence for ages 20-24 y) less-educated respondents(Fig 1). The HighTotal of graduate or less 45.0 45.3 +0.03 0.22 .921 ...t rate of decline in prevalence ranged Man 52.4 45.7 -0.60 0.34 .133 0.39 from 0.19 percentage points per year for Women 38.8 44.4 +0.51 0.21 .062 0.53 the least educated to 0.91 percentage Some college or more points per year for the most educated Total 28.4 16.4 -1.10 0.22 .005 0.86 (Table 3). Men 34.0 12.4 -1.97 0.25 .001 0.94 For college graduates,smoking prev- Women 23.5 19.7 -0.34 0.30 .306 ..4 alence decreased markedly from 28.5% in 1974 to 18.4%in 1985.The linear mod- 19'Data weighted and age standardized;adults,20 years and ower;data from 1974, 1976, 1978,1979, 1980, 83,and 1985 included. el(Table 3)predicted a 1987 prevalence tCalculated using Student's t test with null hypothesis;slope not significantly different from zero. of 16.6%.This projection is nearly iden- *Slope not significantly different than zero(P>.1)and therefore no R2 is reported. tical to the recently available 1987 §National Health Interview Study observed values are shown in Figs 1 through 3. NHIS estimate of 16.3%. For persons without a high school diploma, preva so lence has decreased from 36.3%in 1974 to 34.2%in 1985. The linear model pre- dicted a 1987 prevalence of 33.8%;how- ever, the provisional NHIS estimate of prevalence for that year was 35.7%, suggesting that even the modest decline 60 ■ in prevalence observed from 1974 to 1985 may have leveled off. Smoking Cessation ° The quit ratio increased across all ¢ .40- ° education categories during the period -=��'-■ 1974 to 1985,with the most highly edu- ° cated persons consistently quitting at a higher rate than those with less educa- High School<High School Some College tion (Fig 2). Linear regression models .20 Graduate Graduate College Graduate+ fit the data well,with R'values between .71 and .88 (Table 3). The quit ratio Observed Values • o ° ■ increased at a rate of 1.03 percentage Regression Line I� g 0.54 0.59 0.67 1.08 points per year for college graduates vs Slope R' .88 .79 .71 .83 0.54 percentage points per year for high school graduates. ° For college graduates, 45.2% of 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 people who had ever smoked were for- Year mer smokers in 1974 compared with .57.1% in 1985. For those who did not Fig 2.-Weighted and age-standardized quit ratios(former smokerstever smokers),by educational catego- complete high school, 37.2% of ever ry,among those aged 20 years and older,1974 to 1985,United States.Both observed values and regression smokers were former smokers in 1974 lines are shown.Slope reflects change in percentage points per year.From National Health Interview Surveys compared with 43.7%in 1985. in 1974,1976,1978,1979,1980,1983,and 1985,National Centers for Health Statistics. Smoking Initiation sample sizes. Among males who go be- group corroborates that this leveling off yond high school education(Fig 3, bot- has occurred among young men with The observed smoking prevalences tom)prevalence dropped rapidly(2 per- more than a high school education. for respondents aged 20 through 24 centage points per year), from 35% in Among women who attend some col- years can be taken as an indicator of 1974 to 15% in 1985. This 1985 figure lege, prevalence decreased at a more smoking initiation(Fig 3,Table 3). The was well above the linear model esti- modest rate than among men,from 26% regression statistics indicate that there mate of 12.4%,suggesting that the rap- in 1974 to 17%in 1985. This latter ob- is much more volatility in these data, id decline may have leveled off. The served prevalence was well below what some of which may be a result of smaller 1987 provisional estimate of 16.3 far this was expected from the linear trend.The 58 JAMA,Jan 6,1989-Vol 261,No.1 US Smoking Trends-Pierce et al 1 10 The trend among females who did not so proceed to college was the mirror image of those who did attend college; thus, the slope of the combined trend in initia- tion by women would be close to zero. 50 Similarly, combining males and fe- males, the trend in initiation among N those who do not attend college was 40 close to zero(Table 3). The provisional R NHIS estimate for 1987 for women who did not attend college was 37.6%. This a indicates a rapid departure from the 30previous trend for this group and repre- sents the first clear sign of a decline in initiation for young women who did not d 20' attend college. 8, E COMMENT High School Graduate or Less Males Females From 1974 to 1985, smoking preva- 10 lence declined within each educational Regression Line go Observed Values • ■ category of the US population. How- Slope pp Slope R' -0.60 0.51 ever, the rate of decline across educa- .39 .53 tional groups was highly variable. 01 1 1 1Smoking prevalence decreased 4.8 1974 1975 1976 1977 1976 1979 1980 1961 1992 1983 1984 1985 times faster among college graduates Year than among people with less than a high so school education. As a result, by 1985 only 18% of college graduates were smokers whereas 34%of those who did not complete high school were smokers. 50This increasing gap in smoking preva- lence across educational categories is N partly explained by trends in smoking 0 cessation. By 1985, the majority(60%) 40 N of college graduates who had ever smoked had quit whereas only 44% of those with some college education and 30 40%of those with a high school educa- tion or less had quit. Furthermore,the quit ratio among college graduates in- s creased at a rate of 1.08 percentage x 20 _ points per year compared with a rate of Some college+ 0.59 percentage points per year among those who did not complete high school. Males Females Smoking cessation was not confined 10 observed values • ■ to any particular educational category. Regression Line Each group had a large number of peo- Slo 19' -1.97 -0.34 ple who had quit smoking,and the quit 0 -- '' ratio within each educational group has 1974 1975 1976 1977 1978 1979 1980 1981 1992 1983 1984 1985 been increasing in a linear fashion. Year However,the rate of increase in the quit ratio among college graduates is 80% higher than that for the other three edu- Fig 3.—Weighted and age-standardized smoking initiation(as measured by prevalence among those aged cational groups.This disparity suggests 20 to 24 years)for less-educated(12 or fewer years of education completed)(top)and higher-educated(13 or that markedly different influences more years of education completed)(bottom)males and females,1974to 1985,United States.Both observed values and regression lines are shown.Slope reflects change in percentage points per year.Where no 192 is against smoking exist among the differ- reported slope is not significantly different from zero.From National Health Interview Surveys in 1974,1976, ent educational categories. 1978,1979,1980,1983,and 1985,National Centers for Health Statistics. Smoking initiation among young men who attended college decreased during 1987 provisional estimate also was much The 1987 provisional estimate was the study period by more than half to lower than expected at 15.1%.This sug- 43.8%, which was only slightly below 15%in 1985. During most of this time, gests that the rate of decline in initiation that estimated from the linear model this drop seems to have been linear,at a among these young women has in-. (44.5%). Among females in this group, rate of 2 percentage points per year; creased in recent years. smoking prevalence actually increased however,the most recent data available Among those who did not proceed be- during the period,from 40%in 1974 to suggest that this rapid rate of decline yond high school, smoking was much 44%in 1985, thus giving males and fe- has leveled oft This remarkable decline more prevalent. Among males it de- males in this educational category the in smoking initiation did not occur creased from 52%in 1974 to 44%in 1985. same smoking prevalence in that year. among young men who did not attend JAMA,Jan 6,1989—Vol 261,No.1 US Smoking Trends—Pierce et al 69 college.Almost half of this group(46%) ing antismoking programs at this vides timely data that are important for were regular smokers in 1985.The year- group, women in the lower-educated public health planning. However, find- ly change in prevalence over the period category still are 2.4 times more likely ings from that study need to be con- was moderate (-0.60 percentage to take up smoking than those who go on firmed with more representative sam- points per year) compared with the to college. ples, such as the youngest reliable age change for more-educated young men Thus,although smoking in the United group from the NHIS. (-2.0 percentage points per year). By States is becoming a behavior less de- In summary,educational level has be- 1987,provisional NHIS estimates indi- fined by sex roles,' it is increasingly come the major demographic predictor cated that there was almost a threefold becoming an issue defined by education- of whether an individual will smoke cig- difference in prevalence between these al level. However, actually spending arettes.There is considerably more ces- two groups. more time in school does not seem to be sation activity occurring among higher- For 1974 through 1985,the initiation responsible for the observed differ- educated than among less-educated of smoking among young women who ences. Using longitudinal data, Farrell groups, and the gap is widening over attend college declined at a rate of 0.3 and Fuchs'found that the differences in time. Further, regardless of gender, a percentage points per year. Recent smoking prevalence were discernible person is more than twice as likely to data suggest that the decline in initia- before people completed their educa- take up smoking if he or she does not tion among these young women has in- tion,a finding corroborated by the High attend college compared with the per- creased markedly.The 1987 provisional School Seniors Study.'Adolescents who son who does.However,the recent,and estimates indicate that initiation among later go on to higher education smoke apparently strong,.decline in initiation higher-educated females in recent years less than those who do not; thus, the among young women offers consider- has been lower than that among higher- number of years of higher education is able support for the concept that specif- educated males. not causally associated with smoking is targeting of health promotion efforts The prevalence of smoking among prevalence. can be effective in changing behavior in young women with less than a college IFiore et al'identify that the toner- the targeted population. In the future, education actually increased from 1974 gene of smoking prevalence between antismoking messages need to be based to 1985, at a rate of 0.5 percentage males and females can be attributed to a much more on educational status. points per year, to an all-time high of much larger decline in smoking initia- The authors gratefully acknowledge the contri- 44.4%in 1985.That year,smoking prev- tion among young men. In this article, butions of the many professionals at the National alence was about the same for young we report that this decline in initiation Center for Health Statistics for providing the data men and women in this educational has not been consistent across the sexes that served as the source for these analyses. In particular,we thank Jake Feldman,Joel Kleinman, group. The provisional 1987 estimate for the differing educational levels. Be- and Ron Wilson for consulting with us on the com- indicates a remarkable six-percentage tween 1974 and 1985 the increasing up- plex design issues and methods related to analyses point decline in prevalence for this take level among lesser-educated young of these data.The authors also thank Kantilal Patel u of young women. Given the close women nullified the decline seen among for statistical support and Sonia Balakirsky for see- group ec- groP Y $ $ retariai support. proximity of the 1987 observed data higher-educated young women. How- References with the trend estimate in all other tate- ever, these trends in initiation by the gories and the large sample size in 1987, different sex/educational groups seem 1. US Dept of Health, Education, and Welfare: this change in initiation is unlikely to be to be changing rapidly. Between 1985 Smoking and Health: A Report of the Surgeon 9. artifactual. Final confirmation of the and 1987 there was a much larger de- 2. BiaGeneral.Public Health PM:Less s $ 2. Bjartveit K,Lochsen PM:Less smoking among size of the reversal in trend must await cline in initiation for women than for the well-educated. World Smoking Health 1979; the 1988 survey data. men.This was a consequence of the lev- 4(1):27-30. Thus, the finding in the accompany- eling off of the decline among higher- 3. warner K:Smoking and schooling:In search of in article' that there has been no educated young men and a large decline the missing link.J Health Econ ng1982;1:291-298. an:29 health: g Y $ 4. Farrell P,Fuchs VR:Schooling and health:The change since 1974 in the proportion of in initiation among lesser-educated cigarette connection. J Health Econ 1982;1:217- young women who are becoming smok- young women.Given this change in pat- 230. ers needs to beualified.The identified terns of initiation and the fact that these 5. Grossman M: The correlation between health q and schooling, in Tereckyj NE (ed): Household lack of difference masks the divergent patterns have been the major influence Production and Consumption.New York,Colum- trends that have occurred between less- on the differential trends in smoking bia University Press,1975,pp 147-223. educated and more-educated women. prevalences,'it might be expected that 6. Fiore MC,Novotny TE,Pierce JP,et al:Trends Initiation of smoking among young we may not reach the stage where fe- in smoking in the United States:The changing in- fluence of gender and race.JAMA 1989;261:49-55. women with more education decreased male smoking prevalence is ever higher 7. Johnston LD,O'Malley PM,Bachman JG:No- from 1974 to 1985 whereas that among than male prevalence in the United tional Trends in Drug Use and Related Factors less-educated women increased to an States. Among American High School Students and all-time high. When these two groups These findings vary from those re- Instil Adults,1975.1986.Rockville,Md,National are combined, the overall smoking ported by the High School Seniors Institute on Drug Abuse,1987. g, Kovar MG,Poe GS:The National Health Inter- prevalence among young women re- Study.'This may relate to the fact that view Survey Design,1978-1984,and Procedures, mained essentially constant from 1974 approximately 15% of both men and 1975-1988,US Dept of Health and Human Services to 1985. However, the 1987 provisional women aged 20 to 24 years report that Publication (PHS) 85-1320. National Center for Health Statistic1985. data strongly suggest changes in this they did not complete high school." 9. The Statistical Design of the Health Household- pattern. Young women from both edu- These people would not have been in- interview survey by staff of the Us National cational categories seem to be taking up cluded in the High School Seniors Study Health Survey and the Bureau of the Census,Pub- smoking at a much lower rate than has and it is among the women in this group lic Health Service(PHS)publication 5834-A2.Na- been the case.This change in pattern is that changes in the initiation pattern 10. P Center for Health.5'tatlatlCS,ty S, $ PP 10. Pierce JP,Aldrich RN,Hanratty S,et al:Up- particularly striking for young women appear most striking. This difference take and quitting smoking trends in Australia, who do not proceed beyond high school. underlies the importance of having a 1974-1984•Prev Med 1987;16:252-260. In 1987 initiation of smoking began to representative group when undertak- 11. Bureau of the Census: Educational Attain- decline in this u While this is excit- in a population-based survey of behav- R Attain- ment in the United States: Current Population group. $ POP Y Reports:Population Chamcteristica,series P-20. ing news to those responsible for target- ior.The High School Seniors Study pro- US Dept of Commerce,1985. 60 JAMA,Jan 6,1989-Vol 261,No.1 US Smoking Trends-Pierce et al Trends in Cigarette Smoking in the United States Projections to the Year 2000 John P.Pierce,PhD;Michael C.Fiore,MD;Thomas E.Novotny,MD; Evridiki J.Hatziandreu,MD;Ronald M.Davis,MD Data from National Health Interview Surveys from 1974 through 1985 are used the same pattern. Smoking prevalence to project cigarette smoking prevalence to the year 2000. Smoking prevalence has been on the decline in the United in the United States has declined at a linear rate since 1974. If this trend States for a number of years.'Assuming continues,in the year 2000,22%of the adult population(40 million Americans) that an inverse,S-shaped curve applies will be smokers.By the year 2000,the major inequalities in prevalence will occur to this behavior decline, the decline in among educational categories.At least 30%of those who have not proceeded smoking prevalence would have started beyond a,high school education will be smokers, whereas less than 10% of slowly.This initial rate of change would then quicken and approximate a linear college graduates will smoke.Among the other sociodemographic subgroups, trend. Accompanying articles demon- smoking prevalence is expected to decrease by the year 2000 to 20%among strate that this linear stage has been men,to 23%among women,to 25%among blacks,and to 21%among whites. reached for smoking in the United Between 1974 and 1985, approximately 1.3 million persons per year became States for the major sociodemographic former smokers, indicating considerable success in public health efforts to subpopulations." However, such a lin- encourage people to stop smoking.However,in the early 1980s,approximately ear trend cannot continue but will at 1 million new young persons per year were recruited to the ranks of regular some point begin to slow down and ap- smokers.This is equivalent to about 3000 new smokers each day.Public health proach an asymptote(which may or may efforts need to focus more on preventing young people from starting to smoke, not be zero). and such prevention efforts should particularly target less educated socioeco- Population dynamics also suggest that the current linear trend observed nomic groups. in smoking data in the United States (JAMA 1989;261:61-65) will slow down at some time in the fu- ture. Reports have demonstrated that declines in smoking prevalence are RECENT trends in cigarette smoking ing prevalence. Such knowledge will strongly dependent on quitting behav- behavior are a primary source for deter- also help to evaluate current interven- ior."The continuance of a simple linear mining future prevalence. Projecting tion priorities and to ensure that they trend suggests that the same propor- future smoking prevalence, given cur- are targeting the most appropriate tion of the total population must quit rent societal influences,is critically im- groups in society. each year to maintain this decline, portant in determining appropriate An approximately S-shaped curve, regardless of the number of smokers public health smoking policy.With such which need not be symmetrical, has who are left in the population to quit. knowledge, challenging yet realistic been used to characterize a variety of Thus, a steadily increasing proportion goals can be set to further reduce smok- population changes,ranging from infec- of the smoking population must quit tious disease epidemics"and changes in each year to maintain this linear trend. From the Office on Smoking and Health,Center for health behavior'to the implementation A more likely senario is that the declin- Chronic Disease Prevention and Health Promotion, of agricultural innovations.' g The ubi ui- in trend in smoking will begin to slow Centers for Disease Control,Atlanta.Dr Fiore is now � q 1;1 with the Department of Medicine, Center for Health ty of this curve in describing both bio- down as prevalence falls to a very low Sciences,University of Wisconsin,Madison. logic and social change in populations level. Reprint requested Chief,Epidemiology Branch,Rock- suggests that population changes in This article assesses the r - Of- fice on Smoking end Health,5600 Fishers Lane,Rock- gg P Ppp opriate- ville,MD 20857(Dr Pierce). smoking behavior will probably follow ness of the linear model for projections JAMA,Jan 6,1989—Vol 261,No.1 Trends in Smoking—Pierce et al 61 1 Table 1.-Validation of Projections From Linear Models* 1987 2000 Smoking Prevalence,% Smoking Prevalence,% Population Observed Linear Logit Linear Group Value Projection P Projection Projection P Total 29.1 29.2±0.98 .92 22.8 21.7±1.33 .41 sex Male 31.7 31.7±1.06 1.00 22.1 19.9±1.58 .16 Female 26.8 26.9±1.00 .92 23.1 22.7±1.44 .79 Race Black 34.0 33.3±1.99 .73 24.3 24.5±3.52 .96 White 28.8 28.8±0.98 1.00 22.6 21.5:t 1.35 .42 Education Did not graduate from high school 35.7 33.8±0.54 .0005 31.5. 31.4±0.85 .91 High school graduates 33.1 33.8±1.13 .54 30.3 30.0±1.72 .86 Some college 26.1 26.4±1.33 .82 18.4 16.3:t 2.18 .34 College graduates 16.3 16.6±1.95 .88 9.6 4.7±3.08 .11 *Data are from National Health Interview Surveys conducted in 1974,1976,1978 through 1980,1983,1985,and 1987. of smoking prevalence to the year 2000. available from the National Health In- from the projection from the linear mod- A strict test of the validity of a model is terview Survey (NHIS) tobacco use el is reported. to test how closely a projection from supplements for 1974, 1976, 1978 To calculate population estimates that model approximates an observed through 1980,1983,1985,and 1987(pro- (Table 2), actual values were derived or a criterion value. In this case, the visional estimate only). For purposes of from weighted and age-standardized projection for 1987 can be compared this analysis,only persons aged 20 years NHIS values. These estimates are for with the recently released observed and older are included, as information the civilian, noninstitutionalized popu- data for that year. The similarity of was not collected on younger persons in lation of the United States in the year of these two values can be used to assess any consistent fashion over this time the survey. To control for changing de- the appropriateness of the model in de- period. mographic patterns between 1974 and scribing population changes in behav- All estimates were weighted and age- 1985, age standardization to the 1985 ior. However, the proximity of 1987 to standardized to the 1985 US population population estimates was undertaken the last data point(1985)limits the pre- age distribution. The current smoking using the direct method.Both the actual dictive power of the model for the year prevalence is reported for the overall and adjusted population estimates are 2000. In the absence of a criterion value adult population and for the sample presented.The change in the number of for the year 2000, another test of the stratified by sex,race(black and white), smokers in the population is a function appropriateness of the linear model is to and education (those with less than a of the number who quit smoking, the compare the estimate from this model high school diploma [<12 years of number who take up smoking, and the with that of an alternative model. As education completed],high school grad- number of current smokers who die or discussed above, the alternative model uates[12 years of education completed], migrate. Survey data allowed for esti- used must incorporate a slowing down those with some college[13 to 15 years mates of the first two of these variables. in the yearly rate of change.A log-linear of education completed],and those with model, for example, assumes that the a bachelor's degree or additional RESULTS trend in the data will approach an as- education [_-16 years of education ymptote. If the prediction from such a completed]).A line of best fit using ordi- Validity of Linear Model model is close to that from the simple nary least-squares regression was mod- Table 1 presents information on the linear model, the latter can be consid eled for each of these measurements. validity of the simple linear model as a ered to have adequate predictive pow- Slopes,SEs of the slopes,and R'coeffi- predictive tool. In all categories but er.If this is the case,the linear model is cients were computed for each linear one,the 1987 observed value fell within to be preferred,as it is much simpler to model. Values for all of these measure- 1 SE of the predicted value. For two use and understand. ments are presented in accompanying categories, whites and men, the model In this article we use the linear model articles." predicted the observed value exactly to estimate smoking prevalence to the The validity of these linear models (P=1.0). The simple linear model was year 2000 for each major demographic was tested in two ways(Table 1):(1)The not an accurate predictor of the 1987 category. We also evaluate the appro- 1974 to 1985 data were used to project prevalence for people who had less than priateness of the linear model in assess- the 1987 prevalence level.The 1987 pro- a high school education, for whom the ing future smoking behavior. Finally, visional NHIS estimate of smoking observed value was significantly higher we quantify the recent changes seen in prevalence is the criterion for compari- than the predicted value (P=.0005). smoking prevalence by the two major son of this model estimate. Using the Thus,even the modest 0.19 percentage components that could be affected by normal distribution based on the ex- point per year decline predicted by the public health efforts, the increasing pected value,we report the probability linear model in the accompanying arti- number of former smokers and the de- that this criterion is not different from cle'was an overestimate for this group. creasing number of new smokers. the expected value. (2) Similarly, the The 1987 observed value was the same METHODS linear model prediction for the year 2000 as that observed in 1985, suggesting was compared with the projection value that there may not be a declining trend Population estimates of smoking using a logit model.'Again, the proba- in smoking prevalence for this group. prevalence in the United States are bility that this criterion is not different Significantly higher logit than linear 62 JAMA,Jan 6,1989-Vol 261,No.1 Trends in Smoking-Pierce et al Table 2.—Population Estimates* Total Current Smokers Former Smokers New Smokers(Aged 20 y) Year Populatlont Actual Adjusted* Actual Adjusted* Actual Adjusted* 1974 125 084 000 46622 000 59676000 25 250 000 32320000 1324 000 1694 000 1976 126400 000 46036 000 58 005 000 25968 000 32 719 000 1243 000 1566 000 1978 139334 000 48073 000 55284 000 28 761000 33 075 000 1293 000 1487 000 1979 141091000 47861000 54 083 000 29 301000 33111000 1358 000 1535 000 1980 144590000 48942000 54326000 30064000 33371000 1206000 1327000 1983 158654000 51351000 51865000 35634000 35990000 1 141 000 1152000 1985 160298000 48791000 48791000 39918000 39918000 1077000 1077000 'Data are for adults aged 20 years and older taken from National Health Interview Surveys conducted in 1974,1976,1978 through 1980,1983,and 1985. tCivilian,noninstitutionalized population. #Adjusted to 1985 population. the percentage of whites who smoke is 50 projected to fall to about 27% in 1990 and 21%by the year 2000(Fig 2). The Men greatest discrepancies in smoking prev- alence were observed across education- 40 al levels (Fig 3). If the 1974 through 1985 trends continue to the year 2000, projected rates of smoking will be about Women 31% for persons with less than a high 30 school diploma, 30% for high school graduates, and 16% for persons with C some college education. For those with a bachelor's degree or more education, L. 20- the linear projection should be inter- preted with caution. Comparison be- tween simple linear and logit projec- tions suggests that trends in smoking 10- prevalence in this higher-educated group will slow down before the turn of the century. This probable departure from linearity has been represented in Fig 3 as a dotted line. The projected 0 difference in smoking prevalence by 1975 1980 1985 1990 1995 2000 educational category may well be an un- (est) (est) (est) derestimate given that the small decline Year from 1974 through 1985 among persons with less than a high school diploma did Fig 1.—Smoking prevalences for men and women with projections to year 2000.Lines were computed via not continue to 1987. linear regression based on observed values from National Health Interview Surveys by National Center for Health Statistics in 1974,1976,1978 through 1980,1983,and 1985.Slopes(percentage point change per Number of Smokers In Population year)are—0.91:t 0.06 for men and—0.33.t 0.06 for women.est indicates estimate. Actual and adjusted population esti- mates of current, former, and "new" (aged 20 years)smokers by survey year are shown in Table 2. The effect of an increasing population base can be seen projections for the year 2000 would indi- among adults(aged 20 years and older) in the actual population estimates, tate that the linear trend is likely to in the United States is projected to fall while little change was noted in the total slow down before the year 2000.Table 1 to about 28%by 1990 and to 22%by the number of smokers.The adjusted popu- indicates that only the projections for year 2000. The male prevalence is pre- lation estimates, however, indicated college graduates approached statisti- dicted to decrease to about 29%by 1990 that the number of smokers had actually cal significance.The logit projection for and to 20%by 2000,whereas the female declined.According to these estimates, this group is double that of the linear prevalence will decrease to 26%in 1990 the decline in smoking prevalence be- projection. In all other instances, the and to 23%by 2000 (Fig 1). If current tween 1974 and 1985 was equivalent to 1 logit and the linear projections provided trends continue, smoking prevalence million fewer smokers each year. By statistically similar estimates. among men and women will be about 1985,the number of adult current smok- Projec tions of Prevalence equal in 1995;thereafter, a larger pro- ers(aged 20 years and older)had fallen portion of women than men will be to approximately 48 million persons. Assuming that changes in smoking smokers. Using adjusted population estimates, prevalence will continue to follow the The percentage of blacks who smoke we found that the number of,former same pattern as has occurred from 1974 is projected to fall to about 32%by 1990 smokers increased by the equivalent of through 1985, smoking prevalence and to 25%by the year 2000, whereas 688 000 persons per year over the period JAMA,Jan 6,1989—Vol 261,No.1 Trends in Smoking—Pierce et al 63 from 1974 to 1985,yielding a total of just 50 under 40 million former smokers in the population in 1985. The number of 20- Blacks year-old smokers(the entry point in this database)decreased slightly from 1974 40 until 1985,when an estimated 1.08 mil- Whites lion 20-year-olds were current smokers. A similar decline occurred in the age- specific demographic base of the popula- 30 tion over this time period,"suggesting m that at least part of this decline in num- bers did not come from a decline in smoking initiation. If one assumes an I20 equal distribution of smoking uptake throughout the year, this yearly figure is equivalent to 3000 new smokers start- ing each day in the early 1980s. This figure is an estimate of the number of 10 new smokers each day who become reg- ular smokers. COMMENT 0 Accompanying reports have demon- 1975 1980 1985 1990 1995 2000 strated that cigarette smoking in gen- (est) (est) (est) der, race, and educational groups de- Year clined in a linear fashion between 1974 and 1985 in the United States."'During Fig 2.—Smoking prevalences for blacks and whites with projections to year 2000.Lines were computed via this period the linear model accounted linear regression based on observed values from the National Health Interview Surveys by National Center for almost all of the variation in the ob- for Health Statistics in 1974,1976,1978 through 1980,1983,and 1985.Slopes.(percentage point change per served data.The validity of these linear year)are—0.67±0.15 for blacks and—0.57±0.05 for whites.est indicates estimate. projections was tested by comparing the 1987 projection from the linear mod- el with the provisional estimate for that year from the NNIS.In all instances but one, the observed and expected smok- Fig 3.—Smoking prevalence by educational status with projections to year 2000.Lines were computed via ing prevalences were extremely close, linear regression based on observed values from National Health Interview Surveys by National Center for the majority being within 1 SE of the Health Statistics in 1974,1976,1978 through 1980,1983,and 1985.Slopes(percentage point change per year)are 0.19±0.03 for persons who did not graduate from high school, —0.30±0.07 for high school expected value. graduates, —0.78±0.09 for persons with some college, and —0.91±0.13 for college graduates. est The literature on patterns of social indicates estimate. change is considerable, ,"' This litera- ture strongly suggests that the long- 40 term pattern of change will follow an S- High School Graduates shaped curve. Assuming the same pattern applies to smoking behavior, Did Not Graduate From High School the linear trends observed in smoking Some College prevalence can be expected to level off at some stage in the future. A logit 30 transformation of the data before linear modeling provides for such a contingen- cy. Accordingly, comparison between oe College Graduates linear and logit projected values for the year 2000 allows for an assessment of m 20 the appropriateness of the linear model (D for short-term predictions. This com- a parison demonstrated that predictions from both models for the year 2000 were essentially the same,with the possible 10 exception of the highest education cate- gory. While these two tests of validity suggest that the simple linear model is adequate to project trends to the year 2000, other factors may influence the 0 validity of such projections. Tb begin with,the linear model does not take into 1975 1980 1985 1990 1995 2000 account the changing demographic pro- (est) (est) (est) file of the United States during the lat- Year ter part of the 20th century. Specifically,two demographic factors 64 JAMA,Jan 6,1989—Vol 261,No.1 Trends in Smoking—Pierce et al may work to accelerate the decline in who are former smokers) increased 1. Smoking is decreasing at a steady smoking prevalence in the United from about 35% to 44% at a rate of rate, and there is no current evidence States: the decreasing proportion of +0.76 percentage points per year."In that the decline in smoking prevalence persons who are entering the usual age actual numbers,this percentage trans- is beginning to slow in the overall of smoking initiation and the increasing lated into approximately 1.3 million ad- population. proportion of persons who are reaching ditional former smokers per year. By 2. The overall decline in smoking the ages during which smoking cessa- 1985,almost half of the adult Americans prevalence has been unequal across so- tion usually occurs. In this respect,our who had ever smoked were classified as ciodemographic subpopulations of soci- linear projections are conservative and former smokers. ety. Differences in smoking patterns may overestimate actual smoking prev- In contrast,efforts to prevent smok- (prevalence, quitting, and initiation) alence for the year 2000. ing initiation in the United States have that are based on educational status are Two other factors may affect future been less successful. While the overall becoming more substantial than those smoking patterns and the accuracy of adult prevalence declined 0.59 per- based on gender and race. these projections: (1) The cumulative centage points per year from 1974 to 3. Efforts to prevent smoking initia- impact of years of social disapproval of, 1985,the prevalence among young per- tion have been less effective than efforts smoking may grow in the 1990s (and sons aged 20 to 24 years declined at a to promote cessation. beyond)and may result in higher quit- rate of only 0.41 percentage points per These findings have two major public ting rates than those among the current year.'The implication from this is that health implications: (1)We need to ori- smoking population(many people began in 1985,approximately 3000 new young ent our current general emphasis and to smoke before attitudes toward smok- persons who were to become regular resources more toward the prevention ing had changed).On the other hand,it smokers started smoking each day. of smoking among young Americans. is possible that "hard-core" smokers Certain sociodemographic subgroups More than 1 million young persons start (heavily dependent or disinclined to actually showed an increase in smoking smoking each year in the United States. quit)may slow down the rate of decline prevalence between 1974 and 1985 from This population must be addressed in prevalence as they become a greater within this youngest age group. Smok- more vigorously if we are to maintain proportion of the smoking population. ing among young women who did not momentum toward the goal of a smoke- However, heavier smokers (those who proceed beyond high school education free society. (2) Public health preven- smoke 25 or more cigarettes per day) actually increased from about 39% in tion and cessation efforts must be care- did not increase in proportion to total 1974 to 44% in 1985!Although young fully targeted to reach those groups smokers (even when stratified by sex, men in this education group showed a with the highest prevalences and slow- race,and age)from 1974 to 1985." decrease in prevalence(from about 52% est rates of decline in smoking.Adoles- Although smoking rates are pro- to 46%), the overall result was that cents and, particularly, less educated jected to decline, the actual number of smoking among young persons in this persons have fallen behind in the battle persons who smoke may decrease only educational category remained essen- to reduce smoking prevalence. Public slightly over the next 11 years.The US tially unchanged, at about 45%, from health interventions must be made population of adults(aged 20 years and 1974 to 1985.Educational level has been more relevant to these more slowly older)is projected to rise to 193 million's repeatedly shown to be a marker for changing groups. Only when such tar- in the year 2000.Assuming that the non- socioeconomic status.1e This unchanged geted efforts are successful will smok- civilian and institutionalized population status suggests that smoking in the ing have the potential to lose its distinc- makes up a proportion of the total popu- United States is increasingly becoming tion as the number one preventable lation in the year 2000 similar to that in a behavior primarily of the less ed- cause of death in the United States. 1985,the 22%of adults who will smoke ucated and the socioeconomically disad- in that year will represent approximate- vantaged. Moreover, this finding indi- ly 40 million American smokers. Smok- cates that the public health effort to ing prevalence rates vary as the rates of prevent smoking initiated among mem- we thank the many pstic sio at the National smoking initiation and smoking cessa- bers of these groups has been relativelyCenter for Health Statiatica for p aproviding the re- smoking apected,high-quality data that served as the source tion change. Our analysis has shown unsuccessful. for these analyses, including Jake Feldman,Joel that cessation activity is occurring at an Important public health findings re- Kleinman,and Ron Wilson,who consulted with us impressive rate in the United States. sult from this series of reports on trends on the complex design issues and methods related From 1974 to 1985 the overall adult quit in cigarette smoking in the United to the analyses of the data.We also thank ldrsk l ' q g g Patel for statistical support and Sonia Balakirsky ratio (the proportion of ever smokers States: for secretarial support. References 1. Sartwell PE:The distribution ofincubation peri- 6. Health United States 1987,US Dept of Health Press,1975. ode of infectious disease. Am J Hygiene 1950; and Human Services publication(PHS)88-1232. 10. Bandura A: Social Foundations of Thought 51:310-318. National Center for Health Statistics,1988. and Action:ASocial Cognitive Theory.Englewood 2. Sartwell PE:The incubation period and the dy- 7. Fiore MC,Novotny TE,Pierce JP,et al:Trends Cliffs,NJ,Prentice-Hall International Inc,1986. namics of infectious disease. Am J Epidemwl in cigarette smoking in the United States: The 11. Reducing the Health Consequences of Smok- 1966;83:204-216. changing influence of gender and race.JAMA 1989; ing,US Dept of Health and Human Services.Office 3. Armenian HK,Lilienfeld A:Incubation period 261:49-55. on Smoking and Health,Center for Health Promo- of disease.Epidemiol Rev 1983;5:1-15. 8. Pierce JP,Fiore MC,Novotny TE,et al:Trends tion and Education,1989. 4. Rogers EM,Shoemaker F:Communication of in cigarette smoking in the United States:Educa- 12. Current Population Reports:Population Esti- Innovations: A Cros8-ndturial Approach, ed 2. tional differences are increasing. JAMA 1989; mates and Projections,US Dept of Commerce pub- New York,Free Press,1971. 261:56-60. lication P25 952,1984. 5. Rogers EM: Diffusion of Innovations, ed 3. 9. Bishop YM,Fienberg SE,Holland PW:Discrete 13. Social Indicators III,US Dept of Commerce. New York,Free Press,1983. Multivariate Analysis. Cambridge, Mass, MIT Bureau of the Census,1980. JAMA,Jan 6,1989-Vol 261,No,1 Trends in Smoking-Pierce et al 65 Birth Cohort Analysis of Prevalence of Cigarette Smoking Among Hispanics in the United States Luis G.Escobedo,MD,SM,MPH,Patrick L.Remington,MD,MPH To investigate historical trends of cigarette smoking among Mexican-Ameri- Data Collection cans, Cuban-Americans, and Puerto Rican—Americans, we conducted a birth The HHANES included a household cohort analysis of smoking prevalence by using smoking histories of 8286 adults interview and a subsequent three-hour and adolescents from the 1982-1983 Hispanic Health and Nutrition Examina- medical examination conducted in a mo- tion Survey. We constructed smoking prevalence curves for men and women bile examination center. Demographic among successive ten-year birth cohorts. Birth cohort-specific prevalence and smoking data were collected from rates were higher for men than for women. Rates,however,decreased among adults(20 through 72 years of age)dur- successive cohorts of men. Conversely, rates increased among successive ing the household interview and from cohorts of Cuban-American and Puerto Rican-American women.For example, adolescents(13 through 19 years of age) peak rates among the 1911 through 1920 cohorts were 26%(Cuban-American in the course of the visit to the mobile ) p pe examination center. Individuals were women)and 25%(Puerto Rican-American women compared with peak rates asked "Have you smoked at least 100 of 43%and 52%, respectively, among comparable groups from 1951 through cigarettes in your entire life?" Those 1960.These results demonstrate that despite a reduction of cigarette smoking who answered yes to this question were among successive cohorts of Hispanic men, Hispanic women have made little considered smokers and were asked progress or have actually increased their cigarette smoking. "About how old were you when you first started smoking cigarettes fairly regu- (JAMA 1989 slats-s9) larly?"Responses to this question were used to determine the age at which the respondents began smoking. Former smokers were asked "About how long has it been since you last smoked ciga- SINCE 1964,when the Surgeon Gener- a unique opportunity to conduct a retro- rettes(fairly regularly)?"Responses to al's first report on smoking was pub- spective cohort analysis of trends in cig- this question were used to determine lished, the American public has been arette smoking among Hispanics in the the age at which the respondents quit surveyed repeatedly to determine cur- United States. smokin rent and past smoking patterns. g Unfortunately,data on smoking behav- METHODS Study Population for among minority groups in this coun- The target population for this report try,including Hispanics, are not as ex- Survey Design was the 7100 adults who were inter- tensive as data on the smoking behavior Details of the HHANES sample de- viewed for the HHANES (4218 Mexi- of the general population.`Without this sign as well as the plan of operation, can-Americans, 1193 Cuban-Ameri- information, it is difficult to plan, selection process, data collection, and cans, and 1689 Puerto Rican—Amer- implement,and target successful smok- quality control are presented else- icans) and the 2543 adolescents who ing intervention programs. In lieu of where.''In brief,the HHANES used a were scheduled for interview (1555 national data on smoking among His- stratified,four-stage,cluster sample of Mexican-Americans,261 Cuban-Amer- panics from previous surveys, we used households and had selection probabili- icans, and 727 Puerto Rican—Amer- data from the (1982-1983) Hispanic ties proportional to the size of the popu- icans). Health and Nutrition Examination Sur- lation at each stage. The four stages of Among all three groups, 512 adults vey(HHANES). This survey provided selection were as follows:(1)counties or and 61 adolescents who were not His- small groups of adjacent counties mak- panic were excluded from the analysis. From the Epidemiologic Studies Branch,Division of ing up the primary sampling units, (2) Of the remaining 6588 adults,eight had Reproductive Health(Dr Escobedo),and the Epidemi- clusters of households making up the missing smoking data, and 110 either ology Branch, Division of Nutrition (Dr Remington), segments, households, and (4) per- had unknown values for smokingvari- Center for Health Promotion and Education,Centers for sons Because the Cuban-American irregularly. Disease control,Atlanta. ables or smoked irre larl . Of the re- Presented in part as a poster at the 37th Annual sample included only Dade County,Fla, maining 2482 adolescents,24 had miss- Epidemic Intelligence f 988 aService d i paConference, nf Deice Atlanta, tl us- this sample had a three-stage selection ing or unknown smoking data or smoked Mexico Border Health Association Conference, Chi- procedure.Although the areas sampled irregularly, and 335 could not be inter- huahua,Mexico.June 5-8,1988. in the HHANES are not representative viewed. The study population was thus Reprint requests to Mailstop COs,Room 4044,Bldg of all Hispanics in the United States, reduced to 6470 adults and 2123 1,Centers for Disease Control,1600 Clifton Rd,Atlanta, GA 303M(Dr Escobedo). they include a substantial percentage. adolescents. 66 JAMA,Jan 6,1989—Vol 261,No. 1 Smoking Among Hispanics—Escobedo&Remington Data Analysis Prevalence of Current Cigarette Smoking Among Hispanic Americans in 1962-1963 A statistical package for analyzing Prevalence,% complex sample survey data,' SESU- DAAN was used to compute the Sample 96%Confidence 95%Confidence P Ethnic Group Size Melee Interval Females Interval prevalence of current cigarette smoking Adults by ethnic group, sex, and adult or ado- Mexican-Americans 3777 43.6* (40.1-47.2) 24.5 (22.0-27.0) lescent status, and 95% confidence Cuban-Americans 1095 41.8* (37.7-46.0) 23.1 (20.3-25.9) intervals. Puerto Rican-Americans 1495 41.3 (37.0-45.5) 32.6t (28.2-37.1) We also computed the prevalence of Adolescents smoking for ten-year birth cohorts Mexican-Americans 1262 12.9* (11.2-14.5) 7.9 (4.9-10.9) (1911 through 1920, 1921 through 1930, Cuban-Americans 168 7.9 (2.6-13.1) 8.1 (1.4-14.7) 1931 through 1940, 1941 through 1950, Puerto Rican-Americans 489 21.71 (17.4-26.0) 19.1t (12.0-26.2) 1951 through 1960, and 1961 through *Rate among males is significantly greater than among females. 1970)of men and women in each ethnic tRate is significantly greater than among Mexican-Americans. group, beginning from the last birth year of the cohort and ending with the year of interview. For example, for the 1911 through 1920 cohort, smoking 60' prevalence rates were calculated for ev- ery year from 1920 through 1983 by us- 70 ing each individual's smoking history so 1911-1920 and sampling weight. The numerator , represented the number of individuals o 1921-i 930 smoking in each specific year, and the 501 denominator represented the entire co- m 1931-1940 hort.'Because the sampling weight for 40LD each respondent represented a large d. 1941-1950 number of persons from the population 30 l in the area sampled that varied among ' i 1951-1960 respondents (unequal selection proba- PO , bilities), the sampling weight for each ' respondent was used to provide correct to- % 1961-1970* estimates of the prevalence rate. The sampling weight also incorporated a .. . nonresponse adjustment.Smoking pre- o - valences for the 1961 through 1970 co- 1920 1930 1940 1950 1960 1970 1960 1990 horts may have underestimated the Year peak rate for these cohorts since some members may not have reached an age to begin smoking. Fig 1.-Birth cohort-specific smoking prevalence among Mexican-American men by year in Hispanic Health Because our analysis focused on com- and Nutrition Examination Survey sample.Asterisk indicates that smoking histories are probably incomplete. parisons of ten-year birth cohorts,only those born from 1911 through 1970 were included. We therefore excluded from 80- the analysis those born from 1907 through 1910(n=102)and in 1971 and 70- 1972(n=205). RESULTS 60 Among adults, the Table shows that the rates of current smoking among men a 50- were similar, that proportionally more Puerto Rican-American women smok- 0-0 ed than did Mexican-American or Cu- - ban-American women,and that propor- 30 tionally more men smoked than did women. Among adolescents, the Table 20 1931.1940 also shows that proportionally more 1921-1930 isa�-1"s5o - Puerto Rican-American boys and girls 10 1911-1920 /, 1951-1960 smoked than did Mexican-American or i�� Cuban-American boys and girls and 0; 1991-1970* that proportionally more Mexican- 1920 1930 1940 1950 1960 1970 1980 1990 American boys smoked than did Mexi- Year can-American girls. Visual examination of birth cohort curves among both sexes for adults and - Fig 2.-Birth cohort-specific smoking prevalence among Mexican-American women by year in Hispanic adolescents combined revealed impOr- Health and Nutrition Examination Survey sample.Asterisk indicates that smoking histories are probably tent differences in the rates and trends incomplete. JAMA,Jan 6,1989-Vol 261,No. 1 Smoking Among Hispanics-Escobedo&Remington 67 1930), 30% (1931 through 1940), 42% so (1941 through 1950), and 43% (1951 through 1960).The rate of smoking inti- ,'— ation increased after the 1921 through 7o i 1930 cohort, and smoking cessation / rates remained slow among successive 60 v....... .' cohorts. �.� Peak smoking rates among succes- 50- i sive cohorts of Puerto Rican—American r ' men diminished slightly(Fig 5). Unlike d 40-. 1911-1920 ;� ^t Mexican-American and Cuban-Ameri- 1921-i93o ; i can men,who displayed higher rates of 30, 1931-1940 i smoking cessation among successive 1941-1950 ;' t cohorts, Puerto Rican—American 20 i 1951-1960 men displayed no such trend. Peak smoking rates among successive co- 10 /�� �; ts61111s7o horts of Puerto Rican—American wom- en increased substantially(Fig 6).Peak 0 ___ rates were 25% (1911 through 1920), 1920 1930 1940 1950 1960 1970 1980 1990 37% (1921 through 1930), 42% (1931 through 1940), 37% (1941 through Year 1950), and 52% (1951 through 1960). Smoking initiation rates increased and Fig 3.—Birth cohort-specific smoking prevalence among Cuban-American men by year in Hispanic Health smoking cessation rates remained slow and Nutrition Examination Survey sample.Asterisk indicates that smoking histories are probably incomplete. among successive cohorts, just as among Cuban-American women. COMMENT 80- Inspection of birth cohort—specific 70- curves of smoking rates revealed that smoking patterns for men and women 60 within each ethnic group were begin- ning to converge over time. The preva- 50- lence of cigarette smoking among successive cohorts of men declined g 40 i =�'� markedly among Mexican-American m 1951-1960 men and slightly among Cuban-Ameri- 30 1941-1950 /' ,...._............... can and Puerto Rican—American men. Conversely, rates of cigarette smoking 20- 1 931-1940 ------ among successive cohorts of Mexican- 1 s21-1 s3o American women changed little and t o 191 ,1 1- s2o ; actually increased markedly among �' ;' 1961-16ioi Cuban-American and Puerto Rican- 0 ........ ........... ican—o- _.—.% American women. 1920 1930 1940 1950 1960 1970 1980 1990 Comparison of cohort curves among Hispanics in this population with cohort Year curves for the general population"re- vealed contrasts:(1)Declines in rates of Fig 4.—Birth cohort-specific smoking prevalence among Cuban-American women by year in Hispanic cigarette smoking among successive Health and Nutrition Examination Survey sample.Asterisk indicates that smoking histories are probably cohorts of Cuban-American and Puerto incomplete. Rican—American men were slight com- pared with declines among men in the of smoking. First, for all comparable rates of smoking initiation among general population. (2) Rates of cig- cohorts, Mexican-American men had successive cohorts of Mexican-Amen- arette smoking among successive co- higher lifetime smoking rates than can women increased,although rates of horts of Cuban-American and Puerto women (Figs 1 and 2). Among succes- smoking initiation among these women Rican—American women increased mar- sive cohorts of men, the rates of smok- increased more gradually than among kedly compared with the increases ing initiation(the upward slope of these Mexican-American men. among successive cohorts of women in curves)decreased slightly, and higher- Among successive cohorts of Cuban- the general population. (3) Visual ex- rates of smoking cessation (the down- American men, peak smoking rates di- amination of the downward slope of the ward slope of the curves) occurred at minished slightly,except for those born cohort curves revealed that men and progressively earlier ages.The highest from 1941 through 1950(Fig 3).Higher women in the general population expe- lifetime smoking rate among Mexican- rates of smoking cessation occurred at rienced relatively more rapid and sub- American women occurred among those progressively earlier ages. Peak smok- stantial smoking cessation rates than born from 1931 through 1940(35%). In ing rates among successive cohorts of Hispanic men and women in this contrast with the peak rates for men, Cuban-American women increased sub- population. peak rates among successive cohorts of stantially(Fig 4). Peak rates were 26% This retrospective analysis had cer- women diminished little. Moreover, (1911 through 1920),25%(1921 through tain potential biases: (1) There was a 68 JAMA,Jan 6,1989—Vol 261,No.1 Smoking Among Hispanics—Escobedo&Remington 1 younger cohorts but not between men 80- and women or among ethnic groups in the same cohorts. 70- Several preliminary conclusions were --------- suggested by our findings.The previous two decades' passive and active cam- so paigns aimed at curbing cigarette smok- ing among the public in general ap- 50 peared to have had some impact on the smoking habits of Mexican-American _d ao i - ! men. No comparable improvements 1941.1950 30 1931-1940 1951-1990 ! were apparent in the smoking patterns (1921-19;30 1 ! of other Hispanic men and women;ciga- 1911-1920 �� 1961-1970* rette smoking actually became more 20 ; prevalent among Cuban-American and i Puerto Rican-American women. 101 ,' ' ,i Historical data on smoking preva- lence can be useful qualitatively for esti- 0 mating future trends of diseases attrib- 1920 1930 1940 1950/ 1960 1970 1960 1990 utable to smoking. For example, the diminishing proportion of Mexican- Year American men who smoked among successive cohorts suggests that the in- Fig5.—Birth cohort—specific smoking g y year in Hispanic cidence of diseases attributable to p g prevalence.Asterisk Puerto Rican—American men b smoking among younger cohorts will be Health and Nutrition Examination Survey sample.Asterisk indicates that smoking histories are probably g g y g incomplete. lower in the future compared with the incidence among older cohorts. In con- 80- trast, the increasing proportion of smokers among successive cohorts of 70- Cuban-American and Puerto Rican- American women suggests that the inci- so; dence of diseases-attributable to smok- ing among younger cohorts will be greater in the future than the incidence 50 i - among older cohorts. c Because the proportion of smokers 40 _ ' among successive cohorts of Cuban- ' `1 American and Puerto Rican-American 0 30 women appeared to be increasing mark- i edly and because the proportion of cur- 20- ' ' ! rent cigarette smokers among Puerto 1951-1960 Rican-American adolescents was hi h 1931-1940 1941-1950 / g 10", 1961-1670* intervention efforts must be targeted 1911-1920 1921-1930 _ toward these groups. --� ' We thank Joe Fred Gonzalez, MS, and Olivia 1920 1930 1940 1950 1960 1970 1980 1 g90Carter,MS,of the National Center for Health Sta- Year tistics and Van Munn and Kevin Sullivan,MPH,of the Centers for Disease Control,Atlanta,for their Fig 6.—Birth cohort—specific smoking prevalence among Puerto Rican—American women by year in Hispan- ic Health and Nutrition Examination Survey sample.Asterisk indicates that smoking histories are probably analysis. incomplete. References possibility of selective survival,because through 1910 cohort. However, there 1. The Health Consequences of Smoking for Wom- smokers are more likely to die than non- was no appreciable increase in the cor- en, A Report of the Surgeon General. Office on Smoking and Health,US Dept of Health and Hu- smokers. The prevalence curves re- rected rates compared with the uncor- mar,Services,1980. flected the smoking habits of those who rected rates after 1910. Because we in- z. Gonzalez J,White AA,Ezzati T:Sample Design were alive to be interviewed. The sur- cluded cohorts after 1901 through 1910, for the Hispanic Health and Nutrition Examina- vey therefore may have underestimat- our estimates of smoking prevalence tion Survey,1982-84:Proceedings of the Survey ed how manmembers of the older birth were probably not reduced spuriously. Research Methods Section. Washington, Dc, y p yAmerican Statistical Association,1984. cohorts smoked. But at least among (2)The possibility of recall bias existed 3. National Center for Health Statistics:Plan and those in the general population,this un- because of differences in the ability of Operation of the Hispanic Health and Nutrition der-estimate occurred only among birth younger and older individuals to recall Examination Survey,1982-84.US Dept of Health cohorts prior to 1910.'That is, correc- dates of smoking initiation and cessa- q Shah an Services,1985,SESUDAA5,StaSeries 1, Errors Pro- tion of smoking prevalence for differ- tion. Older individuals may have been gram for Computing of Standardized Rates From ences in mortality rates between smok- less likely than younger individuals to Sample Survey Data. Research Triangle Park, ers and nonsmokers increased the recall the year of smoking initiation or NC Research Triangle Institute,1981. prevalence of smoking appreciably for cessation. Thus this bias may have af- 5.birth cohorts its Cigarette smoking among successive p g pp y r y birth cohorts of men and women in the United cohorts up to and including the 1901 fected comparisons between older and States during 1900-80.JNCI 1983;71:473479. JAMA,Jan 6,1989—Vol 261,No.1 Smoking Among Hispanics—Escobedo&Remington 69 ■ Comparing the Prevalence of Smoking . in Pregnant and Nonpregnant Women , 1985 to 1986 David F.Williamson,PhD;Mary K.Serdula,MD;Juliette S.Kendrick,MD;Nancy J.Binkin,MD The 1990 health objectives for the nation state that pregnant women should be surveillance system is primarily in- only half as likely to smoke as nonpregnant women.To assess progress toward tended to provide state-specific esti- meeting this objective, we used cross-sectional data from the 26 states in the mates of the prevalence of personal Behavioral Risk Factor Surveillance System in 1985 and 1986. We compared health practices related to the leading the prevalence of self-reported smoking among pregnant (N=836) and non- causes of death among adults 18 years of pregnant (N=18 025) women aged 18 to 45 years. Overall, pregnant women age and older. Data are collected by were 70%as like) to be current smokers as nonpregnant women(prevalence telephone interview and include infor- mation on smoking practices, alcohol ratio,0.7;95%confidence interval,0.6 to 0.8),while blacks showed the largest use,blood pressure control,weight con- pregnancy-associated reduction in the prevalence of smoking (prevalence trol,physical activity,and seat belt use. ratio,0.5;95%confidence interval,0.3 to 0.9).Most of the difference in smoking Each state conducts approximately prevalence occurred not because pregnant women were less likely to have ever 100 interviews per month(range, 50 to smoked, but because pregnant women were more likely to have quit smoking 250), for an average of 1200 completed than nonpregnant women. However, unmarried pregnant white women were interviews per year per state (range, 40% more likely to smoke than their nonpregnant counterparts (prevalence 600 to 3000)• All states use an identical ratio, 1.4;95%confidence interval, 1.1 to 1.7).We conclude from this analysis core questionnaire.Centers for Disease that the 1990 health objective for smoking among pregnant women is unlikely to Control staff provide standardized be achieved. Clinicians providing care to pregnant women need to a in- training in survey administration and P 9 P 9 pay interview procedures in each state. creased attention to smoking cessation. Respondents are selected randomly (JAMAisss;2sl.7a74) from each state's noninstitutionalized civilian population with telephones by using a multistage cluster design based THE HARMFUL effects of maternal currently pregnant and nonpregnant on the Waksberg method.' Between smoking have been well documented.' women overall and by age, race, and 1985 and 1986 the median response rate The 1990 health objectives for the na- education. We also examine the smok- (the ratio of completed interviews to the tion'state:`By 1990, the proportion of ing behavior of unmarried pregnant sum of completed interviews and refus- women who smoke during pregnancy women,a group for which data are espe- als) among participating states was should be no greater than one-half the cially limited.`We based our analysis on 84%•A detailed technical description of proportion of women overall who cross-sectional data collected from 1985 the Behavioral Risk Factor Surveil- smoke." However, it was reported in through 1986 from 25 states and the lance System has been presented 1986 that progress toward this objec- District of Columbia,in which pregnant elsewhere. tive cannot be measured because ade- and nonpregnant women were included Zb obtain an adequate number of quate data are not available.' in the same sampling frame. pregnant women for the analysis, we The purpose of this article is to pro- pooled data from the participating vide interim estimates of the difference METHODS states.Between 1985 and 1986 a total of in the prevalence of smoking between Data Source 19124 interviews were completed From the Department of Health and Human services, among women aged 18 to 45 years in the Public Health Service,center for chronic Disease Pre- Data for this analysis are from the 25 states and the District of Columbia ventionand Health Promotion, Centers for Disease Behavioral Risk Factor Surveillance (Alabama, Arizona, California, Con- Control,Atlanta. System, a collaborative system of par- necticut,Florida,Georgia,Hawaii,Ida- Reprint requests to Nutrition Division, Centers for an sae health departments and ho Illinois Kentucky, Massachusetts Disease control,Bldg 3,SB/45-A,Mailstop A-41,Atlan- ticiti p gttp > > ' ta,GA 30333(Dr Williamson). the Centers for Disease Control. This Minnesota, Missouri, Montana, New 70 JAMA,Jan 6,1989—Vol 261,No.1 Smoking Prevalence—Williamson et al York, New Mexico, North Carolina, Table 1.-Demographic Characteristics of Pregnant and Nonpregnant Women Aged 18 to 45 Years, North Dakota, Ohio, Rhode Island, Behavioral Risk Factor Surveillance System,1985 to 1986 South Carolina, Tennessee, West Vir- ginia,Wisconsin,and Utah).Pregnancy Women Women nt status was ascertained at the end of the ' Characteristic No. % Ilio. % interview by asking: `ro your knowl- edge,are you now pregnant?"A total of gTotel sample 836 100.0 1eo2s 100.0 A 8136 women(4.4%)answered "yes"and Age. 262 31.4 3649 20.2 18 025 (94.2%) answered "no."The re- 25.29 302 36.1 3647 20.2 cords of 89 respondents who said they 30-34 174 20.8 3897 21.6 were not,sure(0.5%), 110 who refused 35-45 96 11.7 6832 37.9 to answer(0.6%),and 64 whose respons- Race Es were missing(0.3%) were excluded White 662 79.2 14620 81.1 from analysis. Black 92 11.0 1791 9.9 Before being asked pregnancy status, Other 82 9.8 1614 9.0 respondents were asked the following Education questions about smoking practices: <Hgh school 8o 9.6 1777 9.9 1. Have you smoked at least 100 ciga- High school 304 36.4 6570 36.5 rettes in your life? >Hgh school 216 25.8 4879 27.1 2. If yes, do you smoke cigarettes College graduate 236 28.3 4785 26.6 now? Missing 0 0 14 0.1 3. If yes, on the average, about how manta)status many cigarettes a day do you now Married 684 81.9 10512 58.4 smoke? Unmarried 151 18.1 7491 41.6 Ever smokers were defined as those Missing 1 0.1 22 0.1 who responded"yes"to the fust ques- tion. Current smokers were defined as those who responded"yes"to both the first and second questions. Former Table 2.-Differences in the Prevalences of Smoking Practices Between Pregnant and Nonpregnant Women, Smokers were defined as those who re- Behavioral Risk Factor Surveillance System,1985 to 1986• sponded"yes"to the first question and "no" to the second question. Heavy Prevalence,% Prevalence Prevalence smokers were defined as current smok- once,n Ratio Pregnant Nonpregnant (95% CConfidence (95%Confidence ers who reported smoking 20 or more Smoking Status women women Interval) Interval) cigarettes per day. current 21 30 -9 (±2.8) 0.7 (0.6-0.8) Ever 43 45 -2 (±3.4) 1.0 (0.9-1.0) Data Analysis Fortner 22 15 7 (±2.9) 1.5 (1.3-1.7) We used the prevalence difference Heavyt 38 52 -14 (s 7.3) 0.7 (0.6-0.9) (PD) and the prevalence ratio (PR) to •All estimates were age-adjusted by direct standardization to the age distribution of pregnant women using the compare the prevalence of smoking be- following weights:0.314(18 to 24 years),0.361(25 to 29 years),0.208(30 to 34 years),and 0.117(35 to 45 years). tween pregnant and nonpregnant wom- tTwenty or more cigarettes per day among current smokers. en.The PD reflects the absolute differ- ence in the prevalence of smokers and allows assessment of the public health and more likely to be married than non- Because the sample was obtained by importance of the differences observed. pregnant women(Table 1).We adjusted pooling across states, state-specific The PR expresses the difference in rela- for age in the analysis because (1) the sampling weights were not used in this tive terms to more directly assess the age distributions were different be- analysis. With unweighted analyses, strength of the association between tween pregnant and nonpregnant wom- the sampling design of the Behavioral smoking and pregnancy.' en, (2)age is related to both the initia- Risk Factor Surveillance System has The prevalence of ever smoking tion and the cessation of smoking, and minimal effect on the estimated sam- among pregnant and nonpregnant wom- (3)age is an important biologic determi- piing variances.° Standard methods en was different for some of the sub- nant of whether a woman will become were used to compute aproximate 95% groups studied,reflecting differences in pregnant.We obtained age-adjusted es- confidence intervals for crude and ad- smoking initiation. Therefore, we also timates by direct standardization ac- justed estimates.' Prevalence differ- estimated the PD and PR for current cording to the age distribution of the ences were reported to the nearest smoking attributable to quitting. The sample of pregnant women. whole percent, and PRs to the nearest PD was computed as the prevalence of In the total sample,we estimated the tenth. former smoking among nonpregnant age-adjusted PD and PR between preg- RESULTS women minus the prevalence of former nant and nonpregnant women for cur- smoking among pregnant women. The rent smoking, ever smoking, former The prevalence of current smoking PR was computed as the prevalence of smoking, and heavy smoking. The PD was 9 percentage points lower in preg- former smoking among nonpregnant and PR for current smoking also were nant than in nonpregnant women;in rel- women divided by the prevalence of for- estimated by age group and, after ad- ative terms,pregnant women were 70% mer smoking among pregnant women. justment for age,by race,education lev- as likely as nonpregnant women to be The samples of pregnant and non- el,and marital status.The median num- current smokers (Table 2). There was pregnant women were similar in their ber of cigarettes smoked per day also only a 2-percentage point difference be- distributions of race and education; was estimated for pregnant and non- tween pregnant and nonpregnant wom- pregnant women tended to be younger pregnant current smokers.a en in their prevalence of ever smoking. JAMA,Jan 6,1989-Vol 261,No.1 Smoking Prevalence-Williamson et al 71 Because the two groups were similar in Table 3.-Differences in the Prevalences of Current Smoking Between Pregnant and Nonpregnant Women, their prevalence of ever smoking,near- by Age,Race,Education,and Marital Status,Behavioral Risk Factor Surveillance System,1985 to 1986 ly 80%of the PD in current smoking was Prevalence,% Prevalence Prevalence attributable to quitting rather than to - Difference,% Ratio differences in smoking initiation (un- Pregnant Nonpregnant (95%Confidence (95%Confidence rounded PDs, 6.9%-8.8%). Similarly, Characteristic Women woman Interval) Interval) more than 90% of the PR for current Age y 1e-2a 21 2e -7 (±s) o.s (0.6-1.0) smoking was attributable to quitting 25-29 23 32 -9 (±5) 0.7 (0.6-0.9) (unrounded PRs, 0.68_0.73). Among 30-34 14 29 -15 (±5) 0.5 (0.3-0.7) smokers, pregnant women were less 35-45 26 29 -3 (:t9) 0.9 (0.6-1.2) likely than nonpregnant women to be Race- heavy smokers. The PD for heavy White 22 30 -8 (±3) 0.7 (0.6-0.8) smoking between pregnant and non- Black 16 30 -1a (±8) 0.5 (0.3-0.9) pregnant women was -14 percentage Other 18 25 -7 (±9) 0.7 (0.5.1.2) points and the PR was 0.7. Education* In Table 3.we present the PD and PR <High school 35 47 -12 (±11) 0.7 (0.5-1.0) for current smoking by age and by race, High school 28 36 -8 (±5) 0.8 (0.6-0.9) education, and marital status adjusted >High school 17 25 -8 (±5) 0.7 (0.5-0.9) for age. Within most of the strata the College graduate 14 20 -6 (±5) 0.7 (0.5-1.0) differences in the prevalence of current MaYesrried* 18 27 -9 (±3) 0.7 (0.5-0.8) smoking between .pregnant and non- No 36 34 2 (±8) 1.1 (0.8-1.3) pregnant women were similar to the dif- ferences observed for the total sample. *Estimates were age-adjusted by direct standardization to the age distribution of pregnant women using the There were two exceptions to this pat- following weights:0.314(18 to 24 years),0.361(25 to 29 years),0.208(30 to 34 years),0.117(35 to 45 years). tern, however. First, among women aged 35 to 45 years,the PD for current smoking was only-3 percentage points stratification were imprecise because of smoked by these older pregnant women (PR, 0.9). Second, among unmarried small sample sizes, the results sug- (20 per day)was equal to that of their women, the prevalence of smoking gested that race strongly modified the nonpregnant counterparts. Among among pregnant women was 2 percent- effect of marital status on maternal blacks,the median number of cigarettes age points higher than that among non- smoking. Among unmarried whites, smoked by pregnant women (ten per pregnant women(PR, 1.1).Differences pregnant women had a substantially day) was the same as that smoked by in ever smoking between pregnant and higher prevalence of current smoking nonpregnant women. However, this nonpregnant women did not account for than did nonpregnant women (PD, 13 figure was lower than the median num- these findings. percentage points; 95% confidence in- ber smoked by pregnant white women In two instances, the pregnancy-as- terval, ±11 percentage points:PR,1.4; (15 per day) or by nonpregnant white sociated reduction in current smoking 95%confidence interval, 1.1 to 1.7). By women (20 per day). Within all educa- was much greater than that observed in contrast, among married whites and tion levels,except the college educated, the overall sample. For women aged 30 among both married and unmarried pregnant women smoked fewer ciga- to 34 years,the PD was-15 percentage blacks, the pregnancy-associated re- rettes than nonpregnant women. Mari- points and for black women the PD was duction in the prevalence of current tal status was not related to the median -14 percentage points; in both cases smoking was similar to that in the total number of cigarettes smoked by preg- the PR was 0.5.Much of the pregnancy- sample.Differences in ever smoking be- nant women:both married and unmar- associated difference in smoking among tween pregnant and nonpregnant wom- ried pregnant women smoked the same women aged 30 to 34 years, however, en did not account for these findings. amount(12 per day). could be attributed to differences in The PD for current smoking had a COMMENT ever smoking between pregnant and different relationship with education nonpregnant women. In this age group than did the PR(Table 3). The PD was In this cross-sectional study of a geo- the PD attributable to quitting was -5 inversely related to education, ie, the graphically diverse sample of women 18 percentage points (PR, 0.8). In con- reduction in the prevalence of maternal to 45 years of age, the prevalence of trast, differences in ever smoking be- smoking tended to be largest in those current smoking was 9 percentage tween pregnant and nonpregnant black with the lowest education. However, points lower among pregnant women women accounted for only a small part because women with lower education than among nonpregnant women. In of the pregnancy-associated reduction also tended to have a higher prevalence relative terms,we found that pregnant in current smoking;the PD and PR at- of smoking, the PR remained nearly women were only 70%as likely as non- tributable to quitting were -11 per- constant across educational levels. pregnant women to be current smokers. centage points and 0.5,respectively. To determine whether pregnant Most of this difference occurred because Because of the marked differences in smokers smoked fewer cigarettes than pregnant women were more likely than the marital status of white and black nonpregnant smokers,we examined the nonpregnant women to have quit smok- pregnant women (87% of whites and median number of cigarettes that re- ing and not because pregnant women 46%of blacks were married),we strati- spondents reported smoking per day, were less likely to have ever smoked. fied marital status by white and black stratifed by age, race, education, and Among current smokers, pregnant race to further examine the association marital status (Figure). Overall, preg- women were less likely than nonpreg- between smoking and pregnancy in un- nant women smoked fewer cigarettes nant women to be heavy smokers, and married women. (There were too few (12 per day)than did nonpregnant wom- they smoked fewer cigarettes per day unmarried pregnant women in the"oth- en(20 per day).This was true within all than their nonpregnant counterparts. er" group for stratification.) Although age groups,except those aged 35 to 45 These results suggest that unless ma- the PD and PR estimates obtained after years;the median number of cigarettes jor changes in maternal smoking have 72 JAMA,Jan 6,1989-Vol 261,No.1 Smoking Prevalence-Williamson et al nancy-associated reduction in smoking a 3o T 30 occurred only among whites. The 1982 0 25 0 25 National Survey of Family Growth10 and ga the 1985 National Health Interview m 20 20 Survey" also found that unmarried LD 15 15 women were considerably more likely 10 10 than married women to smoke during opregnancy.A study using the 1984 Mis- z 5 6 5 souri Birth Certificate Registry's found that the prevalence of maternal smok- ° 18-24 25.29 30-34 35-45 ° White Black Other inig was also substantially higher among unmarried women but that marital sta- tus was related more strongly to mater- nal smoking among whites than among blacks. We found that black women showed 30 30 the largest pregnancy-associated re- 25 0 25 duction in smoking prevalence,even af- ter we accounted for differences in 20- 20- smoking initiation. As did a previous 7777;a study," we found that blacks smoked 0 15 m 15- fewer cigarettes than whites.Three na- 10 L) to- tional surveys found a lower prevalence o 5 of maternal smoking among blacks than 0 5 z 5- among whites."'Y However, two of 0F 1 0these surveys also found that whites <High High >High College Married unmarried were more likely than blacks to quit School School School Marital Status smoking after becoming pregnant"A; Education those findings suggest that the lower prevalence of maternal smoking among Median number of cigarettes smoked per day by pregnant and nonpregnant smokers, Behavioral Risk blacks was due to a lower prevalence ofever SmOkln A recent stud found Factor Surveillance System, 1985 to 1986. Dotted bars indicate pregnant women;crosshatched bars, g y nonpregnant women. that although blacks of both sexes are less likely than whites to quit smoking, occurred since 1985 to 1986, the 1990 firmed by the findings of earlier studies. both races seem to have the same preva- health objective for the nation to lower A report on maternal smoking from a lence of ever smoking." Hence, evi- the prevalence of smoking among preg- national sample interviewed in 1980 es- dente to support our finding remains nant women to no more than half that of timated that among married pregnant equivocal. More studies are needed to nonpregnant women will not be met, women aged 20 years or older, 25% of examine ethnic differences in smoking especially among older women and un- whites and 23% of blacks smoked." In cessation before, during, and after married white women. 1980 another report estimated the prev- pregnancy. The results of this study are not di- alence of smoking among women in the Because of the small sample sizes in rectly comparable with the results of general population at 33%." The PD some subgroups, these results need to other studies of maternal smoking. (-8 percentage points [25%-33%1) be interpreted with caution. Estimates Most other studies restricted their sam- and the PR(0.76[25%-33%])that can of the difference in smoking prevalence ple to formerly pregnant women, be estimated from these two studies are between pregnant and nonpregnant whereas the goal of this study was to similar to the estimates from our study. women were imprecise within several of compare pregnant with nonpregnant We found that pregnant women older the subgroups studied. Also, although women. These previous studies esti- than 35 years of age had a smaller reduc- differences in maternal smoking be- mated the prevalence of smoking during tion in the prevalence of smoking than tween subgroups were striking and the most recent pregnancy(regardless other age groups and also smoked the seemed to be exceptions to the overall of when the pregnancy occurred)'° or highest median number of cigarettes pattern,there was not adequate power during the 12 months preceding the per day of any subgroup studied. Al- to demonstrate"statistical significance" birth of a child within the last five though women in this age group make across subgroups.Hence,it must be re- years" or documented, for a specific up a minority of pregnant women,their membered that some of these results year, pregnancy-related changes in older age makes them especially suscep- are only suggestive and should be veri- smoking practices.' In addition, each tible to the deleterious effects of mater- fied by other independent studies. study had different sampling frames, nal smoking.16 The 1980 National Natal- In addition,information on why wom- survey methods, and response rates. ity Survey also found that pregnant en quit smoking was not collected in this For example, some studies included women aged 35 years and older were study.Thus,we were not able to deter- married women only' whereas others least likely to have quit smoking during mine whether pregnant women who re- included only those married women who pregnancy.' ported being former smokers quit gave birth to live-born infants.'s Some In the present study, unmarried smoking because of their current preg- studies employed face-to-face inter- pregnant women were the only sub- nancy or because of other factors, such views"`whereas others used mail and group that did not have a lower preva- as smoking cessation interventions. telephone surveys.' lence of smoking than their nonpreg- Stage of gestation also was not deter- Despite these differences, the infer- nant counterparts; after stratifying by mined;this could be important because ences from this study generally are con- race we found that this lack of a preg- it is known that quitting smoking as late JAMA,Jan 6,1989—Vol 261,No. 1 Smoking Prevalence—Williamson et al 73 as the fourth month of pregnancy has in this study may be too low.In general, results of this study may be too high. benefit for the fetus.' Because this however, we believe that rather than 1b further reduce the prevalence of study was cross-sectional, the number biasing the estimated difference in maternal smoking, the current effec- of pregnant women who ultimately quit prevalences, the above factors have tiveness of smoking cessation programs smoking before delivery also may be more likely led to our underestimating for pregnant women will need to be im- underestimated. the prevalence of current smoking for proved."This will require that efforts to Three characteristics of this multi- both pregnant and nonpregnant wom- properly evaluate smoking cessation state sample also need to be considered en. The overall estimates of the preva- programs be intensified and that those when the results are interpreted.First, lence of smoking for pregnant and non- programs found to be most effective be because this sample included only those pregnant women in this sample were rapidly deployed.Our results and those states that participated in the Behavior- slightly lower than those reported in 'of other studies indicate that unmarried al Risk Factor Surveillance System, it other studies carried out earlier in the mothers are at highest risk of continu- does not constitute a representative 1980s,1,'although these differences in ing to smoke during pregnancy. It has sample of US women.Second,this sam- prevalence may partly reflect the con- been pointed out previously that be- ple was restricted to women from tinuing decline in smoking that has tak- cause such high-risk mothers are most households with telephones.In general, en place since the earlier studies." likely to rely on public agencies for pre- telephone-owning households tend to Because pregnancy status was asked natal care,these agencies have a special have fewer risk factors for ill health only at the end of the interview,a major responsibility to ensure that effective than do households without tele- strength of this study was that the smoking cessation programs be made phones.18 Third, in the Behavioral Risk respondents were unaware that the in- available to all pregnant smokers under Factor Surveillance System, 16%of all formation they reported on smoking their care."In addition,individual pro- households contacted by telephone did practices would be related to their preg- viders of prenatal care can be important not complete the interview, and it is nancy. It has been estimated that in participants in helping pregnant smok- known that smokers are less likely than 1985 between 68%and 85%of US wom- ers to quit,using readily available mate- nonsmokers to respond to health en aged 18 to 45 years were aware that rials' and a variety of intervention surveys.19 maternal smoking increased the risks of modalities.' For these characteristics to have bi- a poor pregnancy outcome.''0 Such ased our estimates of the difference in knowledge may lead women to deny We acknowledge the state coordinators of the Cen- the prevalences of smoking between smoking or to underestimate the uan- Behavioral Risk Factor Surveillance System,hose p g g q fere for Disease Control,Atlanta,without whose pregnant and nonpregnant women, tity smoked-particularly when they help these data would not have been available.In however, these factors would also need are asked about their smoking practices addition, we thank Gary Hogelin, MPA, chief of to be related to pregnancy status. We in relation to their own pregnancy, as Field Services for the Behavioral Risk Factor Sur- veillance System;Eileen Gentry,chief of the Be- believe this is unlikely, although one has been done in most other studies of havioral Risk Factor Surveillance System Statis- potential bias (suggested by an anony- maternal smoking.1 " However, it may tics Section; and Patrick L. Remington, MD, mous reviewer) is that among single be that pregnant women in general are Wisconsin Division of Health, Madison, for their women,single mothers may be less like- less likely to admit to smoking,regard- assistance. we also thank Ronald Davis, MD, 1 to have a telephone.If single mothers less of whether the are aware that James Marks, FreMD, Carol Hogue, PhD, Robert y p g Y Anda,MD,and Frederick Trowbridge,MD,of the have a higher smoking prevalence and their pregnancy status will be asked. If Centers for Disease Control, Atlanta, for their are somewhat underrepresented in this this is the case, then the magnitude of careful review of earlier drafts of this manuscript. sample, then the pregnancy-associated the pregnancy-associated reduction in We are grateful to Kenneth Rothman,DrPh,of the excess prevalence of smoking observed smoking prevalence suggested b the University of Massachusetts for advice on data p g g P gg Y analysis. References 1. Office on Smoking and Health:Pregnancy and methods,and estimates from combined state data. pregnant?in:Smoking Behavior and Policy Dis- infant health, in: The Health Consequences of Am J Prev Med 1985;1:9-14. cussion Paper Series.Cambridge,Mass,Harvard Smoking for Women:A Report of the Surgeon Gen- 10. National Center for Health Statistics,Mosher University Press,1986,pp 1-18. eras.Office of the Assistant Secretary for Health, WD,Pratt WF:Fecundity,Infertility,and Repro- 17. Novotny TE,Warner KE,Kendrick JS,et al: 1980,pp 189-249. ductive Health in the United States,1982,US Dept Smoking among blacks and whites:Socioeconomic 2. McIntosh ID:Smoking and pregnancy:Attrib- of Health and Human Services publication(PHS) and demographic differences.Am J Public Health utable risks and public health implications.Can J 87-1990.Public Health Service,1987. 1988;78:1187-1189. Public Health 1984;75:141-148. 11. National Center for Health Statistics,Schoen- 18. Woltte EM:Characteristics of persons with and 3. US Dept of Health and Human Services: The born CA:Health Promotion and Disease Preven- without home telephones.J Marketing Res 1979; 1990 Health Objectivesfor the Nation:AMideourse tion:United States,1985,US Dept of Health and 16:421.425. Review.Public Health Service,1986,pp 179-180. Human Services publication(PHS)88-1591.Public 19. Madans JH,Kleinman JC,Cox CS,et al:Ten 4. Kleinman JC,Pierre MB,Madans JH,et al:The Health Service,1988. years after NHANES I:Report of initial follow-up, effects of maternal smoking on fetal and infant mor- 12. Prager K,Malin H,Spiegler D,et al:Smoking 1982-1984.Public Health Rep 1986;101:465473. tality.Am J Epidemiol 1988;127:274-282. and drinking behavior before and during pregnancy 20. Fox SH,Brown C,Koontz AM,et al:Percep- 5. Waksberg JS:Methods for random digit dialing, of married mothers of live-born infants and still- tions of risks of smoking and heavy drinking during JAm Stat Assoc 1978:73;40-46. born infants.Public Health Rep 1984;99:117-127. pregnancy:1985 NHIS findings.Public Health Rep 6. Remington PL, Smith MY,Williamson DF,et 13. Kleinman JC, Kopatein A: Smoking during 1986;102:73-79. al:Design,characteristics,and usefulness of state- pregnancy, 1967-1980.Am J Public Health 1987; 21. Lincoln R: Smoking and reproduction. Fhm based behavioral risk factor surveillance: 1981- 77:823-825. Plann Phrspect 1986;18:79-84. 1986.Public Health Rep 1988;103:366-375. 14. National Center for Health Statistics:Health, 22. Davis RM: Uniting physicians against smok- 7. Rothman KJ: Modern Epidemiology. Boston, United States,1985,US Dept of Health and Human irW The need for a coordinated national strategy. Little Brown&Co Inc,1986. Services publication(PHS)86-1232.Public Health JAMA 1988;259:2900-2901. 8. Velleman PF, Hoaglin DC: Applications, Ba- Service,1985. 23. Kottke TE,Battista RN,DeFriese GH,et al: sic8 and Computing of Explaratory Data Analy- 15. Cnattingius S, Axelsson 0,Eklund G, et al: Attributes of successful smoking cessation inter- 8a8.Boston,Duxbury Press,1981. Smoking, maternal age, and fetal growth. Am J ventions in medical practice:A meta-analysis of 39 9. Gentry EM,Kalsbeek WD,Hogelin GC,et al: Obstet Gynecol 1985;66:449-452. controlled trials.JAMA 1988;259:2882-2889. The behavioral risk factor surveys: II. Design, 16. Land GH,Stockbauer JW:Who smokes while 74 JAMA,Jan 6, 1989-Vol 261,No. 1 Smoking Prevalence-Williamson et al The Cost-effectiveness of Counseling Smokers to Quit Steven R.Cummings,MD;Susan M.Rubin,MPH;Gerry Oster,PhD Cigarette smoking is the most important preventable cause of death in the While the cost-effectiveness of a num- United States. Surveys of patients, however,suggest that many physicians do ber of preventive practices has been not routinely counsel smokers to quit. Because physicians may not consider studied, there have been no studies of counseling against smoking to be as worthwhile as other medical practices,we the cost-effectiveness of physician coun- seling against smoking. Is time spent examined its cost-effectiveness.We based our estimates of the effectiveness of during a routine office visit to counsel physician counseling on published reports of randomized trials and our esti- smokers to quit as worthwhile as other mates of its cost on average charges for physician office visits. Our results commonly accepted preventive prac- indicate that the cost-effectiveness of brief advice during routine office visits tices?If so,would a follow-up visit also ranges from$705 to$988 per year of life saved for men and from$1204 to$2058 be worthwhile? for women.Follow-up visits about smoking appear to be similarly cost-effective. METHODS Physician counseling against smoking,therefore,is at least as cost-effective as several other preventive medical practices and should be a routine part of health We herein analyze the cost-effective- care for patients who smoke. ness of physician counseling to quit (JAMA 1989;261:75-79) smoking during a routine office visit. For convenience, we considered the cost-effectiveness of this intervention in a hypothetical group of patients who are CIGARETTE smoking is responsible Follow-up visits with a physician may smokers and are seen during routine for more than 270 000 avoidable prema- also be important in helping patients office visits. We assumed that all pa- ture deaths in the United States every quit and remain abstinent.""The Ca- tients would be counseled,but that only year.' Despite the hazards of cigarette nadian Task Force on the Periodic a fraction of these patients would actual- smoking,33%of men and 28%of women Health Examination,18 for example,has ly quit smoking as a result. continue to smoke.' Controlled trials recommended "scheduling follow-up We estimated the cost of physician suggest that physician counseling can visits to enhance the effect of anti-smok- counseling based on the amount of time influence smokers to quit."Since about ing counselling."More than two thirds that would be spent advising smokers to 38 million smokers in the United States of internists, however,report that they quit and on the cost of self-help materi- visit a physician each year,10 routine never schedule appointments with pa- als about smoking cessation that may be counseling against smoking by all physi- tients primarily to counsel them to quit given to the patient. We measured the cians could potentially help a large num- smoking.19 effectiveness of counseling in terms of ber of smokers to stop. However,while Thus, although physicians consider the difference between the number of most physicians believe that advising smoking to be an important problem, patients who would quit smoking if giv- smokers to quit is important,"-'ba recent counseling smokers to quit is not yet a en such advice and the number who survey"reported that 56% of smokers routine part of medical practice and would quit smoking on their own. We who had visited a physician during the many physicians seem reluctant to de- measured the benefits of intervention in previous year had never been advised to vote time to such counseling. This may terms of the added years of life expec- quit smoking. be due, in part, to the fact that third- tancy experienced by those additional party payers generally do not reim- patients who quit smoking. We ex- burse providers for time spent coun- pressed the cost-effectiveness of coun- From the Division of General Internal Medicine,De- seling patients to quit smoking."' seling in terms of the cost per year of life t of Medicine(Dr Cummings and Ms Rubin and the Clinical Epidemiology Program (Dr Cum-- Furthermore,since only a small propor- saved among all smokers who are coun- and the mings), University of California, San Francisco; and tion of smokers quit as a result of physi- seled, not just those who succeed in Policy Analysis Inc,Brookline,Mass(Dr Oster). cians' advice, physicians may wonder quitting smoking. Reprint requests University California Medical Center,400 Parnassus Ave.A-405,,San Francisco,CA whether counseling smokers is as We assumed a societal perspective in 94143-0320(Dr Cummings). worthwhile as other medical practices. our analysis. We evaluated the net so- JAMA,Jan 6,1989—Vol 261,No.1 Cost-effectiveness of Counseling Smokers—Cummings et al 75 1 10 cial costs of physicians' advice against office visit, we searched the English- Gains In Life Expectancy After its net effectiveness,measured in terms language medical literature (MED- Quitting Smoking of additional years of patient life expec- LINE)and bibliographies of recent re- Tb estimate the gains in life expectan- tancy.Therefore,the cost-effectiveness views""' for randomized trials that cy among those who quit smoking and of this intervention as perceived by any compared rates of smoking cessation . remain abstinent,we used published es- particular patient or payer may differ among patients who were given advice timates of the increase in life expectan- from our estimates. by a physician to quit smoking and those cy for men and women between the ages Tb permit comparison of our findings who received no counseling. We ex- of 35 and 69 years at time of quitting,' to those reported for other preventive eluded studies that did not use random which were based on data from the practices,we expressed our results and assignment of subjects to intervention American Cancer Society's 25-state those of comparable studies in terms of and control groups, as well as studies Cancer Prevention Study. These esti- 1984 dollars. that followed up subjects for less than mates are similar to published esti- Costs of Physician Counseling one year. We identified a total of four mates based on data from the Framing- Costs g studies that met our inclusion criteria, ham study.2B We assumed that delivering brief ad- three of which biochemically validated vice to quit smoking would take four self-reports of smoking cessation.'The Discounting minutes during a routine office visit,the one study that did not biochemically val- Estimates of the gains in life expec- equivalent of about one third the length idate self-reports of cessation reported tancy were discounted;the rationale for of an office visit with an established pa- no significant difference between pa- this has been discussed elsewhere." tient (Table 1)."According to a recent tients who received advice and those An annual rate of 5%was used through- national hrou h national survey of physicians," the who did note Tb estimate the effective- out the analysis, although discount average charge of an internist for an ness of physician advice,we used stan- office visit with an established patient dard techniques'to pool the cessation rates of 3%and 7%also were used in our was $30. We also assumed that a self- rates reported in these studies (Table sensitivity analysis. help booklet would be given to all pa- 2).We calculated that brief advice from Other Medical and Nonmedical Costs tients who received brief advice at a cost a physician increased the cessation rate .We did not consider the use of nico- of$2 per booklet.&e The estimated total at one year by 2.7%. Using the 95% tine gum since the cost-effectiveness of cost of brief advice was therefore $12 confidence limits of this estimate (ie,[t/s x$30)+$2).Because the costs of (±1.7% we alternatively assumed in- this intervention as s adjunct to been ex- office visits vary, we examined the creases in the cessation rate of 1.0%and cian advice and counseling has been in- sensitivity of our results in relation to 4.4%in our sensitivity analyses. elude previously.f We also did not in- sensitivity changes in the assumed cost of elude a number of potential indirect intervention. Relapse to Smoking costs to patients, such as the value of time spent in travel or office visits, be- Effectiveness of Brief Differences in rates of smoking cessa- cause patients were assumed to be coun- Physician Counseling tion between treatment and control seled during a routine office visit.While Tb estimate a rate of smoking cessa- groups at one year may be larger than quitting smoking may decrease the tion for patients receiving physicians' differences in quit rates that would be costs of medical care for smoking-re- advice about smoking during a routine found after a longer period of observa- fated illnesses,it may also increase costs tion. For example, smokers who quit of routine medical care during addition- as a result of physicians' advice might al years of life. Since previous stud- eventually relapse at a rate somewhat iesn,er,' have reported that these costs higher than that among patients who probably offset each other, they also Table 1.—Baseline Assumptions Used in Analyses quit on their own. We are unaware of were not included in our analysis. reliable estimates of smoking relapse Cost office visit vis among medical patients who have ab- Incremental Cost-effectiveness of a Time devoted to brief advice 1/3 of visit stained for one year. Nevertheless, we Follow-up Visit Time devoted to follow-up visit Full visit assumed that 10% of patients who had Since a follow-up visit may increase Cost of self-help booklet abstained or one year as a result of the likelihood that a patient quits smok- (with brieffbif advice only) $2 Y Additional 1-y cessation physician counseling would eventually ing,but would also increase the costs of due to brief advicecessa 2.7% relapse and that the would none of intervention we considered its incre- Addnionall-ycessation I> y g� r due to follow-up visit See text the health benefits of smoking cessa- mental cost-effectiveness,expressed as Additional relapse after 1 y tion.In our sensitivity analyses,we also the ratio of the marginal cost of provid- among those given advice to quit 10°/0 Discount rate 5% examined the effect of assuming a 50% ing patients with an additional visit to rate of relapse. its marginal effectiveness. The latter Table 2.—Randomized Trials of Brief Advice to Quit Smoking No.of Patients Biochemical Net cessation Rate,% source y Intervention' Advice Control Follow-up,y Validetiont (Counseled-Control) Jamrozik at 81,31984 Advice,leaflet,warned of folkrw-up 512 549 1 + 4.4# Russell et al,41979 Advice(1-2 min),leaflet,warned of follow-up 408 340 1 + 4.8# Russell et al 51983 Advice(1-2 min),booklet,warned of follow-up 675 584 1 + 0.2 Stewart and Rosser,31982 Advice,questionnaire,pamphlet 94 128 1 – 1.2 'Controls received no advice. tPlus indicates yes;and minus,no. *P<.05. 76 JAMA,Jan 6,1989—Vol 261,No.1 Cost-effectiveness of Counseling Smokers—Cummings et all Table 3.—Randomized Trials of Follow-up Visits Minimum Biochemical` No.of Additional Net Cessation Rate,% Average Cessation Rate per Source,y Intervention Follow-up Validation Visits (Counseled-Control) Additional Visit,% Fagerstrom,71964 Intervention(n=22) Initial advice and telephone 1 y + 1 12 12 call at 1 wk;follow-up visits at 2 and 4 wk; letter at 3 mo;leaflet Control(n=27) Initial advice;follow-up ... ... ... ... ... visit at 2 wk after cessation;leaflet Richmond and Webster,81865 intervention(n=100) Counseling,questionnaire, 6 mo + 4 30 7.5 diary;follow-up visits at 1,2,3,and 5 wk and ata and 6 mo Control(n=100) 2 visits(1 at entry and at ... ... ... ... ... 6 mo) Wilson at al,a 1982 Intervention(n=106) 3-5-min counseling 6 mo - 3 lit 3.7 session and leaflet; follow-up visits at 1,3, and 6 mo Control(n=105) 3.5-min counseling ... ... ... ... ... session and leaflet 'Plus indicates yes;and minus,no. tP<.05. was expressed as the difference be- Table 4.—Cost-effectiveness of Brief Advice About Quitting Smoking(Dollars per Year of Life Saved) tween the number of patients who would quit smoldng if given brief advice Effectiveness of Advice and a follow-up visit, and the number 1%Cessation Rate, who would quit if given brief advice sex and Intervention Intervention alone. Age Group Basellne' 1% 4.4% Cost,$17 cost,$19, ]G'or patients receiving follow-up ad- Men 60%Relapse vice, we assumed that an entire visit 35-38 988 2667 606 1399 7600 would be devoted to smoking cessation 40-44 837 2260 514 1166 64x1 counseling. Since the cost of self-help 4549 748 2020 459 1060 5758 materials was included in the cost of 50.54 705 1905 433 999 5429 providing initial brief advice, we as- 55.59 726._ 1961 446 1029 5568 sumed that no additional materials 60-64 810 2186 497 1147 6230 would be given to patients as part of a 85.69 950 2564 563 1346 7309 follow-up visit. We therefore assumed Women that the marginal cost of a follow-up 40-39 2056 1283 2515 15633 visit would be$30. 1646 4466 1010 2332 12667 To estimate the marginal effective- 4549 1372 3704 eat 1943 10556 1235 3333 752 1749 9500 50-59 ness of a follow-up visit, we searched 55.54 1204 33W 739 1706 9268 the English-language medical literature 65.59 1266 3219 739 1796 9768 (MEDLINE) and bibliographies of re- cent reviews'lo,r7,w for randomized trials 65.69 1411 3810 866 1999 10857 that compared rates of smoking cessa- *Baseline assumptions used in analysis;one third of office visit devoted to brief advice;cost of office visit,$3o;cost tion after one year among patients in of self-help booklet,$2;total cost of intervention,$12((1!3 x$301+$2);additional one-year cessation due to brief general medical practice who were giv- advice,2.71,6;relapse rate after one year among those given advice to quit,100%;and discount rate,5%. en physicians' advice to quit smoking and those who received physicians'ad- studies," patients in the intervention quitting.We used the same estimates of vice and at least one follow-up visit.Al- groups received several more follow-up increased life expectancy for those who though we found three randomized tri- visits than those in the control group. quit smoking' and discounted future als of the effectiveness of follow-up Since none of the identified studies survival by 5%per year. visits(Table 3),we found no study that met our criteria for inclusion of trials, met all of our criteria. Two of the stud- we analyzed the cost-effectiveness of RESULTS ies"' compared patients given advice follow-up visits across a range of rates of est-effectiveness of Brief Advice with patients given advice and follow- smoking cessation that encompassed up visits, but followed up patients for those reported in the controlled trials Baseline Assumptions.—Brief ad- six months only. The remaining study (1%to 12%).'As with brief advice,we vice and counseling by a physician dur- followed up patients for one year,'but assumed that 10%of those who quit as a ing a routine office visit about quitting used a control group that also received a result of a follow-up visit would relapse smoking costs from$705 to$988 for men follow-up visit. Also, in two of these and gain none of the health benefits of per year of life saved and from$1204 to JAMA,Jan 6,1989—Vol 261,No.1 Cost-effectiveness of Counseling Smokers—Cummings et al 77 Table 5.—Comparison of Cost-effectiveness for Common Preventive Practices 20 000 Cost-effectiveness Practice (Dollars/Year of Lffe Saved)* 1 Brief advice about smoking 1 during routine office visit 15000 2.7%additional cessation 748 t 1%additional cessation 2020 1 Follow-up visit devoted to counseling n 1 about quitting smoking(marginal ° 10000 1 cost-effectiveness) io 1 1%additional cessation 5051 m 1 Nicotine gum added to \` physicians'advioe27 4113 0 5000 �` Treating moderate hypertension; diastolic blood pressure>11033 11 300 Treating mild hypertension; diastolic blood pressure 90-1103' 24 408 0 Treating hypercholesterolemia 2 4 6 B 10 12 (;8.85 mmol/L)with cholestyramine3' 65 511-108189 Average Cessation Rate(%)per Follow-up Visit *Ali analyses based on 45-to 50-year-old men,and a 5%discount rate.All costs adjusted to 1984 US dollars. Cost-effectiveness of follow-up visit about smoking Incremental Cost-effectiveness of a less effective than we assumed it to be at varying rates of smoking cessation for 45-to 49- year-old men(solid line)and women(dashed line). Follow-up Visit (eg,if only 1%of smokers would quit as a Average rates of cessation r follow-upvisit for result),it would remain more cost-effec- three andomizedtrials are 3.7%o,°7.5%,° d 12%.' The incremental cost-effectiveness of tive than treating mild hypertension or Corresponding cost per added year of life expectan- a follow-up visit about smoking cessa- hypercholesterolemia (Table 5). We cy for menAvomen would be $1365/$2503, tion for men and women 45 to 49 years of by also have overestimated the cost $673/$1235,and$421/$772. age is presented in the Figure for mar- y brief advice. In our baseline analysis, ginal rates of smoking cessation ranging we assumed that advice to quit smoking from 1%to 12%. Among men, cost-ef- would take four minutes, while two of $2058 for women, depending on a pa- fectiveness ratios range from$421 (as- the trials we used in estimating the ef- tient's age(Table 4). Cost-effectiveness suming a 12 percentage point increase in fectiveness of advice involved only one is generally higher for men, partially the cessation rate)to$5051(assuming a to two minutes of advice. reflecting heavier smoking among men one percentage point increase only). A recent analysis by Kottke et al" than women and, hence, greater bene- Costs per year of life saved are some- suggests that multiple "reinforcing fits of cessation. what higher among women, increasing contacts" with smokers are an impor- Sensitivity Analyses.—We exam- from$772 to$9259 as the marginal rate tant part of helping medical patients ined the sensitivity of our results in rela- of cessation declines from 12%to 1%. quit smoking, and the Canadian Task tion to changes in a number of key as- We also examined the combined ef- Force on the Periodic Health Examina- sumptions and parameters(Table 4)as fect of several pessimistic assumptions. tion18 recommends follow-up visits with follows:(1)Assuminga 50%increase in Assuming a cost of$45, a one percent- the cost of an office vsit(ie,$17 for brief age point gain in smoking cessation,and smokers to counsel them about quitting. a 50%relapse rate costs per of life We were unable,however,to find aran- advice and materials),costs per year of p Pe y domized trial of follow-up visits com- life saved increase by about 40%.A 50% saved range from$12 857 to$18 000 for pared with brief advice alone that fol- decrease in visit cost causes cost-effec- men and from $21951 to $37 500 for lowed up patients for at least one year. tiveness ratios to decrease by approxi- women. While the three randomized trials that mately 40%. (2) Assuming that advice COMMENT we reviewed suggest that such visits and counseling results in only a one per- may substantially increase the rate of centage point increase in the rate of ces- Our results suggest that physician smoking cessation, we note that these sation, costs per year of life saved in- counseling about smoking during rou- studies may have overestimated the ef- crease by 170%. Alternatively, if the tine office visits is at least as cost-effec- fectiveness of follow-up visits. Two cessation rate increases by 4.4 percent- tive as a number of accepted medical studies followed up patients for only six age points,cost-effectiveness ratios de- practices,such as the treatment of mild months,and differences in cessation be- cline by approximately 40%.(3)At a 3% to moderate hypertension or hypercho- tween the intervention and control rate of discount, costs per year of life lesterolemia (Table 5).',' The cost-ef- groups may have diminished during a saved decrease by between 21% and fectiveness of counseling also appears to longer period of time." The study by 46%,depending on patient age and sex. be similar to that of using nicotine Fagerstrom'involved only a small num- At a 7% rate of discount, these costs gum.' ber of patients, and, consequently, the increase by between 27%and 71%. We note that we might have underes- higher rate of smoking cessation among We combined several of our most pes- timated the effectiveness of brief ad- those who received additional follow-up simistic assumptions: $45 for an office vice.Our assumption that only an addi- visits was not statistically significant. visit, $4 for a self-help booklet, a 1% tional 2.7%of smokers would quit as a Even using pessimistic assumptions cessation rate for brief advice, and a result of brief advice is lower than the about the marginal effectiveness of a 50%relapse rate among those who re- 4%to 10%rate cited by other review- follow-up visit, however, it appears to ceive advice after one year. At a 5% ers.10,",°' Our estimate may be lower be as cost-effective as a number of other discount rate, costs per year of life because we only included studies that accepted preventive practices(Table 5). saved increase by about 670%, ranging were randomized and reported cessa- We also note that the true marginal from$5429 to$7600 for men and from tion rates after one-year follow-up. cost of a follow-up visit may differ from $9268 to$15 833 for women(Table 4). Nevertheless,even if brief advice were that which we assumed. On the one 78 JAMA,Jan 6,1989—Vol 261,No.1 Cost-effectiveness of Counseling Smokers—Cummings et al I hand, while we assumed that an entire reviewed was simple and can be pro- need,however,for more research about office visit would be devoted to counsel- vided by all physicians without special how much such visits add to the effec- ing, in practice it could be combined training. It usually included a message tiveness of brief advice. with the care of other medical problems. that smoking was detrimental to health Despite the health hazards of smok- We also assumed that follow-up visits and firm advice that the patient should ing and the potential value of physician would be with a physician,although vis- quit smoking. Special training in smok- advice,many physicians do not routine- its with other health professionals ing cessation might increase physicians' ly counsel smokers to quit smoking. might be less costly and equally effec- effectiveness,but whether training pro- Some physicians may consider counsel- tive. These possibilities deserve addi- grams increase the cost-effectiveness of ing about smoking to be less worthwhile tional study. On the other hand, we did physicians'advice depends on their cost than other medical practices;some may not include a variety of costs to patients, and the degree to which they increase feel frustrated because only a few of the such as the value of time spent in travel cessation rates among patients. smokers whom they counsel quit smok- to and from medical facilities or during We note that brief advice about smok- ing. Our results suggest that even if visits themselves. ing does not require specific coverage physicians' advice helps only 1% of Nicotine gum might enhance the ef- by third-party payers because it can be smokers to quit,it is as cost-effective as fectiveness of physician advice about incorporated into the routine medical many other accepted medical practices. quitting smoking.While we did not con- care of smokers. Follow-up visits de- Physician counseling about smoking sider the use of nicotine gum as an ad- voted entirely to counseling smokers to should be a routine part of the health junct to physician advice and counsel- quit, however, would require such cov- care of smokers. ing, we note that Oster et al'reported erage. Most third-party payers do not that a prescription for nicotine gum was pay for physician visits for smoking ces- likely to increase the cessation rate,rel- sation and state laws that regulate This study was supported by grant CA38374A alive to advice and counseling alone,by health insurers do not mandate that from the National Cancer Institute,Bethesda,Md, points, and the cost-ef- the and the Henry J.Kaiser Family Foundation,Menlo 1.6 percentage P $ p y pay for such visits. Our analysis Park,Calif;Faculty Fellowship in General Internaall fectiveness of nicotine gum therapy suggests that additional physician visits Medicine(Dr Cummings). compared favorably with other widely devoted to helping patients quit smok- we appreciate the critical suggestions of Dr Da- vid Burns,PhD,the support of Drs Thomas Glynn, accepted medical practices(Table 5). ing may be cost-effective and may war- PhD, and Joseph Cullen, PhD, and the technical The advice provided in the trials we rant third-party payment. There is a assistance of Daniel M.Huse,MA. References 1. Rice DP,Hodgson TA,Sinsheimer P,et al:The primary-care practitioners. N Engl J Med Center for Health Statistics,Jan 23,1987. economic costs of the health effects of smoking, 1983;308:97-100. 23. Center for Health Policy Research:Socioeco- 1984.Milbank Q 1986;64:489-547. 13. ASurvey of Physicians,Dentists,Nurses,and nomic Characteristics of Medical Practice 1985. 2. National Center for Health Statistics:Health, Pharmacists:Their Behavior and Attitudes Con- Chicago,American Medical Association,1985. United States,1986, Dept of Health and Human cerning 7bbacco,Dept of Health,Education,and 24. Pederson LL:Compliance with physician ad- Services publication(PHS)87-1232.Public Health Welfare, Atlanta, National Clearinghouse for vice to quit smoking:A review of the literature. Service,December 1986. Smoking and Health,Centers for Disease Control, Prev Med 1982;11:71-84. 3. Jamrozik K, Vessey M, Fowler G, et al: Con- 1975. 25. L'Abbe KA, Detsky AS, O'Rourke K:Meta- trolled trial of three different antismoking 14. Coe RM,Brehm HP:Smoking habits of physi- analysis in clinical research. Ann Intern Med interventions in general practice. Br Med J cians and preventive care practices. HSMHA 1987;107:224-233. 1984;288:1499-1502. Health Rep 1971;86:217-221. 26. Cummings KM,Jaen CR,Giovino G:Circum- 4. Russell MAH,Wilson C,Taylor C,et al:Effect 15. Fortman SP,Sallis JF,Magnus PM,et al:Atti- stances surrounding relapse in a group of recent of general practitioners'advice against smoking.Br tudes and practices of physicians regarding hyper- exsmokers.Prev Med 1985;14:195-202. Med J 1979;2:231-235, tension and smoking:The Stanford Five City Proj- 27. Oster G, Huse DM, Delea TE, et al: Cost- 5. Russell MAH,Merriman R,Stapleton J,et al: ect.Prev Med 1985;14:70-80. effectiveness of nicotine gum as an adjunct to physi- Effect of nicotine chewing gum as an adjunct to 16. Anda RF, Remington PL, Sienko DG, et al: cian'. advice against cigarette smoking. JAMA general practitioners'advice against smoking. Br Are physicians advising smokers to quit?The pa- 1986;256:1315-1318. Med J 1983;287:1782-1785. tient's perspective.JAMA 1987;257:1916-1919. 28. Taylor WC,Pass TM,Shepard DS,et al:Cho- 6. Stewart PJ,Rosser WW:The impact of routine 17. Kottke TE,Battista RN,DeFriese GH,et al: lesterol reduction and life expectancy.Ann Intern advice on smoking cessation from family physi- Attributes of successful smoking cessation inter- Med 1987;106:605-614. cians.Can Med Assoc J 1982;126:1051-1054. ventions in medical practice:A meta-analysis of 39 29. Warner K,Luce B:Cost-benefit and Cost-effec- 7. Fagerstrom KO: Effects of nicotine chewing controlled trials.JAMA 1988;259:2882-2889. tiveness Analysis in Health Care: Principles, gum and follow-up appointments in physician- 18. Canadian Task Force on the Periodic Health Practice,and Potential.Ann Arbor,Mich,Health based smoking cessation. Prev Med 1984;13:517- Examination:The periodic health examination:II. Administration Press,1982. 527. 1986 update.Can Med Assoc J 1986;134:724-727. 30. Weinstein M, Stason WE: Hypertension: A 8. Richmond RL,Webster IW:A smoking cessa- 19. Cummings SR, Stein MJ, Hansen B, et al: Iblicy Perspective. Cambridge, Mass, Harvard tion programme for use in general practice.Med J Smoking counseling and preventive medicine: A University Press,1976. Aust 1985;142:190-194. survey of internists in private practice and an 31. Oster G,Colditz GA,Kelly NL:The economic 9. Wilson D,Wood G,Johnston N,et al:Random- HMO.Arch Intern Med,in press. costs of smoking and benefits of quitting for individ- ized clinical trial of supportive follow-up for ciga- 20. Weshsler H,Levine S,Idelson RK,et al:The ual smokers.Peet,Med 1984;13:377-389. rette smokers in a family practice.Can Med Assoc J physician's role in health promotion:A survey of 32. Warner KE:Health and economic implications 1982;126:127-129. primary-care practitioners. N Engl J Med of a tobacco-free society. JAMA 1987;258:2080- 10. Ockene JK:Physician-delivered interventions 1983;308:97-100. 2086. for smoking cessation:Strategies for increasing ef- 21. Ockene JK, Aney J, Goldberg RJ, et al: A 33. Weinstein MC, Fineberg HV: Clinical Deci- fectiveness.Prev Med 1987;16:723-737. survey of Massachusetts physicians'smokinginter- sion Analysis. Philadelphia, WB Saunders Co, 11. Cummings KM,Giovino G,Emont SL,et al: vention practices.Am J Prev Med 1988;4:14-20. 1980. Factors influencing success in counseling patients 22. McLemore T,DeLozier J(eds):1985 Summa- 34. Oster G, Epstein AM: Cost-effectiveness of to stop smoking. Patient Ed Counsel 1986;8:189- ry:National Ambulatory Medical Care Survey antihyperlipemic therapy in the prevention of coro- 200• Advance Data From Vital and Health Statistics, nary heart disease.JAMA 1987;258:2381-2387. 12. Wechsler H,Levine S,Idelson RK,et al:The No.128,Dept of Health and Human Services pubb- physician's role in health promotion:A survey of cation (PHS) 87-1250, Hyattsville, Md, National JAMA,Jan 6,1989-Vol 261,No.1 Cost-effectiveness of Counseling Smokers-Cummings et al 79 ReducingIllegalthe Sale of Cigarettes to Minors . David G.Altman,PhD;Valodi Foster,MPH;Lolly Rasenick-Douss;Joe B.Tye,MBA This study reports on an effort to stop the illegal sale of cigarettes to minors. In stop the illegal sale of tobacco to minors. Santa Clara County,Calif,412 stores and 30 vending machines were visited by In Avon, England, four minors under 18 minors aged 14 through 16 years with the intent to purchase cigarettes;they the age of 16 years were able to pur- were successful at 74%of the stores and 100%of the vending machines.After chase tobacco from 91 of 100 tobacco- an aggressive six-month campaign using communitywide media, direct mer- nists. One year later,after a press con- chant education, contact with the chief executive officers of chain stores and ference was held and a letter was sent to franchise operations owned b major companies, and grassroots work with merchants clarifying the law,minors re- franchise Y J P 9 turned to 50 of the stores and were able community organizations,the percentage of stores with illegal over-the-counter to purchase tobacco in only 44% of sale of cigarettes to minors was reduced to 39%.Sales from vending machines them.`Similarly,the Decatur,Ill,chap- were not reduced. While much remains to be accomplished in stopping the ter of Doctors Ought to Care found that illegal sale of tobacco to minors, data from this study illustrate that a well- a 14-year-old girl and an 11-year-old boy designed community and merchant education campaign can significantly re- were able to purchase cigarettes in 78% duce such sales. and 63%of stores,respectively.Letters (JAMA 1989;Ml:80-83) were then written to all of the store managers giving them feedback about their performance, a copy of the state law, and a letter from the chamber of THE SALE of tobacco products to mi- of tobacco to minors is,therefore,a crit- commerce urging them to adhere to the nors is a major public health problem. ical step in any effort to prevent tobacco law. This intervention reduced the We Although most states have laws regu- use.' of illegal cigarettes by 18V An unpub- lating the access of minors to tobacco, Easy access to tobacco is obviously a lished Canadian project called `Busi- they are rarely enforced."In California prerequisite to maintaining a tobacco ness for Kids"has also been conducted it is against the law to sell tobacco prod- addiction. The earlier a young person to stop the We of tobacco to minors(Bill ucts to anyone under age 18 years. By begins using tobacco,the less likely it is Howard, written communication,June all accounts,tobacco is readily accessi- that he or she will be able to quit later. 22,1988). ble to minors. In field trials, minors Over half of high school seniors (53%) The purpose of the current study was who smoke at least half a pack per day to determine whether a voluntary mer- have made at least one serious but un- chant education program combined For editorial comment see p 99. successful attempt to quit smoking,47% with a media campaign and grassroots say they would like to quit at the current community organization would be effec- time, and almost three fourths of daily tive in reducing the sale of cigarettes to have successfully purchased tobacco smokers in high school still smoke seven minors in a wide range of stores and from stores and vending machines 709o' to nine years later,even though in high communities. to 100%of the time.'It is alarming that school only 5% thought they would be METHODS minors have little difficulty obtaining daily smokers five years later.' tobacco, since most smokers begin Peer pressure to smoke coupled with In January 1988, 412 stores in Santa smoking regularly as teenagers. Na- societal influences such.as tobacco com- Clara County,Calif,that sold cigarettes tional survey data indicate that 57%of pany advertising that glamorizes smok- over the counter and 30 outlets that had high school seniors who report daily ing'influence young people to perceive cigarette vending machines were visit- smoking began by the age of 14 years smoking in a positive light. Unfortu- ed by 18 minors aged 14 (N=9), 15 (ie,the ninth grade).'Stopping the sale nately, our society has done very little (N=5),and 16(N=4)years.Half of the to counteract this by decreasing the minors were girls. No attempt was From the Center for Research in Disease Prevention, availability of tobacco through enforce- made to recruit minors who looked older Stanford University School of Medicine(Dr Altman), ment of existing laws and the enactment than their chronological age. Stores and the Center for the Study of Families,Children and of meaningful penalties for violation of were located in the socioeconomically Youth,Stanford University oci Raseneasaouss),Palo these laws. If minors have a difficult diverse towns of Cupertino, Mil ltas Alto,Cafif;HeatthWorks Associates,Pleasant Hill,Calif l� p s (Ms Foster);and the Baystate Medical Center,Spring- time obtaining tobacco, they may be Mountain View, Palo Alto, San Jose, field,Mass(Mr Tye). prevented from experimenting with Santa Clara,and Sunnyvale. Reprint requests to Center for Research in Disease and later becoming addicted to it. The stores visited included Prevention, Stanford University School of Medicine, g grocery 1000 Welch Rd,Palo Alto,CA W04-1 885(Dr Altman). There have been a few attempts to stores(N=89), liquor stores(N=103), 80 JAMA,Jan 6,1989—Vol 261,No. 1 Illegal Sale of Cigarettes—Altman et al Table 1.—Variables Measured During Interaction Table 2.—Cigarette Sales to Minors by Type of Store With Store Merchants %of Stores Where Variable Responses Minors Bought Sale made Yes or no No.of Cigarettes Merchant asks age of minor Yes or no Store Type Stores First Test Second Test P. %Changet Merchant asks who the Yes or no Vending machine 30 100 100 cigarettes are for 9 NS 0 Warning sign posted about Yes or no Gas station 50 84 55 .001 —35 tobacco sales to minors s as required by state law Liquor store 103 83 46 <.0001 —45 Sex of merchant M or F Convenience store 103 81 48 <,0001 —41 Age of merchant as <30 y or>30 y Pharmacy 67 60 28 .001 —53 estimated by the minor Grocery store 89 60 22 <,0001 —63 . Total(excluding vending machines) 412 74 39 <.0001 —47 *NS indicates not significant. convenience stores (N=103), pharma- t(°%in first test—%in second test)/%in first test;the percentage changes between the first and second tests did cies(N=67), gas stations IN=50), and not differ significantly among store types. businesses that had vending machines (N=30). Stores that sold cigarettes over the counter included company- tobacco sales posted on the front door or by small media,such as service club and owned chain operations (43%), fran- window or at the cash register, and chamber of commerce newsletters. chise operations (27%), and indepen- asked for a package of Marlboro ciga- The community education campaign dent stores(30%). Chain and franchise rettes. The minors were instructed to included distribution of materials de- operations were preselected from local be honest if the merchant asked them scribing the project as well as numerous phone books,while independent stores any questions. For example,if the mer- presentations to local community were selected by the minor and an adult chant asked the minor's age,the minor groups. As a result of these contacts, escort when the chain and franchise op- was instructed to respond truthfully many community leaders wrote letters erations were visited (with the excep- (eg, 14, 15,or 16 years old). If the mer- of support, and the Kiwanis club, par- tion of 24 chain stores and one franchise chant asked who the cigarettes were for ent-teacher association, and county store that were selected by the minor the minor was told to respond"For me." board of supervisors passed formal res- and adult). Most of the independent If a sale was made, the minor left the olutions in support of the project.Com- stores selected were in proximity to the store with the package of cigarettes and munity organizations seemed to be at- chain and franchise operations(eg,in a returned to the car.Data were collected tracted to the project because of its shopping center). Industry data(1985) on six characteristics of the interaction focus on youth, do the enforcement of on cigarette sales by retail outlets show with merchants(Table 1). existing laws, and on the use of educa- the following distribution: supermar- tional approaches to merchant contact. kets (30%, convenience stores (18%, intervention Components Direct Merchant Education.—Us- grocery stores (13%, gas stations From January through July 1988, ing a table of random numbers, all (10%, drug stores (9%), vending ma- combinations of three interventions stores were allocated randomly to one of chines (8%), discount stores (7% and were implemented:(1)community edu- three interventions:(1)no personal con- other(5%):9 cation, (2) direct education with mer- tact(eg,only community education),(2) Procedures and Measures chants who sold tobacco,and(3)contact mailed information,or(3)personal visit with the chief executive officers of the from project staff. The personal visit Minors attempted to purchase ciga- companies that owned the major chain (five to ten minutes)consisted of project rettes from stores in January 1988(be- and franchise operations visited. staff telling store managers about the fore intervention)and six months later Community Education. — Santa project and giving feedback as to wheth- in July 1988(after intervention).A peri- Clara County communities were made er they had sold cigarettes to minors. od of six months between tests was se- aware of the problem of tobacco avail- They were given a merchant education lected to allow businesses sufficient ability through mass media, small me- kit that included background facts and time to implement new tobacco sales dia, and presentations to community figures on tobacco and youth, tips for policies and to ensure that the communi- groups.A press conference announcing training employees not to sell tobacco to ty education campaign had been fully the results of the test before interven- minors,a copy of the state law that was implemented.At the second test,an ef- tion initiated the community education required to be posted, warning labels fort was made to send minors back to campaign. A 60 s television public ser- for cash registers and store windows the same stores they visited at the first vice announcement was taped for the (the labels stated"State Law. It is ille- test, recognizing that the store mer- project by US Surgeon General C. Ev- gal to sell cigarettes or chewing tobacco chant at the second visit might be differ- erett Koop, MD, and distributed to 13 to anyone under the age of 18."),a list of ent from the store merchant at the first television stations. Two 60 s radio ad- 34 project supporters,and a copy of the visit. Four of 18 minors were unable to vertisements produced by media expert resolution of commendation given to the participate in the second test;two were Zbny Schwartz (New York) were dis- project by the county board of supervi- out of town, one was unavailable, and tributed to 23 radio stations. The proj- sors. Stores in the mailed information one parent did not want her daughter to ect was covered by five television sta- group received the same merchant participate. Four stores went out of tions on local news segments, by five education kit but did not receive a per- business between tests. special segment news—life-style televi- sonal visit. Stores in the no personal Minors were driven to the stores by sion programs, by four local newspa- contact group received only community adults who remained in the cars and out pers, and by numerous radio stations, education. of view of merchants.The minor walked including a one-hour feature story by a Contact With Chief Executive Offi- into the store, determined whether local affiliate of the Public Broadcasting cers.—Certified letters were-sent to there was a visible warning sign about System. The project was also covered the chief executive officers of 24 compa- JAMA,Jan 6,1989—Vol 261,No.1 Illegal Sale of Cigarettes—Altman et al 81 10 Table 3.—Cigarette Sales to Minors by Chain,Franchise,and Independent Stores letter within a month were phoned. By %of scores where July 1988, 23 of 24 chief executive offi- Minors Bought Cigarettes cers or their designees had responded to Store No.of the letter, and all promised to take Category Stores First Test Second Test P. %Changet action. Chain 174 62 27 <.0001 -56 Franchise 110 81 50 <.0001 -38 Statistical Analysis Independent 121 85 47 <.0001 -45 The McNemar nonparametric test for *Chain stores also had significantly lower percentages of sales to minors than franchise and independent stores the significance of changes was used to In both the first and second tests(P<.001). analyze nominal data for two related t(%in first test-%in second test)/°/.in first test;the percentage changes between the first and second tests samples(eg, differences from first test did not differ significantly among store categories. to second test in same type of store). Analysis of variance was used to exam- Table 4.—Changes in Merchant Behavior ine differences across variables(eg,dif- ferences from first test to second test %(No.of stores) between store types). Variable First Test Second Test P. %Changet RESULTS Was warning sign posted as required by law? 3 (13) 23 (92) <.0001 +667 Was minor asked age? 24 (100) 48 (196) <.0001 +100 Cigarette sales to minors by type of Was minor asked who cigarettes were for? 5 (22) 6 (23) NS +20 store before and after intervention are Sex of merchant who sold tobacco to minor# Shown in Table 2. Sales to minors M 78 (185) 46 (96) <.0001 -41 dropped in all categories except for F 67 (115) 33 (64) `.0001 -51 vending machines. The overall rate for *NS indicates not significant. over-the-counter tobacco sales dropped t(%in first test-%in second test)/°%in first test. from 74%at baseline to 39%at the sec- $Male merchants were more likely to sell tobacco to minors than female merchants at both the first(P<.02)and and test. Vending machine sales were second(P<.o1)tests. not reduced from the preintervention level of 100%. Table 5.—Cigarette Sales to Minors by Age and Sex of Minor Stores were also categorized as com- pany-owned chains,franchises,or Inde- %(No.)of stores where pendent stores.As noted in Table 3,all Minors Bought Cigarettes Characteristic - showed decreases in sales to minors. of Minor First Test second Test P %Change* Although chain stores had significantly Age,yt lower sales than either franchises or in- 14 69 (134) 32 (62) <.0001 -54 dependent stores at baseline and at the 15 80 (92) 47 (53) <.0001 -41 second test (P<.001), the percentage 16 75 (76) 45 (45) <.0001 -40 differences between store categories sext M 71 (180) 35 (87) <.0001 -51 were not significant.Franchises and in- F 77 (122) 47 (73) <.0001 -39 dependent stores did not differ from each other in the percentage of sales at *(%in first test-%in second test)/%in first test;the percentage changes between the first and second tests did baseline or at the second test. not differ significantly among the three age groups or between boys and girls. There was an increase in the posting tSignificantly fewer 14-year-olds than 15-year-olds(P=.01)and 16-year-olds(P=.05)were able to purchase g signs,ofwarnin si s as required b Califor- cigarettes at the second test. y $Significantly more girls than boys were able to purchase cigarettes at the second test(P=.02). nia state law, and an increase in the frequency minors were asked their age (Table 4). Compared with stores that Table 6.—Changes in Warning Signs Posted by Type of Contact With Merchants did not post warning signs, stores that %of Stores With Warning did were less likely to sell cigarettes to Signs Posted minors at both the first (1% vs 99%, Type of Merchant P<.0001) and second (15% vs 85%, Contact first Test Second Test* Pt %Change* P<.005)tests.Although both male and No personal contact 5 7 NS +40 female merchants decreased their sales Mail 2 33 <.0001 +1555 Personal visit 3 2s `.0001 +730 of cigarettes to minors,male merchants were more likely to sell to minors than *At the second test,significantly fewer merchants with whom we made no personal contact posted warning signs female merchants(Table 4). than merchants contacted by mail(P<.0001)or by personal visit(P=.0002). As noted in Table 5, 14-, 15-, and 16- tNS indicates not significant. ear-olds were all less successful In pur- chasing in first test-%in second test)/°%in first test. Y p chasing cigarettes after the interven- tion. Furthermore, at the second test, ny-owned chain and franchise opera- to post notices informing the public of 14-year-olds were less successful in pur- tions with stores in Santa Clara County. this policy,as required by state law;(3) chasing cigarettes than either 15-or 16- The letter requested that the company implement employee training programs year-olds, although the percentage dif- take four actions to address the problem to ensure that employees have the ferences between age groups were not of illegal tobacco sales: (1)issue a com- knowledge and skills to comply with the statistically significant. Table 5 also il- panywide directive alerting store man- law;and(4)develop a monitoring mech- lustrates that female minors had a sig- agers and employees that the company anism to ensure compliance with corpo- nificantly higher percentage of pur- will not tolerate illegal tobacco sales to rate policy and state law. Chief execu- chases than male minors at the second minors; (2)require all company outlets tive officers who did not respond to this test,but the percentage differences be- 82 JAMA,Jan 6,1989—Vol 261,No. 1 Illegal Sale of Cigarettes—Altman et al tween male and female minors were not rettes illegally to minors. Clearly, the ing signs alerting customers(and store significant. norms in our communities about minors' employees)that tobacco sales will not be Stores that received an education kit access to and use of tobacco must made to minors. (6) Cigarette vending in the mail or a personal visit from pro- change. Society appears to accept early machines should be eliminated.If this is ject staff increased the posting of warn- tobacco use more easily than it accepts not politically feasible, alternatives ing signs significantly more than did early alcohol use. To decrease youth such as that recently adopted by King stores that did not receive a kit (Table smoking rates, people need to better County should be implemented. (7) A 6). There was no differential effect on understand that tobacco use among mi- national campaign should be mounted to sales as a function of the type of contact nors is illegal, existing laws regulating inform parents, health professionals, project staff had with merchants(ie,no tobacco sales must be strengthened, community organizations,and the busi- personal contact,mail contact,personal laws must be enforced, and, in states ness community about the problem and visit). without laws regulating access, laws to provide them with resources to re- COMMENT must be passed. These measures com- duce it. bined with effective community and If we are ever to achieve a tobacco- Data from this study corroborate merchant education will do much to re- free generation we must eliminate the data from other trials showing that mi- duce the availability of tobacco to youth. sale of tobacco to minors. While much nors around the country can readily pur- No single strategy will be completely remains to be accomplished in achieving chase cigarettes over the counter and effective in stopping tobacco sales to this goal,the findings of this study illus- from vending machines. The present minors. As described below, several trate that an aggressive community and study illustrates that an aggressive complementary actions could be taken. merchant education program can be merchant education program combined The model law proposed by DiFranza et effective. with community organizations applying ale would do much to reduce sales to pressure on merchants to change their minors. A law in unincorporated King Addis Thion s project 1bwas cco(STAT) and byothe Santa practices and a broad-based media cam- County, Wash (effective February Clara County Chapters of the American Cancer paign can reduce illegal over-the- 1989), should also serve as a model for Society,American Heart Association,and Ameri- counter cigarette sales to minors in all other communities (Michael Lippman, can Lung Association.It was supported by grants types of stores. Indeed, our data MD,oral communication,Oct 23, 1988). to STAT from the Henry J.Kaiser Family Founda- showed that, overall, over-the-counter This law will have the followin rovi- tion,Menlo Park,Calif;from the Cancer Research g p Foundation of America, Alexandria, Va;from El sales were reduced by almost 50%.The sions: (1) Vending machines must be - Camino Hospital,Mountain View,Calif;from the fact that minors were six months older electronically disabled until a store Kiwanis and Rotary clubs of Northern California; at the second test strengthens our find- clerk is certain the purchaser is of age. and from members of STAT. ings. These encouraging findings must (2)All tobacco merchants must have a We thank the minors and their parents,Thom- be balanced by what remains to be license(at a cost of$210),and the license J.Curran,MBB A,and the numerous community organizations that provided volunteers and support accomplished. fee will be used to pay for four blind for their contributions to the project.We also thank The best methods for working with inspections every year in which a super- Irwin Altman, PhD, Nathan Maccoby, PhD, and merchants remain unclear. We expect- vised minor will attempt to buy tobacco. manuscript.mare PhD, for their comments on the ed but did not find that mail or personal (3)The sale of individual cigarettes will Stop Teenage Addiction to Tobacco is developing visits to merchants combined with com- be banned. (4)The penalties for selling a nationwide campaign to eliminate the illegal sale munity education would be more effec- tobacco to minors will be $100 for the of tobacco to minors.For further information about this campaign and to receive the 7bbacco and tive in reducing sales than community first offense,$500 and a 90-day suspen- youth Reporter, contact: STAT, PO Box 60658, education alone. One explanation for sion of the tobacco license for the second Longmeadow,MA 01116. why we did not realize effects of our offense,and$1000 and suspension of the direct contact with merchants is that we license for nine to 18 months for the References may not have given stores, especially third offense. (5)Minors caught buying large chains and franchises, enough cigarettes can be subjected to communi- 1. Kirn TF:Laws ban minors tobacco purchases, time to implement the wide-ranging ty service and participation in a smoking but enforcement is another matter. JAMA 1987; 3324. changes we were promoting.In the case cessation program. 2. DiFranza JR,Norwood BD,Garner DW,et al: of personal visits, it may also be that Other steps that could be taken to Legislative efforts to protect children from tobac- five to ten minutes with store managers stop the sale of tobacco to minors are as co.JAMA 1987;257:3387-3389. is not sufficient.Alternatively,our con- follows:(1)We could encourage federal 3. Stanwick RS,Fish DG,ManfredaJ,et al:Where tact with chief executive officers of legislation vin states strop incen- Manitoba children obtain their cigarettes.Can Med � giving g Assoc J 1987;137:405-408. chains and franchises and the actions tives to establish a single minimum age 4. Naidoo J,Platts C:Smoking prevention in Bris- they subsequently took may have of 21 years for the sale of both alcohol tol:Getting maximum results using minimum re- washed out the effects of our direct con- and tobacco.This would allow for a uni- sources.Health Educ J 1985;44:39-42. 5. Minors easily buy tobacco. DOC News Views, tact with store managers. form enforcement mechanism to regu- Spring 1988,p 10. Our inability to reduce vending ma- late the sale of these products to minors. 6. Johnston LD,O'Malley PM,Bachman JG:Ma- chine sales suggests that elimination is (2) Businesses that sell tobacco, like tional Trends in Drug Use and Related Mors the only effective way to prevent minors businesses that sell alcohol, should be Among American High School Students and from purchasing cigarettes through this licensed, as will soon be req uired in Human Services publiYoung Adults,1975-1986US Dept of Health and caatt ion(ADM)87-1535. Na- medium. As noted in this and other unincorporated King County, Wash. tional Institute on Drug Abuse,1987. studies, however, vending machines (3) Fines for selling tobacco to minors 7. The Health Consequences of Smoking:Nicotine are not the only source of tobacco for must be increased substantially. (4) Ad of the Surgeon General. US Addiction, anHuman Services publication8S minors, since cigarettes can be pur- Stronger enforcement of tobacco access 8406,1988. chased illegally over the counter in all laws is critical. Income generated from S. Altman DG,Slater MD,Albright CA,et al:How types of stores. the licensing process and from fines for an unhealthy product is sold:Cigarette advertising Even after a massive community and illegal tobacco sales could pay for the in magazines, 1960-1985. J Commun 1987;37:95- 106. merchant education program, almost costs of enforcement. (5) Merchants 9. Cigarette sales by retail outlet.USDistribution four often stores continued to sell Giga- should be required to post visible Warn- J1988;115:12. JAMA,Jan 6,1989—Vol 261,No.1 Illegal Sale of Cigarettes—Altman et al 83 1 Recall and Eye TStudy of Adolescents Viewing Tobacco Advertisements Paul M.Fischer,MD;John W.Richards,Jr,MD;Earl J.Berman,MD;Dean M.Krugman,PhD The warning on tobacco advertisements was required by the federal govern- Minnesota Court of Appeals, a three- ment, presumably as a health message to educate the public about the risks judge panel unanimously ruled that the associated with tobacco use.Despite its potential public health role,there have federally mandated warnings do not been few published studies on the effectiveness of these warnings as a health protect tobacco companies from claims message.The present study used well-accepted market research methods to that they have failed to warn consumers examine adolescent viewing of tobacco advertisements.Sixty-one adolescents of the dangers of smoking. participated in the study. Eye tracking was used to study how participants It remains unclear whether the cur- viewed five different tobacco advertisements.The average viewing time of the rent warnings adequately warn the 9 9 public or are effective in discouraging warning amounted to only 8%of the total advertisement viewing time. In 43.6% the use of tobacco products.Research in of cases, the warning was not viewed at all. Following the advertisement the area of information disclosure has viewing,participants were asked to identify the observed warnings within a list delineated the difference between "in- that included other simulated warnings. Subjects did only slightly better than formation provision" and "information random guessing in this test of recognition. Using market research criteria,the impact.'"The latter is the central issue federally mandated warning must be viewed as an ineffective public health for whether warnings really warn, an message in so far as adolescents are concerned. issue for which even the Surgeon Gener- (JAMA 1989;261:89-89) al has expressed concern! The effectiveness of the warnings to adolescents is a special problem. This group continues to demonstrate high IN 1965,as a result of numerous scien- changed in 1984 when four rotating mes- rates of tobacco use.9 ' Furthermore, tific studies on the health hazards of sages were mandated by legislation, adolescents are the group from which smoking,legislation was passed requir- each describing a different health is- most new smokers are recruited.Great- ing that a warning be placed on all ciga- sue."In addition,a warning for smoke- er than 90%of smokers begin to smoke rette packages."' Threatened in many less tobacco advertisements and prod- as teenagers.11•12 states with legislation that would have uct packages was mandated in 1986.4 While tobacco manufacturers claim Because of the legislated wording of that they do not advertise to children or See also p 90. these various warnings, they are often adolescents(Time,April 9, 1984,p 91), referred to as the Surgeon General's there is considerable evidence that their required harsher warnings on adver- warnings. marketing is "youth oriented"20 Re- tisements,the tobacco industry worked It is the contention of the tobacco in- search has demonstrated that tobacco out a compromise in 1972 with the Fed- dustry that these warnings provide ade- advertising is well recognized by chil- eral Trade Commission (FTC) regard- quate information about the potential dren and adolescents."•"In addition, it ing the specific wording and the format hazards of tobacco use and that the man- has been demonstrated recently that of the warning.' The warning was ufacturers are therefore exempt from adolescent recognition of tobacco adver- product liability!Support for this legal tising correlates in a dose-response From the Department of Family Medicine, Medical position has come from cases argued manner with cigarette use. College of Georgia,Augusta(Drs Fischer,Richards, successfull b the tobacco indust be- This stud was conducted to examine and Berman);and the School of Journalism and Mass Y Y i'Y Y Communication, University of Georgia, Athens (or fore the US Court of Appeals in Geor- whether adolescents "read" and "re- Krugman). gia, Pennsylvania, and Massachusetts called"the Surgeon General's warning Reprint requests Department of Family Medical College oftGeorgia,Augusta,GA 30912'(or (Timer Sept 7, 1987,p 43).°However,in when viewing tobacco advertisements. Fischer). a more recent case argued before the It involved the recording of eye tracking 84 JAMA,Jan 6,1989—Vol 261,No,1 Tobacco Advertisements—Fischer et al ri F Fig 1.—Eye tracking apparatus in use. data during advertisement viewing, as Fig 2.—Recorded image of subject15 field-of-view.Time line and cursor(circled) are superimposed on well as standard postviewing recall videotaped image.(Simulated for publication purposes.) techniques used in market research.'` MATERIALS AND METHODS The study data were collected in July ence Laboratory,Waltham,Mass,mod- reading activity.`A 10%random sam- 1987. Participants were 61 paid adoles- el 3100H). This system obtains an eye pling of the recorded data was reana- cent volunteers (aged 13 to 17 years) image by using lightweight optical com- lyzed to measure the intraobserver reli- who were recruited from a YMCA mem- ponents mounted on a headband that is ability of the frame-by-frame analysis. bership list. Subject and parental worn by the subject (Fig 1). A coaxial This check of reliability showed that ad- written consent were obtained prior to camera and near-infrared light source vertisement total viewing times and to- enrollment in the study and after ex- within the optics produce a bright pupil tal fixations were within 0.1 s in 92%and plaining that the study involved mea- image and a corneal reflection.Also fit- 90%of cases,respectively. suring eye tracking while viewing ad- ted unobtrusively on the headband was Following the eye tracking measure- vertisements. Information about age, a field-of-view video camera.The head- ments, each subject was asked ques- sex, school grade, and smoking status band configuration permitted freedom tions about the advertisements they was obtained with a written question- of head motion and allowed a natural had seen using a masked recall tech- naire. The self-report smoking status reading posture. nique. This market research technique categories were similar to those that we The pupil and corneal reflection infor- is often utilized in situations where a low have previously used."The reliability of mation was videotaped.The signal con- amount of learning is thought to have the smoking status data was enhanced taining the field-of-view was simulta- taken place." The advertisements by employing the bogus pipeline tech- neously recorded. The recorded data were"masked"to cover all specific ref- nique.14•$25 were then computer processed off-line erences to the brand of cigarette, the Five advertisements (four cigarette to generate a data tape with a cursor, principal advertisement heading, and and one chewing tobacco)were selected representing eyepoint of regard,super- the Surgeon General's warning(Fig 4). from current magazines. The tobacco imposed on the field-of-view image(Fig The subjects were shown the masked brands chosen reflected the brands re- 2). This permitted detailed analysis of advertisements in the exact sequence ported by adolescents as most popular the viewing behavior(Fig 3). viewed during the corneal tracking in our previous study."The advertise- The tape was then analyzed frame by study period. Each was asked to recall ments were mounted in a three-ring frame to determine total advertisement what was covered by the masked areas binder to allow subjects to view them in viewing time,viewing time of the warn- of the advertisement. The responses a manner consistent with reading a ing, and the frequency and duration of were recorded verbatim and were later magazine.The viewing order for each of visual fixations within the warning scored for the pack, the heading, and the five magazine advertisements was area. Fixations are periods of relative the warning. As indicated in Table 1, randomized for each of the study sub- eye stability during which information credit was given for identifying the gen- jects. Participants were instructed to is taken in for processing.. These are eral wording of the heading or warning. view the pages in the same way that distinguished from rapid motions be- Finally,each subject was given a list they would view a magazine. Instruc- tween fixations(saccades)during which of ten Surgeon General's warnings tions were scripted to ensure uniformi- very little information is acquired. Sig- (Table 2). Five of these were the warn- ty. Participants were not told that they nifxcant fixations were operationally ings on the five advertisements that had would later be questioned about the ad- defined in this study as the absence of been viewed. Five other simulated vertisements.No time limit for viewing any eye movement greater than 10 of the warnings were also included. Each sub- was given. visual angle for at least three frames(ie, ject was asked to circle the five Eye tracking was measured using an 0.10 s). Fixations of this length have warnings that they remembered seeing Eye View Monitor system(Applied Sci- been reported to be characteri$tic of during the advertising viewing period. JAMA,Jan 6, 1989—Vol 261,No.1 Tobacco Advertisements—Fischer at al 85 r ,fw, W 71 ceme to 1 -.1boro Countar H ' JA0 a uro SURGEON GENERAI'SWARNING Srtakitp Now Gfeatly Reduces Serious Risks Health, Fig 3.—Eye movement(arrow)and significant fixation points(dots)of typical Fig 4.—'Masked"advertisement as used in recognition study.Pack,heading, advertisement viewing. and warning areas are covered. Table 1.—Masked Recall Scoring Table 2.—Surgeon General's Warnings Used in Aided Recall Score Real Warnings Masked Area 1 p 3 1. Surgeon General's Warning:Quitting smoking now Pack Unable to identify Identified as a pack Identified specific greatly reduces serious risks to your health. brand 2. Surgeon General's Waming:Smoking by pregnant Heading Unable to identity Identified as a heading Identified general women may result in fetal injury,premature birth,and low wording of heading birth weight. 3. Surgeon General's Waming:Cigarette smoke con- Waming Unable to identity Identified as a warning Identified general tains carbon monoxide. theme 4. Surgeon General's Waming:Smoking causes lung cancer,heart disease,emphysema,and may com- plicate pregnancy. 5. Waming:This product may cause gum disease and tooth loss. These data were used as a measure of showed that 30 (49%) had never Simulated W Warning: Ci Warnings 1. Surgeon General's Warning:Cigarerie smoke con- warning recognition. smoked, 22 (36%) had experimented tains arsenic. The data were analyzed using X'test, with cigarettes, and nine (15%) were 2. Surgeon General's Waming:Quitting smoking can help prolong your life. Kruskal-Wallis analysis of variance, regular smokers (more than one ciga- 3. Surgeon General's Waming: Smoking can cause and Spearman's correlational analysis. rette per week). wrinklingof yraduinykeyanossr older. 4.wg product you The Scheffe test for multiple compari- When asked about their favorite mag- 5.Waming:This product may cause mouth cancer. sons was used to analyze viewing time azine, the most frequent responses differences among advertisements.The were Sports Illustrated(43.6%),Made- data are reported as mean ±1 SD. moiselle (14.5%), Newsweek (9.1%), RESULTS and Cosmopolitan(7.3%). Of the 13 fa- The eye tracking tape for 16 of the vorite magazines mentioned, only two subjects was unacceptable for analysis The mean age for the 61 teenagers (Seventeen and Boys Life)do not carry because of technical problems (eg, participating in the study was 15±1.3 cigarette advertisements. These two drooping eyelid,inability to focus on the years. Sixty-four percent were male were reported to be the favorite maga- ocular reflexes). Two hundred two ad- and 95%were white. The self-reported zine by one female and two male sub- vertisements viewed by the remaining Smoking status of the participants jects,respectively. 45 subjects were available for analysis. 86 JAMA,Jan 6,1989—Vol 261,No.1 Tobacco Advertisements—Fischer etai .s.. Table 3.-Surgeon General's Warning Characteristics and Viewing Times Warning Time as Percent Total Advertisement Warning of Total Advertisement Location on Page Shape Wording of Warning Viewing Time,s Viewing Time,s Viewing Time Winston Left lower comer Rectangle "Cigarette smoke contains carbon 8.47±4.56 0.79±0.93 8.5 (n=44) monoxide." Marlboro Left lower comer Rectangle "Guitting smoking now greatly 7.12±3.29 0.55±0.87 7.7 (n=35) reduces serious risks to your health." Red Man Left upper comer Arrow-into-circle "This product may cause gum 8.21±4.42 0.62±1.2 7.5 (n=42) disease and tooth loss." Virginia Slims Left upper comer Rectangle "Smoking causes lung cancer,heart 11.76±7.74 0.55±1.02 4.7- (n=39) .7(n=39) disease,emphysema,and may oomplicte pregnancy." Camel Right upper comer Rectangle "Snaking by pregnant women may 9.11±6.72 1.25±1.50 13.7 (n=42) result in fetal injury,premature birth,and low birth weight." These 45 subjects included 22(49%)who Table 4-Masked Recognition Scores(N=61) had never smoked, 18 (40%) who had experimented with cigarettes, and five Score (11%)who wereregular smokers. Advertisement Pack Heading Warning The mean viewing time for a single Winston 1.8±0.65 1.1=0.32 1.9±0.54 advertisement was 8.95±5.75 s.Of this Marlboro 2.4±0.83 1.9±0.86 2.1±0.51 ± time,only 0.75±1.17 s was spent view- Red Man 2.40.82 1.2±0.44 1.4-0.76 ing the warning(ie, 8%of the total ad- Virginia Slims 2.3±0.93 1.6±0.88 2.3±0.70 vertisement viewing time). Camel 2.1±0.63 1.2±0.46 2.0±0.52 The differences between the mean al advertisements 2.2• 1.4• 1.9" viewing times for the five advertise- ments were not statistically significant .P<.001,F test for analysis of variance with Scheff6 test(multiple comparisons of means). (Table 3). Neither the warning's con- tent, its position on the page, nor its shape appeared from these limited data Table 5.-Masked Recognition by Smoking Status(N=61) to influence the time spent viewing the warning. Score Previous research has studied view- Smoking Status Pack Heading Warning ing behavior based on fixation time and Nonsmoker(n=30) 2.01 1.31 1.87 content recall.1 0'For this study,view- Experimenter(n=22) 2.32 1.44 2.02 ing behavior was classified as`looking" Regular smoker(n=9) 2.60 1.69 2.04 activity, consisting of eye movement P <.001 .02 NS* and short fixations (<0.10 s), and "reading" activity, which was charac- 'Ns indicates not significant. terized by visual fixations equal to or greater than 0.10 s. For all advertise- ments,the mean`looking"time for the recall for Marlboro and Red Man, the and heading. The relationship between warning was 0.52±0.80 s. The mean high heading recall for Marlboro and smoking status and the recall of the duration of the warning"reading"time Virginia Slims, and the extremely low warning was not significant, but the was 0.23±0.51 s. recall for the warning in the Red Man trend was in the expected direction. The 202 advertisement viewings advertisement.The mean masked recall In the final phase of the study, the were further analyzed according to the score of 1.9 for the warning correlates participants were asked to choose five type of viewing of the warning.In 43.6% with identification of this masked area actual warnings from a list that also in- of cases there was no viewing of the as a"warning"but without recall of its eluded five simulated warnings (Table warning.In 19.8%of cases,the warning general wording. In comparison, the 2).The five actual warnings were those was looked at but not read. In 36.7%of pack recall was higher (mean score, that participants had viewed earlier in cases,some fixations equaled or exceed- 2.2), while the heading areas were not the advertisements. The possible score ed 0.10 s, indicating that parts of the recalled as well(mean score,1.4). on this aided recognition test ranged warning were read. The participants' masked recall from 0 to 10. Random guessing would Following the eye tracking study, scores for the warning were significant- have led to an average score of 5. For each participant was asked to identify ly associated with total viewing time of the 61 study participants,the mean aid- masked areas of the five advertise- the warning(r=.252, P<.001). An in- ed recognition score was 7.25±1.59. ments(Fig 4).These masked areas cov- crease in this score was also associated The score on the this test was signifi- ered the cigarette pack,the main adver- with both the warning `looking" time cantly associated with the warning tisement heading, and the Surgeon (r=.232, P<.001) and the warning `looking" time (r=.167, P=.025) and General's warning.The recall scores for "reading"time(r=.211,P<.005). the "reading"time (r=.158, P=.025). the five separate advertisements were As indicated in Table 5, there was a The aided recognition score was nega- significantly different (Table 4). This significant association between the par- tively associated with age (r=-.273, was true for the pack,the heading,and ticipants! smoking status and their P<.001) (Table 6). This score was also the warning. Of note was the high pack masked recall score for both the pack associated with smoking status (non- JAMA,Jan 6,1989-Vol 261,No.1 Tobacco Advertisements-Fischer et al 87 Table 6.—Warning Recall Score by Age(N=61) there have been few studies examining 1986 after the 1981 FTC report indicat- Surgeon Generals Warning the effectiveness of the warnings. In ed that it might lead to higher visability Age,y awed aacall* 1981, the FTC published a report on than the rectangular warnings already 13 (n=7) 8.00.t 1,08 cigarette advertising that did discuss in use.'Within the visual environment 14 (n=17) 7.53±1.47 this issue!The FTC staff'claimed that of the Red Man advertisement that was 15 (n=14) 7.33±1.06 the original warning was y' y pe 18 (n=12) 7.14±1.26 ging g probably mef- used in this study, shape did not 17 (n=11) 6.36:t2.41 fective because it was (1) overexposed elicit a high level of eye kation. In the and worn-out; (2) no longer novel; (3) masked recall study, this warning also 'r=.273,�c op1. abstract and therefore difficult to re- generated a significantly lower mean member; and (4) not likely to be per- recall score than the other four warn- ceived as personally relevant. Most of ings(P<.001). smokers, 7.27; experimenters, 6.91; the data relating to the warning's effec- ' The masked recall scores indicate regular smokers,8.0;P<.0001). tiveness were omitted from the pub- that adolescents are more likely to iden- COMMENT lished FTC report because of its"confi- tify the pack than either the heading or dential"nature, the warning. Tobacco advertising is It has been eight years since Blum' The federally mandated warning is a principally picture and image based; first drew attention to the potential small, colorless, cognition-based mes- therefore,it is hardly surprising to find harm of tobacco advertising. Since sage that is usually at the periphery of that the worded elements in the adver- then, several types of evidence have an advertisement.This placement is un- tisements were not remembered as associated tobacco advertising with ad- der the control of those preparing the well. olescent smoking.'"'These studies in- advertisements for tobacco manufac- Seventy-four percent of participants dicate that adolescents and even very turers. The warnings compete with were able to identify the masked warn- young children are aware of the specific large, colorful, image-based messages ing as a health message but only 19% content of tobacco advertisements. In within the advertisement that relate were able to recall even the warning's one study, 309'0 of British 6-year-old smoking to romance, athletic success, general theme. Of interest was the fact children were able to identify that an excitement, and fun. The warning's ef- that the mean recall score for the warn- advertisement was for cigarettes even fectiveness must be measured within ing was higher than that of the heading. though it contained no smoking clues." the environment of the total advertise- We believe this is because subjects were Another study has shown that adoles- ment and compared with the effective- able to recognize the masked warning's cent recognition of tobacco advertising ness of the image-based advertising size, shape, and peripheral location on is closely associated with individual message with which it competes. the page. smoking status." A stronger causal We employed well-accepted market For the masked recall scores,a statis- reiationship between smoking and ad- research techniques to examine the tically significant dose-response rela- vertising will not likely be established viewing of tobacco advertisements by tionship was seen between smoking sta- since it would be unethical to examine adolescents. These included studies of tus and recall for both the pack and the this association in an experimental eye tracking,the recall of masked areas heading.This association has been iden- study and impossible in an observa- of the advertisements,and an aided rec- tified previously in adolescent popula- tional study to isolate the advertising ognition test for the Surgeon General's tions in the United States, Australia, effect from the multiple confounding warnings.While largely unknown with- and Great Britain."'In contrast,the factors in the environment. Neverthe- in the medical literature,these methods recall of the warning,while in the same less,a strong case can now be made that are frequently employed by market re- direction,was not significantly associat- the advertising of tobacco products con- searchers in the development of print ed with smoking status.This suggests a tributes to the initiation and mainte- advertisements."Eye tracking data different degree of interaction between nance of tobacco habits."' are considered to be some of the most the participants'smoking status and the The federally mandated tobacco valid measures of the acquisition of in- body of the advertisement compared warnings were designed to serve as a formation."Only those informational el- with smoking status and the warning. health message to tobacco users and po- ements that are visually fixated on can Despite the fact that Americans regu- tential users of these products.'There be picked up and transferred to the larly observe tobacco warnings in their are indications that the warning may be short-term memory. The recognition environment (ie, in newspapers and ineffective among teenagers,the group and recall testing then measures the ex- magazines and on billboards),the aided from which nearly all new smokers are tent of the processing and retainment of recognition data indicate that the recruited.The most telling of these indi- the information picked up." adolescents in this study were unable to cators is the fact that adolescents great- In the eye tracking study,the adoles- reliably identify actual warnings, ly underestimate the health conse- cent subjects viewed tobacco advertise- Scores on this aided recognition test quences of smoldng.',"A recent study ments for an average of 8.95 s,of which were lower for older adolescents and for has shown that 44% of adolescents ei- only 0.75 s was spent viewing the warn- those who experimented with tobacco ther experiment with or regularly use ing.In nearly half of the cases,subjects products. This may reflect cognitive tobacco products." failed to even look at the warning. In dissonance by those adolescents who The issue of the warning's effective- 36.7%of cases there was reading of the experiment with tobacco. Previous re- ness plays a prominent role in the Habil- warning but the mean duration of the search has indicated that both heavy ity of cigarette manufacturers for the total reading time was only 0.23 s.This and light smokers tend to discount the injuries incurred by persons using their is sufficient time for a reader of average negative impact of smoking." products.The importance of liability lit- speed to read only about one third of the The warnings used in the tested ad- igation as a cancer-control strategy has words in the warning." vertisements averaged 3.2%of the total been recently reviewed! Of interest is the mean viewing time space on the advertisement page. It Despite its potential importance as a for the Red Man warning(0.616 s).This might be expected that more regular health education message and its cen- warning's shape is an arrow into a circle. viewing of the warning would occur if tral role in the current liability debate, This shape was dictated by legislation in warnings were larger or were better 86 JAMA,Jan 6,1989—Vo1261,No.1 Tobacco Advertisements—Fischer et al 3 integrated into the advertisement. Re- will be required to validate this study's ings are the most frequently encoun- search on warning disclosure for pre- results using other populations. tered health messages in America. It is scription drugs has shown that risk in- Second, it is obvious that the study impossible with the present design to formation could be more successfully design produces some artificiality in separate the effect of prior exposure processed by consumers if the informa- terms of advertisement exposure. We from the exposure during the study tion was integrated into the advertising attempted to minimize this by not tell- viewing. The high level of previous en- message.' ing the participants at the time of the vironmental exposure to tobacco warn- Some subjects looked at the warning advertisement viewing that they would Ings makes the low recognition and re- but failed to fixate long enough for be questioned later about the advertise- call scores observed in this study all the reading to occur. Image-based warn- ment's content. Nonetheless,it is prob- more alarming. ings,like the skull and crossbones seen able that the advertisements were stud- In conclusion, our data indicate that on poison products,are more likely to be ied in more detail than would occur adolescents often do not see the warning effective than written warnings when during the routine viewing of a maga- in tobacco advertisements. Even when viewing is only for short periods of time. zine. It is therefore likely that, in nor- seen,there is little,if any,reading of the Iceland has employed a series of image- mal situations, the viewing time of the warning. In addition, adolescents are based warnings on its tobacco warning is much shorter.This strength- unable to recall the content of observed products." ens the major conclusions of the study. warnings or to correctly recognize An alternative option to increase the We have chosen fixations of 0.10 s as warnings from a list.If the warnings are warning's effectiveness might be to representative of reading behavior. not seen,or are seen but not processed, mandate freestanding counteradvertis- There is evidence that the shortest fixa- they are extremely unlikely to be effec- ing. Such counteradvertising would tion for reading comprehension is 0.20 tive in countering the promises of pow- have the advantage of not directly com- s.'If this more generous fixation time er,romance,and fun implied by tobacco peting within the context of tobacco were to be used,then only 18%of warn- advertisements. advertisements. ings would have been classified as hav- The results of this study should be ing been read at all. The minimal fixa- This study was supported in part by a grant from interpreted in light of several obvious tion time required to read the Surgeon DOC(Doctors Ought to Care),Augusta,Ga. limitations. First,the study population General's warning is unknown. It is, was small and had not been selected to however, likely to depend on the com- We would like to thank James J.Burke II;Pete be representative of all adolescents. plexity of the worded message, the Parramore;Joe Warwick, Jr; Katherine Guinan; The complexity of psychophysiologic reader's educational level, as well as and Velveeta Tanksley for their help on this study. advertising measurements tends to lim- previous exposure to the warning.$ We also appreciate the presubmission review of the it sample size.In one review of this type Finally, the masked recognition and manuscript by Alan Blum, MD; Don Shopland; Rollieof market research samples ranged aided recall tests suffer from environ- MPH;Harp;Richard ark,DayMd,JD;Nancy Neeh PMPH; Christopher Shank, MD;Jose Velez;Josh from six to 48 subjects.'Further work mental contamination. Tobacco warn- Borah;Leonard Scinto,PhD;and Ed Popper,PhD. References 1. Smoking and Health:Report of the Advisory Relative Risks of Reduced Exposure. New York, 26. Heller D: Reading text understanding, and Committee to the Surgeon General of the Public American Cancer Society,1980. their underlying skills:Eye movements in reading, Health Service, publication 1103. US Dept of 14. Mittelmark MB,Murray DM,Luepker RV,et in Groner R,Fraisse P(eds):Cognition and Eye Health,Education,and Welfare,1963. al:Cigarette smoking among adolescents: Is the Movements.New York,North-Holland Publishing 2. Federal Cigarette Labeling and Advertising rate declining?Prev,Med 1982;11:708-712. Co,1982,pp 139-154. Act,15 USC 1331.Amended 1965,1970,1984. 15. Aitken PP, Leathar DS, O'Hagan FJ: Chil- 27. Scinto LFM,Pillslamarri R,Karsh R:Cogni- 3. Myers ML,Iscoe C,Jennings C,et al:Federal drents perceptions of advertisements for ciga- tive strategies for visual search. Acta Psychol Trude Commission Staff Report on the Cigarette rettes.Soc Sci Med 1985;21:785-797. 1986;62:263-292. Advertising Investigation,May 1981(public ver- 16. Aitken PP,Leathar DS,Squair SI:Childreni; 28. Blum A:Medicine vs Madison Avenue:Fight- sion), Division of Advertising Practices, Federal awareness of cigarette brand sponsorship of sports ing smoke with smoke.JAMA 1980;243:739-740. Trade Commission. and games in the UK.Health Educ Res 1986;1:203- 29. Chapman S,Fitzgerald B:Brand preferences 4. Comprehensive Smokeless Tobacco Health 211. and advertising recall in adolescent smokers:Some Education Act of 1986,USC 1574-2. 17. Goldstein A0,Fischer PM,Richards JW,et'al: implications for health promotion. Am J Public 5. Anonymous:Cigarette smokers can sue for lung Relationship between high school student smoking Health 1982;72:491494. cancer:Is the'Surgeon General's warning'inade- and recognition of cigarette advertisements.J Pe- 30. Palmer v Liggett Group,Inc.US Court of Ap- quate?Law Alert 1984;63:4-7. diatr 1987;110:488491. peals for the First Circuit,doc. 86.1525, opinion 6, Daynard RA: Tobacco liability litigation as a 18. Krugman HE:Memory without recall,expo- written by Judge Brown, Senior Circuit Judge, cancer control strategy.JNCI 1988;80:9-13. sure without perception. J Advertising Res Aug 25,1987. 7. Jacoby J,Chestnut RW,Silberman W:Consum- 1977;17:7-12. 31. Silvis GL,Perry CL:Understanding and de- er use and comprehension of nutrition information. 19. Krugman HE:Processes underlying exposure terring tobacco use among adolescents. Psdiatr J Consumer Res 1977;4(Sept):119-128. to advertising.Am Psychol 1968;23:245-253. Clin North Am 1987;34:363-379. S. Blum A: Confronting America's most costly 20. Treistman J, Gregg JP: Visual, verbal and 32. Krohn MD,Naughton MJ,Lauer RM:Adoles- health problem:A dialogue with Surgeon General sales responses to print ads. J Advertising Res cent cigarette use:The relationship between atti- Koop.NY State JMed 1983;83(Dec):1260-1263. 1979;19:4147. tudes and behavior. MMWR 1987;36(suppl 9. Warner KE:A ban on the promotion of tobacco 21. Kroeber-Riel W:Activation research:Psycho- 4S):25S33S. products.N Eugl J Med 1987;316:745-747. biological approaches in consumer research.J Con- 33. Young E:Visibility achieved by outdoor adver- 10. Davis RM:Current trends in cigarette adver- sumer Res 1979;5:240.250. tieing.JAdvertising Res 1984;4:19-21. tising and marketing.N Engl J Med 1987;316:725- 22. Leckeriby JD,Plummer JT:Advertising stim- 34. Loken B:Heavy smokers',light smokers',and 732. ulus measurement and assessment research:A re-' nonsmokers' beliefs about cigarette smoking. J 11. McGinnis JM,Shopland D,Brown C:Tobacco view of advertising testing methods. Curr Issues Appl Psychol 1982;67:616.622. and health:Trends in smoking and smokeless tobac- Res Advertising 1988,pp 135-165. 35. Morris LA,Ruffner M,Klimberg R:Warning co consumption in the United States,in Breslow L, 23.Jones EE,Sigall H:The bogus pipeline:A new disclosures for prescription drugs. J Advertising Fielding JE, Love LB (eds): Annual Review of paradigm for measuring affect and attitude.Psy- Res 1985;25:25-32. Public Health. Palo Alto, Calif, Annual Reviews chol Bull 1971;76:349-364. 36. Bldndal T, Magndsson G: Innovation in Ice- Inc,1987. 24. Evans RI, Hansen WB, Mittelmark MB:In- land:Graphic health warnings on tobacco products. 12. Shopland D,Brown C:Smoking and Health:A creasing the validity of self-reports of smoking be- NY State J Med 1985;85(July):405406. National Status Report to Congress, US Dept of havior in children.JAppl Psychol 1977;62:521-623. 37. Watson PJ,Gatchel RJ:Autonomic measures Health and Human Services publication(PHS)87- 25. Berman EJ,Fischer PM,Richards JW,et al: of advertising.JAdvertising Res 1979;19:16-26. 8396.Atlanta,Centers for Disease Control,1987. Use of smokeless tobacco among adolescents. 38. Spache GD:Is this a breakthrough?The Read- 13. Dangers of Smoking,Benefits of Quitting and JAMA 1986;255:3245. ing Teacher 1962;15258-263. JAMA,Jan 6,1989-Vol 261,No, 1 Tobacco Advertisements-Fischer et al 89 i I The Surgeon General 's Warnings in Outdoor Cigarette Advertising Are They Readable? Ronald M.Davis,MD,Juliette S.Kendrick,MD Outdoor advertising media represent a major vehicle by which cigarette compa- four warnings on cigarette packages nies promote their products. We investigated the readability of the Surgeon and in cigarette advertising media.The General's warning in cigarette advertisements (ads) in two outdoor media: warnings required on outdoor bill- billboards and taxicab ads. In an experiment in metropolitan Atlanta under boards are slightly shorter versions of typical driving conditions,observers were able to read the entire health warning those required for other media (Table on 18(46%)of 39 street billboards but on only two(5%)of 39 highway billboards. 1)' The size of the warnings also has In contrast,the content of the ads(ie, brand name,other wording,and notable changed. The 1972 FTC consent order imagery)could be recognized under the same conditions on more than 95%of specified the size-of the warnings in the billboards. In a similar study of 100 taxicab cigarette ads in New York City, newspaper, magazine, and other peri- observers were unable to read the health warning in any of the ads but were able odical ads by assigning a Univers 47 j to identify the brand name in all ads and notable imagery in 95%of the ads. (Fdy)type style(eg, 10-pt type)to the Significant differences between the readability of the warning and identification warnings in ads of varying dimensions. of the advertising content persisted even when partially read wamings were For billboard ads, the required size of considered to have been read.We conclude that the Surgeon General's warn- the warnings was specified in inches.'In ing is not readable in its current form in the vast majority of billboard and taxicab 1975, the US government filed a com- ads.Factors contributing to unreadability include the small size of the letters,the plaint in the US District Court for the length of the warnings,the distance between the viewers and the ads, District of Columbia for alleged viola- excessive len 9 9 tions of the consent order.These alleged and movement between the viewers and the ads. violations included printing the warning (.JAMA 1989;261:90-94) in letters smaller than required on bill- boards and improperly placing the warning in some ads.'This action ulti- mately led to consent judgments in 1981 DURING the past quarter century,the "Caution: Cigarette Smoking May Be against the six major cigarette compa- health warnings on cigarette packages Hazardous to Your Health." In Nov- nies, requiring them to, among other and advertisements(ads)have changed ember 1970 this warning was strength- things,use larger lettering in billboard ened slightly to read: "Warning: The warnings."Under this settlement, the See also p 84. Surgeon General Has Determined That format and size of the warning were Cigarette Smoking Is Dangerous to specified in acetate exhibits (main- Your Health"(Public Health Cigarette tained on file at the FTC) for ads of several times. The Federal Cigarette Smoking Act of 1969 [Public Law 91- varying dimensions. Public Law 98-474 Labeling and Advertising Act of 1965 222]). In 1972,the Federal Trade Com- again increased the size of the letters, (Public Law 89-92) imposed the first mission (FTC) negotiated consent or- but in the case of billboard ads,it did so warning requirement. Effective Jan 1, ders with the cigarette industry, only by requiring that all letters be up- 1966, all cigarette packages were re- requiring that all cigarette advertising percase.This Act was the first to codify quired to carry the following warning: display clearly and conspicuously the into law the requirement for and the same warning required by Congress on sizes of warnings on ads.The sizes of the From the Office on Smoking and Health(Dr Davis) cigarette packages.' lettering now required are shown in and the Pregnancy Epidemiology Branch,Division of The Comprehensive Smoking Educa- Table 2. Reproductive Health(Dr Kendrick),Center for Chronic p g Disease Prevention and Health Promotion,Centers for tion Act of 1984 (Public Law 98474) The purpose of this study was to as- Disease Control,Atlanta. replaced the preexisting single warning sess the readability of the Surgeon Gen- Reprint requests to Office on Smoking and Health, , Centers for Disease Control,5600 Fishers Lane,Rock- with a rotational warning system that erals warnings on outdoor cigarette ad- ville,MD 20857(Dr Davis). requires cigarette companies to rotate vertising. Advertising in two media 90 JAMA,Jan 6,1989—Vol 261,No.1 Surgeon General's Warnings—Davis&Kendrick Table 1.-Health Warnings Required on Cigarette Packages and Advertisements Under the Comprehensive to closing and opening the eyes,to avoid Smoking Education Act of 1984(Public Law 98-474) the temporary reduction in visual acuity Cigarette Packages and AdveMeemeMa cigarette advertisements that occurs immediately after opening (Excluding Outdoor Billboards) on Outdoor Billboards the eyes.At the point of optimal visibili- SURGEON GENERAL'S WARNING:Smoking SURGEON GENERAL'S WARNING:SMOKING ty,the recorder instructed the observer Causes esLung Can r,Heary Complicate t ase,Emphysema, CAUSES EMPHYSEMA. CANCER,HEART DISEASE,AND to begin the observation attempt.After SURGEON GENERAL'S WARNING:Quitting SURGEON GENERAL'S WARNING:QUITTING three seconds of observation had tran- Smoking Now Greatly Reduces Serious Risks to SMOKING NOW GREATLY REDUCES SERIOUS spired on a stopwatch,the recorder in- Your Health. HEALTH RISKS. structed the observer to end the at- SURGEON GENERAL'S WARNING:Smoking by SURGEON GENERAL'S WARNING:PREGNANT tempt, at which time the observer PPremature Birth and nant Women L.ow Birth Weight. PPResult in Fetal Injury, RE AT RE BIRTH. replaced EN WHO SMOKE RISK FETAL INJURY AND replaced the posterboard obstruction. SURGEON GENERAL'S WARNING:Cigarette Smoke SURGEON GENERAL'S WARNING:CIGARETTE Three seconds was considered to be the Contains Carbon Monoxide. SMOKE CONTAINS CARBON MONOXIDE. maximum time during which most peo- ple riding in a car would look at a partic- ular billboard. The choice of this time period is supported by the findings of a Table 2.-Heights of Letters Required for Health Warnings on Cigarette Advertisements, United States, study of cigarette billboards in Austra- 1985 to the Present lia,which showed that health warnings, Approximate Height when legible to passing motorists,were Exhibit arse of Advertisement f C all L n* legible for 0.8 s to 2.3 s(average,1.5 s).' Warnings, However,the size of the letters in these 1 or 1(a) 0 to 65 sq in(0 to 4.19 dm2) 0.13 (3.2 mm) warnings was not mentioned in this 2 or 2(a) >65 to 110 sq in(>4.19 to 7.10 dm2) 0.14 (3.5 mm) article. 3 >110 to 180 sq in(>7.10 to 11.61 dm2) 0.15 (3.8 mm) Immediately after each observation 4 >180 to 360 sq in(>11.61 to 23.23 dm2) 0.16 (4.0 mm) attempt,the recorder asked the observ- 5 >360 to 470 sq in(>23.23 to 30.32 dm2) 0.19 (4.8 mm) er to describe the content of the ad or to 6 >470 to 720 sq in(>30.32 to 46.45 dm2) 0.32 (8.1 mm) cite the Surgeon General's warning, as 7 >5 to 10 sq ft(>46.45 to 92.90 dm2) 0.43 (1.10 cm) appropriate. Recognition of the adver- 8 >10 to 20 sq ft(>92.90 to 185.81 dm2) 0.61 (1.54 cm) tising content was assessed by measur- 9 >20 to 40 sq ft(>1.86 to 3.72 m2) 0.88 (2.24 cm) ing identification of three features of the 10 >40 to 80 sq ft(>3.72 to 7.43 m2) 1.26 (3.20 cm) ad: the brand name, the wording (be- 11 >80 to 160 sq ft(>7.43 to 14.86 m2) 1.77 (4.50 cm) sides the brand name),and any notable 12 >160 to 350 sq it(>14.86 to 32.52 m2) '5.50 (13.97 cm) imagery in the ad.Wording besides the 13 >350 to 1200 sq ft(>92.52 to 111.48 m2) 6.00 (15.24 cm) brand name was considered to have 14 >1200 sq it(>111.48 m2) 8.38 (21.29 cm) been identified if the major advertising *Unpublished data derived from measurements of health warnings on acetate forms maintained on file at the slogan or message was read(eg,"Amer- Federal Trade Commission. ica's Best,""Performance Counts,"and "Alive With Pleasure"). In most cases, these short messages were the only were included in the study:outdoor bill- (the "perimeter,"which surrounds the words in the ad besides the brand name. boards and taxicab ads. Cigarette ad- city in a roughly circular pattern) and The imagery was considered to have vertising in outdoor media and the effec- the major highways inside the perime- been recognized if the observer was able tiveness of the health warnings in these ter (highways 20, 75, and 85). Thirty- to describe at least one of the following: media are particularly important for at nine billboard cigarette ads alongside the presence of a model or models in the least three reasons:(1)billboard ads are these highways were found.The first 39 ad(eg, a cowboy),the activity in which more intrusive than other print ads and cigarette billboards found by the inves- models were involved (eg, volleyball), are more difficult to ignore,(2)children tigators while driving on city streets or any objects that were displayed(eg,a are unavoidably exposed to billboard also were studied;these billboards were sailboat or cigarette pack). In many ads, and (3) small billboards represent located alongside 12 different streets in cases, cigarette ads had only one of one of the most effective media for tar- the city. these three features of imagery. geting racial and ethnic minority A team of three persons-a driver,an The Surgeon General's warning was groups.6-7 "observer,"and a data recorder-rode considered to be entirely read,partially in an automobile past each cigarette bill- read,or not read at all.For the warning METHODS board on two separate occasions.On the to be considered partially read, one or Outdoor Billboards first pass, the observer was instructed more words identifying a health risk had to try to identify the content of the ciga- to be read, including "lung cancer," In September and October of 1986, rette ad. On the second pass, the ob- "heart disease,""emphysema,""health we assessed the readability of the Sur- server was instructed to try to read the risks," "injury," or "premature birth" geon General's warning for 78 billboard Surgeon General's warning. As the (Table 1). In the case of the warning cigarette ads in metropolitan Atlanta. automobile approached a test billboard about carbon monoxide in cigarette The ability to identify the brand name, for each pass, the recorder instructed smoke, the entire warning would need wording, and imagery in the same ads the observer to cover his or her line of to be read for the message to have was measured for comparison. vision with a small, white posterboard meaning; thus, partially read carbon These assessments were made in the (which was held approximately 12 in[30 monoxide warnings were considered following manner. All billboard ciga- cm]from the eyes)and informed him or not to have been read at all. rette ads alongside interstate highways her of the side of the highway or street Two observers were used. These in- in metropolitan Atlanta were located. on which the billboard was stationed. dividuals were employees of the Cen- These highways included Highway 285 The posterboard was used, as opposed ters for Disease Control who were JAMA,Jan 6,1989-Vol 261,No.1 Surgeon General's Warnings-Davis&Kendrick 91 asked to participate in the study be- Table 3.—Recognition of Cigarette Advertising Content and Surgeon General's Warning on 78 Billboards, cause they fulfilled three criteria: (1) Atlanta,September to October 1986,and on 100 Taxicabs,New York City,April 1987 they were well-educated adults with surgeon Generals normal vision or normal"corrected"vi- Cigarette Advertisement warning sion and,thus,had the capacity to read Brand Totally At Least legible warnings;(2)they had no preex- wane Wording' Imagery* Reed Partly Readt fisting knowledge of the specific wording - of the four Surgeon General's warnings; Street billboards dswlum %% (NO.) % (No.) % (NO.) % (NO.) % (NO.) and(3)they played no role in the formu- 8-sheet 100 (30/30) 96 (24/25) 100 (30/30) 40 (12/ao) 67 (20/30) lation of the study hypothesis or design. 30-sheet 100 (9/9) 100 (ere) 100 (9/9) 67 (6/9) 78 (7/9) Both were nonsmokers. One observed Total 100 (39/39) 97 (32/33) 100 (39/39) 46 (18/39) 69 (27/39) 53 billboards and the other observed 25. Highway billboards 100 (39/3x) 97 (34/35) 100 (36/36) 5 (2/39) 15 (6/39) The observations were standardized Taal 100 (78178) 97 (66/ss) 100 (77rM 26 (20178) 42 (33178) by applying a number of conditions to Taxicabs 100 (53/53) 51 (24/47) 95 (37)39) 0 (0/47) 0 (0/47) the study. Observations were made at times when visibility was not impaired 'Ads that had no wording or imagery were excluded from the denominators. by weather conditions and when traffic "he ase"Partly read"mpphhythat e a�hea risks," njury�or premature bilth rth." had to be read,including"lung cancer;' was minimized(ie, from 9 AM to 3 PM). The observer sat in the front passenger seat of the automobile and the recorder the appropriate lanes was observed as brands_.There were 14 ads for Vantage, sat in the rear seat.The car traveled in part of the study,with two exceptions: 13 for Winston Lights, ten for Kool the far right lane, within 5 mph of the (1)those for which the view of the ad Milds, and six or fewer for each of the speed limit. When traffic congestion was obstructed by a pedestrian or a mo- remaining brands. About half of the caused the car's speed to be reduced to for vehicle and (2) when two or more street billboards (49%) were on the below the speed limit, a similar speed taxicabs with cigarette ads were too right side of the street and about half of (within 10 mph)was used for the second close together to allow separate obser- the highway billboards (44%) were on pass. vations. This process continued until the right side of the highway. Of the 39 For each ad, the side of the road on 100 taxicab ads had been observed. street billboards,30 were 8-sheet(77%) which the billboard was located and the Two observers were used;these indi- and nine were 30-sheet(23%)(Table 3). size of the billboard were noted. Bill- viduals were nonsmoking employees of All of the 39 highway billboards were boards were either"8-sheet"(5 x 11 ft the New York City Department of 30-sheet. [1.5x3.4 m]) (exhibit 10 [Table 2]) or Health and were chosen based on the The observers identified the brand "30-sheet"(9 ft 7 in by 21 ft 7 in[2.9 x 6.6 same criteria used to select observers name and imagery in 100% of the bill- m])(exhibit 12[Table 2]). for the billboard study. One partici- board ads and the wording(when pre- Taxicabs pated in 66 observations(35 for adver- sent)in 97%of the ads(Table 3). There tising content and 31 for Surgeon Gen- was no significant difference in these Cigarette ads on taxicabs were inves- eral's warnings) whereas the other findings between the two observers or tigated on April 23, 1987, in New York participated in 34 observations (18 for when comparing street vs highway bill- City. These ads are 14 x 48 in(36 x 122 advertising content and 16 for Surgeon boards, 8-sheet vs 30-sheet billboards, cm)(exhibit 6[Table 2])and are placed General's warnings). Each observer or billboards on the right vs the left side on top of the taxicabs along the front-to- was told to observe ads consecutively of the street or highway. back axis. The observation methods for either advertising content or the The Surgeon General's warning was used were similar to those described Surgeon General's warning until in- readable in its entirety in less than half previously herein for the study of bill- structed to switch. Each observer be- of the street billboard ads(46%)and in board ads. Features of the taxicab eval- gan by viewing the ads for advertising only 5% of the highway billboard ads uation that differed from the billboard content,and switches were made after (P<.001). If 8-sheet billboards are study are as follows. viewing five to 15 ads. These switches excluded from this comparison, a An observer and a recorder sat on a were made to ensure that when the ob- statistically significant difference per- sidewalk bench in Central Park along- servers had completed their "shifts," sists(P<.001). The difference between side a six-lane street (Central Park each had observed an approximately the readability of the warning on street West at 65th Street). The two outer- equal number of ads for either advertis- 8-sheet billboards (40%) and on street most lanes were used only for parking. ing content or the Surgeon General's 30-sheet billboards(67%)was not statis- Only taxicabs traveling in the second warning. tically significant.There was no signifi- and third lanes(closest to the observer) Statistical Analysis cant difference in readability of the were included. These lanes were ap- warning between observers or when proximately 25 to 40 ft(7.6 to 12.2 m) Probability values were calculated comparing billboards on the right vs the from the observation bench. Most taxi- using Fisher's exact test, except for left side of the street or highway. cabs and other cars were traveling at matched-pair analyses, for which the A significant difference occurred be- speeds of approximately 5 to 25 mph. McNemar test was used. For P values tween the proportion of ads for which Because a particular`taxicab ad could cited later herein, Fisher's exact test the entire Surgeon General's warning not be viewed twice(as opposed to the was used unless otherwise indicated. was readable and each of the three billboards, each of which was viewed Values greater than .05 were not con- proportions for which the brand name, twice), separate groups of taxicabs sidered statistically significant. wording, or imagery was recognized were used to assess recognition of the RESULTS (P<.001, McNemar test). This differ- advertising content (N=53) and the Outdoor Billboards ence occurred within both the street readability of the Surgeon General's and the highway billboard categories warnings(N=47).During the period of The 78 billboard ads included in the (P<.001, McNemar test). The signifi- observation, each taxicab traveling in study promoted 16 different cigarette cane of this finding persists even when 92 JAMA,Jan 6,1989—Vol 261,No.1 Surgeon General's Warnings—Davis&Kendrick a• comparing the proportion of billboard Of the 78 billboards observed,three ads in advertising ... could result, as a prac- ads for which the warning was read with each appeared ten or more times. tical matter, in an end to all cigarette the proportion for which the brand We attempted to minimize this prob- advertising! name,wording,and imagery were rec- lem in two ways. First, the recorder The health warning in cigarette ads ognized. A significant difference also repeatedly instructed the observer to did not,of course,result in an end to all persists when partially read warnings report only that which he or she had cigarette advertising. In fact, cigarette are included as read (P<.01 for street actually seen or read in the ad currently advertising and promotional expendi- billboards, P<.001 for highway bill- being observed.Second,we used a total tures have increased steadily through boards, and P<.001 for all billboards; of four observers, two for the billboard the years and now exceed$2 billion per McNemar test). investigation and two for the taxicab year."Cigarette advertising has been study.Results were similar for each ob- particularly widespread in outdoor me- Taxicabs server,and findings were similar at the dia. In 1986,cigarette companies spent beginning and at the end of each observ- $337 million for outdoor and transit ad- Of thved for advertising cigarette ads ads er's "shift." We emphasize that the vertising, an amount that represented were for Parliament Lights, o were for same opportunity existed for the ob- 36% of total cigarette advertising ex- Carlton(nine a one type t , and five of servers to learnthe Surgeon Gener- penditures in the print media." Ciga- Carltor), seven were for Marlboro,and al's warnings had the warnings been rette advertising accounts for 22% of one was for Benson and Hedges. readable. In fact, the billboard study total advertising revenue in outdoor The observers recognized the brand was intentionally designed so that the media. Tobacco is the most heavily ad- names o all of these ads and the i randag- warnings were observed on the second vertised product in outdoor media in ery (when present) ins an t the ads pass;thus, observers theoretically had general and on 8-sheet billboards in par- (Table he Recognition of thefthem two opportunities to observe the warn- ticular.'All five companies having the wording ings and only one opportunity to ob- highest advertising expenditures in the ads was substantially lower(51%), g g P mainly due to failure to note the slogan serve the advertising content. In addi- outdoor media in 1987 were cigarette "The Perfect Recess"in ads for Parlia- tion, the number of different warnings companies (Advertising Age, Sept 28, ment Lights. No significant difference (four) was considerably less than the 1988,p 48). in these results occurred between number of different ads on the bill- observers. boards (16). Thus, in this sense the The Effectiveness of study design was biased toward not Health Warnings The Surgeon Generals warning could finding a difference between readability not be read(even partially)in any of the of the warnings and recognition of the Few studies have evaluated the effec- 47 ads for which the attempt was made. advertising content. tiveness of the health warnings on ciga- This proportion (0%) was significantly Our findings are consistent with the rette packages and ads." Such studies lower (P<.001) than each of the three results of a stud of the legibility of are difficult to c proportions of ads for which the brand y � y airy out because warn- name,wording, or imagery was ibrand health warnings on 37 cigarette bill- ing requirements invariably coincide fled. The significance of this difference boards in Australia. Cullingford et al with the institution of other policies and persists when the proportion i ads for constructed a model to objectively as- programs that discourage smoking and which the warning was read is coin- sess legibility of the warnings,based on with changes in social attitudes toward pared with the proportion d ads for the angle that was subtended by the smoking. Thus,the effects of the warn- which the brand name, wording, and image at the retina, the color contrast, ings themselves are difficult to isolate. imagery were identified. and the obliqueness of the message. In "Laboratory"studies offer an opportu- other words,legibility was defined only pity to infer whether the warnings are COMMENT as a function of the optical limits of the likely to have an impact. For example, eye.These investigators found that 51% as reported in this issue of THE JOUR- This study confirms what seems obvi- of the health warnings were not legible NAL, Fischer et al's used an eye-track- ous on casual observation:the Surgeon to passing motorists whereas 98%of the ing technique to show that in a sample of General's warning on billboard ciga- brand names were legible. As men- 61 adolescents, the average viewing of rette ads usually is unreadable under tioned previously herein,the investiga- the warning was only 8% of the total typical driving conditions. By contrast, tors did not report the size of the warn- advertisement viewing time; in 44%of one can recognize the brand name and ing letters in the Australian billboards. the ads,the warning was not viewed at the words or imagery in the ads them- all, selves in almost all cases. The warning Cigarette Advertising After What factors determine the effective- on highway billboards is particularly Health Warnings ness of a warning? The format, color, difficult to read; the warning could be and placement of the warning may af- read in only two of 39 highway bill- When the first cigarette warning feet the degree to which it is noticed and boards included in this study.Similarly, requirement was being debated in Con- read. In 1981,the FTC recommended a the warning could not be read in any of gress in 1964, Bowman Gray, chairman "circle-and-arrow"format for cigarette the 47 taxicab cigarette ads observed in of the Board of Directors of R.J. irm - warning labels."This format was later this investigation. olds Zbbacco Co, testified on behalf of adopted by Congress (Comprehensive A major methodological difficulty in the cigarette industry against a health Smokeless Ibbacco Health Education this investigation was the potential for warning requirement for cigarette ads: Act of 1986 [Public Law 99-252]) for observers to remember the content of health warnings on smokeless tobacco cigarette ads that were observed re- Advertising is basic to the successful diads; however, billboards were ex- peatedly in the study.This problem was distribution sale of any consumer item on empted from the warning requirement a national basis.The right to advertise is an more important in the case of the taxi- essential commercial right and is virtually because opposing sides could not agree cab ads;only five different ads appeared destroyed if one is required in every on an acceptable format for billboard among the 53 observed for advertising advertisement to caution against the use of ads(E.Ripley Forbes,oral communica- content,and one ad appeared 31 times. the product.The requirement for a warning tion, Nov 10, 1987). Equally important JAMA,Jan 6,1989—Vol 261,No.1 Surgeon General's Warnings—Davis&Kendrick 93 is the wording of the warning.The FTC materials,a condition rarely met in bill- McQuade, Diane Orenstein, Robert Pollard, and concluded that the pre-1985 single board viewing.These standards also de- Darenaa J.Rogers.James S.Marks,MD,Kenneth warningfor cigarettes was ineffective fine the letter heights according to the E. Powell, MD, and Rick Hull provided helpful gaze g g comments on the manuscript. because it was overexposed and "worn lowercase letter m;capital letters,such out,"too abstract, and not likely to be as those in the billboard warnings,need 1. Complaint°�° Mor- perceived as personally relevant."The to be larger because they are less ri American Br matter of Inc, wnan William- warnings s should be "linguistically ade- legible.' so Inc,American Brands ,Brown and William- g � Y � son Tobacco Corporation,R.J. Reynolds Tobacco quate';strong warnings are written in Several approaches might remedy Company, Liggett&Myers Inc, Consent orders, simple syntax, in ordinary, everyday the unreadability of the warnings on etc,in regard to the alleged violation of the Federal language,and without qualifiers such as outdoor cigarette ads.The required let- Trade Commission Act, Complaints, March 30, "may"or"could.""Warnings also might ter size for the warnings could be in- 1972, 9 2,8Decisions, 4"� ' Mph 30, 1972. FTC Decis 65. go unnoticed because the imagery in the dexed to the size of the letters in the ad 2: Report to Congress Pursuant to the Federal ad is powerfully distracting.""" All itself. For example,the warning letters Cigarette Labeling and Advertising Act For the these factors are irrelevant,of course,if could be required to be at least half the Year 1980.Federal Trade Commission,1982. 1981). the warningcannot even be read under size of the largest letters a in 3. U.S.A.v Liggett et of s Civ cco Co.NY 0-198 g PP ►g 4. U.S.A.v R.J.Reynolds 7bbacco Co.,1980-1981 usual viewing conditions. the ad. Alternatively, the required let- Trade Cas(CCH)5 63,847(S.D.N.Y.1981). Certain unique features of warnings ter sizes could remain indexed'to the 5. The unique harm of cigarette billboards. NY on outdoor ads make these warnings size of the ad (Table.2) but could be State J Med 1983;83:1321. 6. Be- particularly difficult to see. First, ei- increased substantially. These two ap- ore Advertising of ttee on Products:Hearings Be- ther the viewers(of billboard ads)or the roaches could be used together: the ment HS�eofRep Subcommittee en Health and the Environ- p g f Representatives Committee on En- ads themselves(on taxicabs and buses) first as the primary requirement and ergy and Commerce,99th Cong,2nd seas(July 18 usually are moving.Thus,a short period the second for ads that have no wording. and Aug 1, 1986), serial No. 99-167, pp 772-779 (testiis available for viewingthe ads. Given 7b increase the size of the warnings,the rector,Co of Edward n McMahon,executive di- !�� rector,Coalition for Scenic Beauty). this brief interval, the wording in the length of the warnings would need to be 7. Davis RM:Current trends in cigarette advertis- warnings is much too long to be read, shortened considerably: eg, "Smoking ing and marketing.NEngl JMed 1987;316:725-732. despite the fact that the warnings on Causes Cancer," "Smoking Kills," or 8. Cuuingford R,Da Cruz L,Webb S,et al:Legi- billboard ads are slightly shorter than "Cigarettes are Addicting."(One study �e cigaareetttes Mw�JA t 1988;148on :336-3that 38. those in other print ads(Table 1).Infor- suggests that"Smoking Kills"would be 9. Cigarette Labeling and Advertising:Hearings mation that can be processed as"photo- particularly effective.YO) The format of Before the House of Representatives Committee on graphic images" (eg, large-type ciga- the warning could be modified to in- Interstate and Foreign Commerce,88th Cong,2nd rette brand names and large objects crease the likelihood that it is seen(eg, seas(June o 25'29,and July c Irma pp Board 9 (testimony of Bowman Gray, chairman, Board of such as boats or people)may be all that the circle-and-arrow format used for Directors,R.J.Reynolds'Ibbacco Co). the typical passerby absorbs. Second, smokeless tobacco warnings in newspa- 10. Report to Congress Pursuant to the Federal the warnings on billboard ads(but not in per and magazine ads). Ideally, any Cigarette Labeling and Advertising Act,1986.Fed- other print ads)are required by Public changes in the required size, format, 11. Trade Commission,1988. 11. Review of the Research Literature on the Ef- Law 98474 to be printed in capital let- placement,and wording of the warnings fects of Health warning Labels:A Report to the ters. However, lowercase letters are should be field tested before such re- United States Congress. Office of the Assistant easier to read because their shapes are quirements are set forth on a national Secretary for Health,US Dept of Health and Hu- manServices,19s7.distinct;capital letters resent a mono- basis. 12. Fischer PM,Richards JW,Berman EJ,et al: tone appearance in a line." Third, the The FTC concluded that making bill- Recall and eye tracking study of adolescents view- warning in 30-sheet billboards appears board warnings large enough to be ing tobacco advertisements. JAMA 1989;261:84- in one long line at the bottom of the ad, clearly visible could take up a substan- 89• whereas the warningin smaller ads is tial amount of the totals ace on a bill- 13. Myers misML,Ion C,Jennings C,the al:Federal P 7'hade Commission Staff Report on the Cigarette broken up into more than one line. board.'Thus,the commission proposed Advertising Investigation.Federal Trade Commis- Shorter lines are easier to read." another option: to permit cigarette sion,1981. Fourth, there is often a huge distance manufacturers to design their billboard 14. Dumas BK:The adequacy of federally mandat- ed between viewers and billboards (es ads without a health warning but to re- A cigarette package warnings,in Levi J,walker Pe- g AG(eda):Language in the Judicial Process.New cially those alongside the highway), quire the manufacturers to devote a cer- York,Plenum Press,in press. which makes the size of the letters tain number of their billboards solely to 15. Richards JI,Zakia RD:Pictures:An advertis- inadequate.. the disclosure of the warning in an ex- ere expressway through FTC regulation.Georgia Rev A guide for producing audiovisual panded size."," A possible loophole in e er C:Graphic propositions:The efficacy of materials recommends a minimum let- this strategy is that tobacco companies imagery and the impotence of warnings in cigarette ter height(for the lowercase letter m)of might place the health warnings on bill- advertising. 7bbacco Products Litigation Rep 1/4 in(6.4 mm)for a maximum viewing boards in more obscure locations. 1986;1:4.714.85. distance of 8 ft (2.4 m). For greater Under conditions such as these ci a- Z. Romano F:Practical l Composition From oc to � � g Z.Arlington, Va, National Composition Asaocia- distances,these two values are doubled rette companies might choose not to ad- tion,1983. repeatedly." Thus, a minimum letter vertise in outdoor media.This possibili- 18. Kemp JE:Planning and Producing Audiovi- height of 8 in(20.3 cm)would be recom- ty should not deter the adoption of an sual Materials,ed 2.Scranton,Pa,Chandler Pub- fishingCo,1968. mended for a maximum viewing dis- appropriate public policy. A company's19. Comprehensive Smoking Prevention Eduea- tance of 256 ft(78 m).The height of the voluntary decision to end outdoor ad- tion Act:Appendix to Hearings Before the Subcom- letters in the health warning on the larg- vertising to avoid the display of read-. mittee on Health and the Environment,House of est billboard is 8.4 in(21.3 cm)(Table 2), able health warnings should be seen as a Representatives Committee on Energy and Com- et billboards are stationed u to 400 ft small rice to a for marketing a rod- merce,97th Cong,2nd seas(March 5,11,and 12, Y P P pay 8 P 1982), serial No. 97-107, pp 64119 (report by (122 m)from viewers.19 By these stan- uet that causes more than 300 000 Keenan&McLaughlin Inc). dards, the height of the letters in the deaths among Americans each year." 20. Health warnings.Health Educ J 1%5;44:218- health warnings is clearly inadequate. 219. The authors are indebted to the following per- 21. Smoking-attributable mortality and years of Moreover,these standards assume slow sons for their assistance in conducting this study: potential life lost-United States, 1984. MMWR and careful observation of projected Deidre Downs, Richard A. Lasco, PhD, Susan 1987;36:693-697. 94 JAMA,Jan 6,1989-Vol 261,No.1 Surgeon General's Warnings-Davis&Kendrick j . Special Communication The Making of a Smoke-free Medical Center Richard D.Hurt,MD;Kenneth G.Berge,MD;Kenneth P.Offord,MS;David A.Leonard; Doreen K.Gerlach,RN,MS;Chris Larson Renquist;Michael R.O'Hara The concept of smoke-free medical facilities is in its formative stages,and such personnel 12 months in advance of the policies have received broad support.Although smoking has been restricted at effective date. This lead time was ad- Mayo Medical Center for many years, in 1986 it was decided that the medical vised by those who had pioneered simi- center should become smoke free.This report summarizes the methods used in lar efforts,'and it provided ample time developing and implementing a smoke-free policy. This experience suggests for preparation and publicity and tem- that with proper planning, the implementation can be smooth. The message pered potential negative reactions to being given to patients and staff concerning the health risk of smoking is now the policy. Following the decision to become more consistent,and there has been wide acceptance of the policy by staff and smoke free, a task force consisting of patients.It is concluded that such a policy can be effectively implemented,if it is representatives from consultant(physi- well planned and supported. cians, scientists, and administrative (JAMA 1989;261:95-97) personnel) and paramedical staffs, in- eluding smokers and nonsmokers, was appointed and charged with implemen- tation of the policy. Early discussions ALTHOUGH there is a glaring contra- This report summarizes the methods centered on ways in which the policy diction between the health hazards of used by Mayo Medical Center (Mayo might be made as acceptable as possible smoking and the acceptance of smoking Clinic and its two affiliated hospitals, and on how compliance could be solic- in medical facilities, only recently has Rochester Methodist Hospital and ited. The task force initiated activities the concept of a smoke-free medical fa- Saint Marys Hospital—a combined total to (1) gather information from other cility become a reality,' and an un- of 15 000 employees)in developing and smoke-free health-care facilities,(2)de- derstanding of the necessary ingredi- implementing a smoke-free policy in termine the optimum implementation ents to implement a smoke-free policy 1987. The policy has met with a high date, (3) develop the specific policy has begun to emerge.``Smoke-free poli- level of success as indicated by the statement,and(4)plan implementation cies in health-care facilities have been smooth implementation and wide accep- procedures. It was agreed by the task broadly supported by patients and tance. force that communication would be the staff,'an experience similar to that seen GETTING STARTED key to successful implementation. Oth- in health-care-related industries. By er considerations were seeking input addressing the impact of such a policy Active support from top administra- from smokers and nonsmokers, listen- well in advance, broad support,includ- tion is necessary for the successful im- ing to the issues, carefully considering ing that of union organizations, can be plementation of a smoke-free policy. In the options, and informing the employ- obtained and significant problems can August 1986, the governing boards of ees fully about the decisions that were be avoided.' Mayo Clinic and the affiliated hospitals made. From the Division of Community Internal Medicine approved a recommendation that Mayo A survey of all Mayo Medical Center (Drs Hurt and Berge),the Section of Biostatistics(Mr and the Mayo-affiliated hospitals be- employees had four principal purposes: Offord), and the Section of Communications (Mr come smoke-free facilities, stating that (1)to help identify problem areas that O'Hara),Mayo Clinic and Mayo Foundation;Adminis- 'to tito tration(Mr Leonard)and Educational Services(Ms Ren- continue permit smoking 111 Mayo had not been anticipated and explore quist),Saint Marys Hospital;and Nursing Services(Ms facilities would be inconsistent with our areas of concern for employees and pa- Getrlrach), Rochester Methodist Hospital, Rochester, leadership role in the health field."The tients who smoke,(2)to serve as part of Reprint requests to Mayo Clinic, 200 First St SW, written communication of the decision a total information program aimed at Rochester,MN 55905(Dr Hurt). to become smoke free was sent to all heightening awareness of the policy JAMA,Jan 6,1989—Vol 261,No.1 A Smoke-free Medical Center—Hurt at at 95 Table 1.—Smoke-free Policy of Mayo Medical Table 2.—Smoking-Cessation Services especially those of an addicted smoker, Center attempting to comply with a smoke-free Self-help material environment. Written guidelines were Staffs/students/professional visitors Quitters'Guide—American Cancer Society Smoking will be prohibited in all facilities used by Calling it Quits—American Heart Association developed for nursing staff in the hospi- Mayo Medical Center,including buildings,hospi- Lifetime o/Freedom From Smoking—American Lung tals, including instructions on how to tals,grounds,vehicles,and leased space. Association Smoking-cessation help will be offered for those who Smoking-cessation programs assist a patient experiencing significant wish to stop smoking. 90-minute education/intervention program—Mayo nicotine withdrawal. Compliance with the policy is expected,based on the Section of Patient and Health Education A dedicated newsletter, Smoke Free importance of our example to our patients and the Fresh Start—American Cancer Society public, smokers'Clinic—Rochester Methodist Hospital '87, was distributed monthly to each Outpatients and all other visitors Mayo Medical Center employee for six Smoking will be prohibited in all areas described above. months prior to implementation. The Smoking-cessation programs will be made available, last newsletter before implementation and.parbcipation will be encouraged. There will be no tobacco sales on Mayo Medical The implementation task force devel- contained a note from the chief execu- Center premises. oped a policy statement that was clear, tive officer about the importance of the Communication of the policy will be by notice prior to concise, and contained two important he and the importance of staff sensi- arrival wherever possible,by appropriate signs m y p facilities,and by reminders from Medical Center concepts: (1) concern and compassion tivity to the needs of patients. A final staff when necessary for the individual and (2) the goal to edition was produced three months af- Inpatients and their visitors Smoking will be prohibited in an rooms and through- provide a smoke-free environment(Ta- ter implementation that included an out all hospital buildings and grounds. ble 1).Consistent with other Mayo poli- overview of the implementation pro- A limited exception will be permitted in a few specified common areas available only to psychiatry and cies, enforcement issues were ad- cess, a discussion of problems that had chemical dependence inpatients. dressed only in general terms. occurred,and a reiteration of the firm- Compliance is expected and will be sought with Hess Of the clic compassion and by tact,diplomacy,and the exer- PLANNING FOR IMPLEMENTATION policy. ase of appropriate judgment. Regular staff newsletters and feature Information sessions presented by articles in other internal publications professionals regarding smoking behav- were used.The articles stressed the un- ior (including the addictive nature of derlying reason for the decision and ap- nicotine) were available for all inter- pealed to the fundamental commitment change and informing the staff of imple- ested medical center staff during work that the staff feels toward the institu- mentation plans, (3)to ascertain exist- time and were attended by more than tional mission. A letter from the presi- ing attitudes and provide an outlet for 1500 employees. The intent of these dent of the consultant staff was sent to expression by those who disagreed with one-hour sessions was to(1)establish a all members of that staff prior to imple- the decision, and (4) to determine the common knowledge base of smoking be- mentation urging active support for the smoking prevalence and interest in havior, (2)provide a basic understand- policy and pointing out their important smoking-cessation assistance among ing of the problems employees and status as role models. A statement of the employees.The survey was used to patients might experience when at- the policy was provided for all job appli- involve employees in the development tempting not to smoke,and(3)provide cants, candidates for appointments to of the implementation procedures, but information about smoking-cessation educational programs, and patients it was not allowed to be construed as a assistance available to staff and fam- with future appointments. means to debate the decision to become ilies.A brochure was distributed the IMPLEMENTATION smoke free. information sessions that highlighted The overall response rate to the sur- the major points to be covered and in- At Mayo, the positive elements of vey was 49.3%. There were no major cluded a selection guide to be used to providing a smoke-free environment for differences in the data (eg, smoking self-select the smoking-cessation pro- patients and staff were emphasized.In- rate,age and sex distribution)from the gram most suited to the smoker's indi- ternal publicity was designed to provide three institutions. vidual needs. a crescendo effect prior to the imple- The overall smoking rate of those who Table 2 summarizes the smoking- mentation, with strong endorsement responded was 17% (15% for men and cessation services provided at no cost from the leadership groups. Special 18%for women).There was a 9%smok- but during nonworking hours to em- group meetings were held with medical ing rate reported by respondents ployees and their families during the and administrative staff to emphasize among the consultant staff and a 17% implementation of the smoke-free poli- the importance of their full support and rate by respondents among the para- cy. Prior to and six months after the participation. A news release was pro- medical personnel. Of the current implementation of the policy, 11 em- vided to local media to help alert pa- smokers,52%were under the age of 35 ployees attended the 90-minute pro- tients to the change to a smoke-free sta- years (41%of the men and 56% of the gram,53 attended the Fresh Start pro- tus and to stress the positive aspects of women). Of the current smokers, 49% gram, and 114 attended the Rochester the policy. Signs stating, "Mayo Clinic expressed interest in a smoking-cessa- Methodist Hospital Smokers' Clinic! Is Smoke Free.Smoking Is Not Permit- tion program. The disparity between the number of ted in Any Mayo Facility or on Medical The major issues from the "Com- employees who attended smoking-ces- Center Grounds," were carefully de- ments" section of the questionnaire sation programs and the survey results signed to be relatively unobtrusive but were (1) concern for patients who that indicated 49%of the current smok- strategically placed at each facility en- smoke; (2) concern for coworkers who ers were interested in such programs trance. Signs were also placed in areas smoke;(3)the need for smoking-cessa- remains unexplained. where smoking had previously been al- tion programs; (4)enforcement issues, To meet the more specific needs of lowed that stated, "A Reminder, Mayo including whether there would evolve a direct patient-care staffin the hospitals, Medical Center Is Smoke Free." This double standard for consultant and non- a "smoke-free caring" video was pro- wording was chosen to avoid the possi- consultant staff;and(5)concern that the duced for in-service education.This vid- ble misinterpretation of a "No Smok- policy represents an intrusion on indi- eo presented an overview of hardships ing"sign,which might imply that there vidual rights. encountered by a hospitalized smoker, is some other area where smoking is 96 JAMA,Jan 6,1989—Vol 261,No.1 A Smoke-free Medical Center—Hurt et al y` allowed.On the day of implementation, acceptance of the policy)and its resolu- The message being given to patients the change was emphasized with fresh tion. These records were reviewed by and staff concerning the health risk of flowers on tables in the cafeterias and the Smoke-free Task Force and were smoking is now more consistent, and it coffee rooms. useful in developing a consistent ap- is believed that the policy will have a PROBLEM AREAS proach to policy enforcement. Excep- positive long-term impact. tions to the policy, inasmuch as they Although there are some differences The experience at Mayo has been tend to create confusion,have been and between Mayo and other medical orga- positive,but there have been problems. will continue to be rare.Guidelines were nizations, we believe that the methods A few patients have canceled appoint- developed whereby a decision to allow described here are transferable. Mayo ments and a few have left the hospital an exception to the policy for a hospital- hospitals operate in basically the same because of the policy. Overall there has ized patient must involve agreement by way as other hospitals, and Mayo pa- been no noticeable effect on the number (1)the staff physician caring for the pa- tients are like patients at any other of patients seen.Many laudatory letters tient, (2)the nursing service involved, large medical center. Although Mayo is and comments have been received from and(3)the Smoking Cessation Consul- internationally known,it is principally a patients and staff, whereas there have tation Service. The last service is an regional medical center, with over 80% been only a few written complaints. important resource for assessing and of the patients coming from within a People smoking outside the buildings counseling smokers and determining 500-mile radius. Large numbers of pa- on the fringes of the medical center the best course of smoking-cessation tients receive primary care at Mayo. grounds have led to complaints by therapy for the patient. If all three are The consultant and paramedical staffs nbighboring businesses. Obviously, not in agreement,then no exception to are large, diverse, and loyal, but they people are free to smoke in most public the policy is made. Ultimate responsi- are similar to those in other medical areas away from the medical center,but bility for enforcing the policy with the institutions and they express their opin- aA employees were reminded to be patient rests with the primary medical ions about various policies just as freely considerate of others and conscious of or surgical service,with backup by hos- as employees anywhere else. the fact that their actions at all times pital administration as necessary. Perhaps.the best measure of the reflect on the overall image of the The exclusion of the inpatient psychi- transferability of the Mayo experience institution. atry and chemical dependence units is the large number of inquiries (from No employee has resigned because of from complete adherence to the policy hospitals and clinics) that have been the policy,although it was and contin- has caused some difficulty. Since the received regarding the model that we ues to be personally difficult for many implementation of the policy, pilot pro- used. Many of the institutions that who choose to continue to smoke away grams have been conducted in the Ado- requested the information have imple- from the medical center.Anecdotally,it lescent Psychiatric Unit and the Ado- mented a policy based on the principles has been reported that many employees lescent Chen-deal Dependence Unit,and outlined herein.The Mayo experience is have used this as a motivating factor to now both units are smoke free.The posi- included as a case study in a manual attempt to stop smoking. tive impact of the policy itself helped to prepared by the American Hospital As- One of the most difficult areas is the change these areas that were initially sociation for hospitals that wish to im- enforcement of the policy for hospital- excluded. plement a smoke-free policy.'°We Ion- ized patients.Many patients who smoke CONCLUSIONS clude that the principles followed in the have a difficult time adhering to the poli- implementation of the smoke-free poli- cy during this time of crisis and fre- If the proper groundwork has been cy at Mayo are generalizable and appli- quently request to be allowed to smoke. laid in planning,the implementation of a cable to most other medical institutions Initially, records were kept of each pa- smoke-free policy can go smoothly. Pa- in the United States. tient's problem (patient smoking in tients, staff, and visitors seem to be room, nicotine withdrawal, and non- accepting the policy and abiding by it. References 1. Hansen AS:Park Nicollet Medical Center amok- Health Care:A Guide to Establish Smoke-Free 8. Action plan to encourage union support for a ing policy:Practicing what we preach,Bull Park Health Care Fhcilities.Minneapolis,University of smoke-free hospital.Minn Med 1987;70:654-655. Nicollet Med Found 1985;29:105-113. Minnesota,1986. 9. Hurt RD,Offord KP,Hepper NGG,et al:Long- 2. Rosenstock IM,Stergachis A,Heaney C:Evalu- 5. Burtaine J,Slade J:The smoke-free hospital.NJ term follow-up of persons attending a community- ation of smoking prohibition policy in a health main- Med 1988;85:143-145. based smoking-cessation program.Mayo Clin Proc tenance organization. Am J Public Health 6. Kottke TE,Hill C,Heitzig C,et al:Smoke-free 1988;63:681-690. 1986;76:1014-1015. hospitals: Attitudes of patients, employees, and 10. Smoking and Hospitals. ..Are a Bad 3. Welty TK,Tanaka ES,Leonard B,et al:Indian faculty.Minn Med 1985;68:5355. Match!Policy Development and Implementation health service facilities become smoke-free.JAMA 7. Hammond SC,DeCenzo DA,Bowers MH:How Strategies for a Smokefree Environment.Chicago, 1987;258:185. one company went smokeless.Harvard Business American Hospital Association,1988. 4. Knapp J,Silvis G,Sorensen G,et al:Clean Air Rev,November/December 1987,pp 44-45. JAMA,Jan 6,1989—Vol 261,No.1 A Smoke-free Medical Center—Hurt at al 97 I Editorials The Silver Anniversary On Wednesday,Jan 11, 1989, I will release the next Surgeon With respect to patients who smoke,each physician should General's report on the health consequences of smoking.That use every"teachable moment"to advise them—repeatedly— date is the 25th anniversary of the day on which the historic to quit.'A minute or two of counseling by physicians may be all report of the Surgeon General's Advisory Committee on that is necessary for many patients to quit.'Physicians should Smoking and Health was released to Surgeon General Luther obtain, display, and use materials that promote cessation, L. Terry, MD. This anniversary report will review and ana- which are available from many sources.` They should not lyze the major developments related to smoking and health display publications that accept cigarette advertising in their that have occurred in the United States during the past quar- offices and clinics.'Physicians should be aware of smoking- ter century. cessation programs in their communities, to which they Tremendous gains have been made.The prevalence of ciga- should refer patients who may need special assistance in rette smoking among adults has fallen steadily since 1964.Per quitting. The Public Health Service maintains a list of state capita cigarette sales have declined each year since 1973.Even and local programs on smoking and health that is available total cigarette sales have decreased about 2%each year dur- from the Centers for Disease Control's Office on Smoking and ing the past six years,despite growth in population.'Smoking Health in Rockville,Md. Nurses,physicians'assistants,and is no longer socially acceptable, and is prohibited in more other office staff should also be made part of these nonsmok- public places than ever before. ing efforts. Physicians and other health care providers,medical organi- Physicians should contact their elected officials to urge zations,voluntary health agencies,public health officials,and support for policies designed to discourage tobacco use. The antismoking activists should rejoice in these accomplish- American Medical Association has produced an attractive ments—but only for a moment. When we remind ourselves Physician Leadership Kit, "Creating a Tobacco-Free Soci- that 50 million Americans continue to smoke, and that more ety,"which provides suggestions and materials to help physi- than 300 000 continue to die each year of smoking-attributable cians work for the adoption of these policies.These materials disease,we remember why we can never let ourselves become include"obituary"postcards,modeled after cards developed complacent. by the British Medical Association,for physicians to mail to Despite declining sales,the cigarette industry remains one their Congressmen to inform them of a constituent's death of the most profitable and powerful businesses in America. It caused by tobacco use and to solicit their support for legisla- uses its vast economic strength to defend the promotion,sale, tion that would ban tobacco advertising.'Similar postcards and use of tobacco and to punish those who stand in its way. are also available from the Coalition on Smoking OR Health Last April, RJR Nabisco Inc pulled an $80 million account (Washington,DC),the American College of Chest Physicians from the advertising agency that helped Northwest Airlines (Park Ridge,Ill),and Doctors Ought to Care(Augusta,Ga).If promote its new ban on smoking aboard domestic flights. "If all physicians join together to educate their patients and their we're attacked, we're not going to roll over and play dead," elected officials about the deadly toll that tobacco inflicts on warned the vice president of the Tobacco Institute. "The our society,the progress we are making will accelerate. sooner our adversaries,friendly or otherwise,learn that,the But as we move forward in the United States,we must not less difficulty they're going to find themselves in"(Washing- forget the rest of the world.Multinational tobacco companies ton Post,April 7,1988,pp E 1,E4). are aggressively marketing their products abroad,particular- I believe that the collective influence and moral strength of ly in developing countries, as their markets shrink in the the medical profession, if fully tapped, can overcome the industrialized world.'"These companies,based in the United tobacco industry's attempts to maintain this nation's addiction States and in the United Kingdom, are exporting disease, to nicotine. By providing patients who smoke with constant disability, and death to the people of developing countries, encouragement and assistance to quit,and by urging elected who have little or no knowledge about the dangers of smoking. officials to support tobacco-control policies, physicians can How will they meet the health costs in the future?We have an make an enormous difference. obligation to help developing countries learn from the mis- 98 JAMA,Jan 6,1989—Vol 261,No. 1 Editorials takes that "developed" countries have already made. The Physician,DHHS publication 862178.Dept of Health and Human Services, World Health Organization,celebrating its 40th birthday this National Institutes of Health,1986. 3. Russell MAH,Wilson C,Taylor C,et al:Effect of general practitioners' year,is to be commended for its strong educational program, advice against smoking.Br Med J 1979;2:231-235. C.Everett Koop,MD,SeD 4. Davis RM:Uniting physicians against smoking:The need for a coordinated Surgeon General national strategy.JAMA 1988;259:2900-2901. 5. Richards JW:A positive health strategy for the office waiting room.NY Public Health Service State J Med 1983;83:1358-1360. Rockville,Md 6. AMA insights.JAMA 1987;258:455. 7. Davis RM: Promotion of cigarettes in developing countries. JAMA 1986;255:993. 1. The Health Consequences of Smoking:Nicotine Addiction,a report of the 8. Chandler WU:Banishing 7bbacco,Worldwatch paper 68.Washington,DC, Surgeon General, DHHS publication 88-8406. Dept of Health and Human Woridwatch Institute,1986.. Services,1988,Appendix A. 9. Nath UR: Smoking: Third World Alert. New York, Oxford University 2. Clinical Opportunities for Smoking Intervention:A Guide for the Busy Press,1986. Routing a Modern Pied Piper of Hamelin In reviewing the current substance abuse literature, it be- behavior—it is part of being an adolescent!Thus, carefully comes evident that our children are using drugs at an alarm- developed advertisements can lure our children into tobacco- ingly high rate.'Illicit drugs are being used at earlier ages and use behavior, then into addiction and perhaps a premature by more youths than ever before.Of the 23 million individuals death,while the tobacco industry continues to reap enormous in the United States aged 12 through 17 years,3.5 million use fiscal returns. It is amazing that we sit back and allow this tobacco products(mostly cigarettes),nearly 3 million smoke modern-day Pied Piper of Hamelin to lure our children away marijuana, 1 million use various stimulants, and 6 million from their good health and to their eventual self-destruction. consume alcohol.'The legal"gateway"drug,tobacco,is used Health care professionals must recognize this danger and help daily by more youths than any other drug:20%of adolescent rescue our children. girls and 16%of boys smoke cigarettes daily,and 12%smoke In this issue of THE JOURNAL, Altman et al' take steps over half a pack per day. toward addressing this problem—they promote worthwhile suggestions to help restrict the access of teenagers to tobacco. In their study of 412 retail stores where cigarettes could be See also p so. purchased, they were able to reduce the sale of tobacco to minors by 39%.A major part of this study was an attempt to educate the local community that selling cigarettes to minors The numerous negative health consequences of tobacco are was not only harmful but illegal.This was not a perfect study, well known,yet our youths continue to be seduced into tobac- yet their interventions represent a step in the right direction, co addiction, and our society, incredibly, tolerates this mas- and the community intervention model is worthy of further sive, potentially suicidal behavior. We allow a multibillion- study. I would like to review some of the shortcomings and dollar advertising campaign to flourish, fueled by the then the positive aspects of this important research effort. American tobacco industry and Hollywood—a campaign that The investigation described was a small-scale study,and it glamorizes tobacco use and irresistibly encourages it among will not be easy to reproduce on a larger scale. The fact that our children.' our society acts as if cigarette smoking is not really dangerous Tobacco abuse by children is a story(unfortunately nonfic- makes any community-wide intervention very difficult to tional)worthy of our best writers.Lonely or inhibited adoles- implement. In my view,little will be accomplished until soci- cents are cleverly encouraged by brilliantly designed adver- ety realizes that all our children,not just those in our nuclear tisements and media "entertainments" to use tobacco: the families,need help to resist the lure of tobacco. promise is adventure, happiness, and improved well-being; There were some methodologic problems with the study. the reality is addiction to one of the most powerfully addictive For example,teenagers could easily he about their age when substances known to mankind.The tobacco industry seems to challenged by the store cashier,and judging age by facial and ignore the official warnings of health care professionals,send- body appearance is difficult. In addition, efforts to prevent ing the message to our youth to just relax and light up or cigarette purchases can easily be circumvented. For exam- chew! Our youngsters are easily drawn into such high-risk ple,teens could identify other teens who work in the stores; JAMA,Jan 6,1989—Vol 261,No.1 Editorials 99 , r these youths would likely be under tremendous peer pressure experts.No doubt,some"experts"will argue for the rights of to sell cigarettes to their friends. Teenagers may also state tobacco producers and abusers. I ask, however, what about that they are buying cigarettes for their parents or other the right of our youth to a healthy,tobacco-free future?We adults,and youths could easily find a store not targeted in this must not remain a nation that fails to wage full-scale war study.Although the youngsters involved in the study claimed against drugs—especially when they involve our children. that some stores became "tougher," this was not proved— Those who contribute to children's tobacco and drug use they may have had a strong desire to alter their request should be punished with costly penalties,and strict enforce- approach so the study results would show a benefit. ment should be ensured. A license fee imposed on vendors Certainly the best method(s)of educating consumers and could provide fiscal coverage for strict monitoring of sales. store owners remains in question. As the authors note, the Cigarettes,alcohol,and other drugs should not be sold to our time spent with these individuals was probably too limited, children! yet how much time and the ideal methods of education remain Is such control of drug sales possible in the"real"world?I to be determined. The selection of a six-month period be- think the answer is definitely yes. However, we must act tween the pretest and post-test intervals is arbitrary. aggressively to reverse the current situation of easy access to While this study is not the final word in solving the problem drugs for minors.Health care professionals,lawmakers,busi- of illegal tobacco sales to youngsters,it clearly is a welcome nesspersons, legal experts, educators, and others who are step.The authors'efforts extending the work of others in this concerned about the future of our society must work together fields could move us a bit closer to our goal of a tobacco-free to bring about the reality of a drug-free(including a tobacco- youth:Their research is grounded in important literature that free)youth.'Tobacco industry profits should be used to keep documents that if we can delay the onset of tobacco(and other our children away from, not lure them toward, tobacco use, drug) experimentation from early to late adolescence, the and these companies should be encouraged to redirect their result will be fewer adults addicted to tobacco.Reducing easy efforts toward this goal. We all must become involved to access to tobacco along with educating youth (and society) promote better health and a better future for our children. about its many dangers is a good beginning! Donald E.Greydanus,MD Part of the success and importance of this study lies in the Des Moines involvement of various community organizations, including 1. Shearin RB,Jones RL:Drug and alcohol abuse:Medical and psychosocial the American Cancer Society, American Heart Association, aspects,in Hofmann AD,Greydanus DE(eds):Adolescent Medicine,ed 2.East American Lung Association, a local hospital, local charities, Norwalk,Conn,Appleton-Century-Crofts,1989,pp 401-430. and others. Diverse expertise can he found in such groups, 2. Holleb AI: Smoking, the ticking time bomb for teenage girls. Cancer 1981;31:44. expertise needed to achieve positive results. The active par- 3. GreydanusDE:Risk-takingbehaviorsinadolescence.JAMA 1987;2M:2110. ticipation of various experts and the media will be necessary 4. Altman DG,Foster V,Rasenick-Douse L,et al:Reducing the illegal sale of to launch an aggressive,cost-efficient education campaign. cigarettes to minors.JAMA 1989;261:80-83. Altman et al also provide thoughtful suggestions on how the 5. DiFranzaJR,Norwood BD,Garner DW,et al:Legislative efforts to protect children from tobacco.JAMA 1987;257:3387-3389. sale of tobacco to minors can be reduced. Of course, these 6. Council on Scientific Affairs:Health effects of smokeless tobacco.JAMA strategies will need review by health, legal, and business 1986;255:1038-1044. 100 JAMA,Jan 6,1989—Vol 261,No.1 Editorials 1 From the Centers for Disease Control .■. Leads From the Morbidity and Mortality Weekly Report Atlanta, Ga State Tobacco Prevention, Control Activities: Results of 1989-1990 Association of State, Territorial Health Officials Survey-Final Report MMWR. 1991; 40: Recommendations forts among targeted populations. Ad- nac was used to supplement informa- and Reports-11(2.4 tables,2 figures omit- ditional sources either supplemented or tion supplied by the states.'Finally,pre- ted) validated state information on tobacco- viously tabulated data were reviewed use control data collected through the and updated by the ASTHO network in Summary ASTHO survey. In the fall of 1989, December 1990. In October 1989,the Association of State ASTHO established a network of health and Territorial Health Officials professionals responsible for communi- Data Collection and Analysis (ASTHO) conducted a survey of state cation between the federal government ASTHO sent the questionnaires to all health department personnel regarding and state health departments on issues 50 states and the District of Columbia. programs,policies,and public health sys- related to tobacco-use prevention and For the purpose of this report,the Dis- tems that stress the prevention and con- control.As the identifiable contacts for trict of Columbia is considered a state trol of the use of tobacco. This survey information transfer on tobacco-related when summary data are presented. In provided detailed data associated with matters, these persons served as re- some cases, supplemental information state tobacco-use control programs and spondents to the ASTHO survey. was obtained by telephone. Responses their essential components (e.g., bud- were tabulated and analyzed using True gets, planning, coalitions, surveillance METHODS Epistat and dBase IV.6,6 systems, smoking cessation programs, The survey's 10 major sections are 1) RESULTS educational activities,legislative actions, background information on tobacco and and health department policies).States tobacco control;2)adult tobacco-use sur- The response rates were 100% for vary widely in the strength and cover- veillance;3)adolescent tobacco-use sur- both the main section and educational age of their programs for preventing veillance; 4) reporting and analysis of sections. and controlling tobacco use.The ASTHO data on the impact of tobacco-related survey data may be used to help plan disease; 5) regulatory activities; 6) co- Background Information and evaluate state health department alitions against tobacco use; 7) special As of October 1990,12 states had devel- programs as part of an effort to prevent populations; 8) community information oped a specific freestanding plan for pre- chronic diseases related to tobacco use. on education activities;9)economic in- venting and controlling tobacco use.In 22 Outcomes of state activities may be eval- centives,deterrents;and 10)educational states,the plan is a part of another plan uated through surveys such as CDC's institutions. ASTHO contacts solicited for controlling chronic disease. Most of Behavioral Risk Factor Surveillance the help of state departments of educa- these plans address areas related to high- System(BRFSS)and the Current Pop- tion to answer questions about tobacco- risk populations, health care, smoking ulation Survey (CPS) of the Bureau of use control activities in educational in- cessation issues, worksite policies, and the Census.Future surveys of state ac- stitutions (public and private schools). other areas in preventing tobacco use. tivities for controlling the use of tobacco This section of the survey assessed the The 12 freestanding plans were all pub- may be included in the evaluation of the ability of each state to measure progress lished after 1980,and most after 1985.' upcoming(1993)American Stop Smok- toward smoke-free schools and the ex- Excluding California, the average ing Intervention Study(ASSIST),which tent to which educational institutions state budget devoted to tobacco-related is cosponsored by the National Cancer addressed antitobacco education. health activities was$70 917. The state Institute and the American Cancer Central data sources were used to sup- funds ranged from no funds (27 states) Society. plement the survey results for the fol- to $151 million in California, where a INTRODUCTION lowing areas of this report: legislative portion of the state cigarette excise tax activities; taxation; and number of is earmarked for health activities.' In The Association of State and Terri- schools,districts,and enrolled students. addition to California, six other states tonal Health Officials (ASTHO) con- Sources used to supplement informa- had earmarked a portion of the excise ducted a survey in October 1989 to as- tion on legislative issues included State cigarette tax for public health activi- sess progress among the states in the Legislated Actions on Tobacco Issues of ties.Additional funds,including grants, public health practice of preventing and the Tobacco-Free America Project'and cooperative agreements,and in-kind ser- controlling tobacco use.The survey was Major Local Smoking Ordinances in the vices, averaged $54 230 per state (in- also conducted to.provide states with United States, National Institutes of cluding California). incentives to create and implement ef- Health.I The Tobacco Institute also pro- The sixteen states growing tobacco forts to control tobacco use.The survey vided state-specific data on taxation.' produced a combined total of covered several components of effective For information related to schools and $2 381 000 000 in tobacco agricultural rev- state programs that address such ef- school districts, the 1990 World Alma- enue in 1989, representing 1.5%of the JAMA, December 11, 1991—Vol 266, No.22 From the CDC 3105 I W total U.S. agricultural farm receipts.' bidity,and Economic Costs(SAMMEC), Restrictions on Tobacco Adverfising The percentage of state agricultural farm to obtain data on smoking-attributable Two states(Massachusetts and Utah) receipts generated by tobacco growing deaths and economic costs." have policies that restrict advertising of ranged from 0.2% (Missouri and Wis- In five states, a record of the dece- tobacco products on state property or consin)to 21.8%(Kentucky). dent's smoking history was required on property under the state's jurisdiction. death certificates.Four states reported Local policies in six states(Arizona,Cal- Surveillance of Adult Tobacco Use data on smoking-attributable hospital ifornia, Colorado, Hawaii, Massachu- CDC's BRFSS is a telephone-based discharges,and eight states have infor- setts, and Nebraska)restrict advertis- system that collects yearly data on to- mation on smoking-attributable state- ing of tobacco products on local govern- bacco use and other health-related be- funded medical care costs. In 33 states, ment property, such as buses, transit haviors among adults 18 years of age maternal smoking history was recorded stations, or sports facilities. and older. In 1990, 46 states partici- on birth certificates. pated in the BRFSS.10 Twenty-one Tobacco-Control Coalitions states collected data on adult smoking REGULATORY ACTIVITIES As of October 1990,50 states had to- prevalence from non-BRFSS sources. Smoking in Public Places bacco-related working groups or coali- Twenty of these states collected data on tions of individuals or agencies concerned adult special target populations(blacks, In 1989,45 states had laws restricting with preventing and controlling tobacco Hispanics, Asians/Pacific Islanders, smoking in public places;in 38 of these use.17 The coalition members represent American Indians,persons with low so- states, the restrictions also applied to the health professions,the general com- cioeconomic status, and women of re- public-sector workplaces. In 17 states, munity, groups concerned with legisla- productive age (15-44 years old). these restrictions extended to private- tion and policy,and educational groups. In addition to the BRFFS,state-spe- sector workplaces.' The Surgeon Gen- Eighty-two percent of these state coa- cific data on tobacco use among adults eral's 1989 report on smoking and health litions carried out public education and 16 years of age and older are available defined extensive regulations as those information activities, 72% addressed from two Current Population Surveys that restricted smoking in the private- legislative efforts, 48% educated pro- (CPS)that were performed by the U.S. sector workplace.15 fessionals,44%worked on developing a Bureau of the Census in 1985 and Local smoking ordinances in cities and plan for tobacco-use control, and 26% 1989.11,11 The 1985 CPS provided state- counties encompassed a wide range of carried out research and evaluation. specific estimates of both smoking prev- public settings, including restaurants, The average funding for coalitions in alence and smokeless-tobacco use. The elevators, hotels, libraries, museums, reporting states (excluding California) 1989 survey provided information only retail stores,schools,public transit,and is $5 536. on smoking prevalence. other enclosed public places. In all,490 local ordinances restricted or prohibited Special Populations Surveillance of Youth Tobacco Use smoking in public places. Special populations targeted for in- No national system exists for moni- Health Department Tobacco-Use tensive tobacco-use prevention and con- toring state-specific tobacco use by ad- trol efforts by the U.S. Department of olescents.However,CDC has developed Policies Health and Human Services include ad- a standard survey(the Youth Risk Be- With the exception of North Carolina olescents, women of reproductive age havior Survey[YRBS])to collect com- and Virginia, all state health depart- (15-44 years old),Asians/Pacific Island- parable school-based data from the ments had a written policy on smoking ers, American Indians, Hispanics, and states.13 By the completion of the sur- in state health department buildings in blacks.18 Forty states have programs vey in January 1990, three states had 1989.Twenty-four(47%)of these states (in addition to the BRFSS)that include participated in the YRBS;19 additional completely banned smoking in state education and information for some or states had participated by the end of health department facilities; 31 states all of these groups. Thirty-three states 1990.From 1986-1990,32 states reported (61%) permit the sale of tobacco prod- had cessation programs, and 26 states collecting data on tobacco use among ucts in health department buildings. collected behavorial data on these high- adolescents from sources other than the Restrictions on Minors' Access to risk populations. YRBS.The respondents were asked fol- Tobacco Products low-up questions to determine if these INFORMATIONIEDUCATION surveys covered the basic question top- As of October 1990,46 states prohib- public Information Activities ics from the YRBS. The surveys exam- ited the sale of tobacco products to un- ined such specific areas as tobacco ex- deraged persons.The minimum age for Twenty-two state health departments perimentation,current tobacco use,age purchasing tobacco varied from 16 years produced public service announcements of initiation of tobacco use, and smoke- of age(Kentucky,Virginia)to 19 years designed to prevent tobacco use.Forty- less tobacco use.Twenty-six states had of age (Alabama, Alaska, Utah); the five states used public service announce- information on experimentation with to- most common minimum age is 18 years ments produced by federal agencies bacco-use, 32 states collected data on of age (37 states).1.16 Nine states re- (such as the Office on Smoking and prevalence of tobacco use,19 states had 'stricted the placement of vending ma- Health and the National Cancer Insti- information on age of initiation of to- chines that contain tobacco products; tute).Thirty-two states initiated public bacco use,and 25 states had information one state(Colorado)banned the sale of information campaigns in their states on smokeless tobacco use. smokeless tobacco in vending machines, within the last 2 years.These campaigns and another(Utah)banned the sale of all used various forms of media(billboards, Tobacco-Related Disease Impact tobacco products in vending machines. radio, television, etc.). Data-Reporting and Analysis Twenty-two states required a state- All 51 state health departments used issued retail tobacco license for vendors Smoking Cessation Programs a software package developed by the selling tobacco products. The fees for Thirty-five states offered smoking Minnesota Department of Health, the these retail licenses ranged from$0 to cessation programs to state health Smoking-Attributable Mortality, Mor- $250(average:$33). employees, and 26 states offered such 3106 JAMA,December 11,1991—Vol 266,No.22 From the CDC 1 programs to members of the commu- Among 12 states that provided infor- among eight states to reduce tobacco nity. mation, approximately 2.8 million use and chronic diseases.21 Economic Incentives and Deterrents (48.5%) public primary and secondary The ASTHO survey is an important students attended smoke-free schools. baseline for monitoring tobacco-use.con- Colorado, Kansas, and Washington DISCUSSION trol programs at the state and local lev- were the only states that had health els. Future surveys may provide data benefits packages with differential rates The. 1989 ASTHO survey provides that can be used to measure the effects for smokers and nonsmokers for state data on the activity of all 51 states re- of planned intervention programs,such government employees.Fourteen states garding the prevention and control of as the National Cancer Institute's reported having third-party payers of tobacco use. States varied greatly in ASSIST, which will begin in 1993.20 medical care that offered differential their approaches to the control of to- These surveys may also provide means rates to consumers, and seven states bacco use. Some states had extensive to measure progress toward the year had third-party payers of medical care surveillance systems and programs in 2000 health objectives for the nation.18 that offered reimbursement for treat- place, whereas others had only limited ment of tobacco addiction. These data programs and funding. Data from the Appendix may be an underestimate,however,be- 1989 ASTHO survey and subsequent Tobacco-related objectives, outlined cause some large national insurers sold surveys may be linked to state-specific in Healthy People 2000:National Health policies in many states,(e.g.,Blue Cross data on smoking prevalence, cigarette Promotion and Disease Prevention Ob- and Blue Shield Company of Southwest- consumption, and smoking cessation. jectives,provide a national guide for as- ern Virginia).19 State tobacco excise These state-specific data (from CDC's sessing progress in preventing and con- taxes ranged from 2 cents per pack in BRFSS and the Bureau of the Census' trolling tobacco use. The following is North Carolina to 41 cents per pack in CPS)may be used to assess the outcome reprinted from Healthy People 2000.11 Texas.3 The average state excise tax of recent state activities in preventing The Year 2000 Objectives for the Na- collected per pack was 23 cents. The and controlling tobacco use. A national tion call for the following tobacco-re- lowest tax rates were primarily in the guide that may direct state progress in lated objectives: tobacco-producing states. these and other areas of concern is 3.1 Reduce coronary heart disease Healthy People 2000:National Health deaths to no more than 100 per 100 000 Educational Institutions Promotion and Disease Prevention Ob- people. Thirty-nine states had state laws that jectives, which lists 16 tobacco-related 3.2 Slow the rise in lung cancer deaths restricted tobacco use in schools. objectives for the year 2000.18 to achieve a rate of no more than 42 per Twenty-seven states banned smoking Little information about programs and 100 000 people. for students; only eight states banned policies to prevent tobacco use among 3.3 Slow the rise in deaths from smoking for both students and staff. In young persons is available either to the chronic obstructive pulmonary disease 16 states, the state department of ed- ASTHO tobacco-control network or to to achieve a rate of no more than 25 per ucation reported having formal policies state departments of education. Fewer 100 000 people. on tobacco use in schools. Only two than half of the states reported any in- 3.4 Reduce cigarette smoking to a states, Ohio and Nevada, provided in- formation related to the education por- prevalence of no more than 15%among formation on private primary and sec- tion of the survey. Consequently, se- people aged,20 and older. ondary schools. lective reporting from certain states may 3.5 Reduce the initiation of cigarette There are 15323 school districts in overstate the percentage of smoke-free smoking by children and youth so that the United States.4 Among the 25 states schools. In addition, those states that no more than 15%have become regular reporting information on policies in reported 100% prevalence of tobacco- cigarette smokers by age 20. school districts, 2 311 (30.8%) of the use education assumed,but did not ver- 3.6 Increase to at least 50 percent school districts in these states banned ify, total compliance to state require- the proportion of cigarette smokers aged smoking for both students and staff. ments. Efforts to collect these data are 18 and older who stopped smoking cig- Among the states with information important in assessing the overall pub- arettes for at least one day during the on smoking policies in public primary he health approach to preventing and preceding year. schools, 4 468 (33.9%) of these schools controlling tobacco use. 3.7 Increase smoking cessation dur- banned smoking for both students and Because the 1989-90 ASTHO survey ing pregnancy so that at least 60% of staff.Among the reporting states,21097 provided baseline information on broad women who are cigarette smokers at (96.2%n)schools completely banned smok- activities to prevent and control tobacco the time they become pregnant quit ing for students(i.e.,students could not use, subsequent surveys may be useful smoking early in pregnancy and main- smoke on school grounds). Within the in assessing states' progress. To con- tain abstinence for the remainder of their 26 states that provided data on tobacco- duct such assessments,state-specific ob- pregnancy. use education, 18 588 of 21129 (87.9%) jectives should be established,and a sys- 3.8 Reduce to no more than 20%the public primary schools taught tobacco- tem for measuring states' progress in proportion of children aged 6 and use prevention. these objectives should be implemented. younger who are regularly exposed to Among the states reporting infor- The evaluation could initially be applied tobacco smoke at home. mation on smoking policies in public to the different control activities cov- 3.9 Reduce smokeless tobacco use by secondary schools, 1368 (21.2%) of ered by the survey(such as education, males aged 12 through 24 to a preva- 6459 schools completely banned coalitions,and surveillance).An overall lence of no more than 4%. smoking for both students and staff, measure for each component for con- 3.10 Establish tobacco-free environ- and 7481 (83.1%) completely banned trolling and preventing tobacco use ments and include tobacco use preven- smoking for students. For the 23 should then be developed. The Rocky tion in the curricula of all elementary, states that provided information on Mountain Tobacco-Free Challenge has middle, and secondary schools, prefer- tobacco-use education, 7 623 of the included an evaluation ofstate activities ably as part of quality school health ed- 9456 public secondary schools (80.6%) on tobacco-use control.20 Initiated in ucation. taught tobacco-use prevention. 1988, this program is a regional effort 3.11 Increase to at least 75%the pro- JAMA,December 11, 1991—Vol 266, No.22 From the CDC 3107 i portion of worksites with a formal smok- References learned and directions for the states:n the 1990s.Annu Rev Public Health(in press). ing policy that prohibits or severely re- 1.Tobacco-free America: state legislated actions on 13.Harel Y,Kann L,Collins J,Kolbe L.Implement-' ' stricts smoking at that particular work- tobacco issues.Washington,DC:Legislative Clearing- ing the Youth Risk Behavior Surveillance System:a house,1990. progress report. In: Agenda of the Fifth National place. 2.National Institutes of Health.Major local smoking Conference on Chronic Disease Prevention and Con- 3.12 Enact in 50 States comprehen- ordinances in the United States:a detailed matrix of trol:From 1990 to 2000. Detroit,Michigan,October S1Ve laws On Clean indoor air that pro- the provisions of workplace, restaurant, and public 1990:47. places smoking ordinances.Washington,DC:US De-- 14.CDC.Smoking and health:a national status report: hibit Or strictly limit smoking in the partment of Health and Human Services, Public a report to Congress.2nd ed.Rockville,Maryland:US workplace and enclosed public places(in- Health Service, National Institutes of Health, 1989; Department of Health and Human Services, Public cludin health care facilities Schools (DHHS)publication no.(NIH)90-479. Health Service,1990;DHHS publication no.(CDC)87- g , , 3. Tobacco Institute. Cigarette Tax Data. Monthly 8396, and public transportation). state cigarette tax report,July 1990.Washington,DC: 15.CDC.Reducing the health consequences of smok- 3.13 Enact and enforce in 50 States Tobacco Institute,1990. ing:25 years of progress.Report of the Surgeon Gen- 4.The World Almanac 1990. New York:Newspaper eral.Rockville,Maryland:US Department of Health laws prohibiting the sale and distribu- Enterprise Association,1990. and Human Services, Public Health Service, 1989; tion of tobacco products to youth younger 5.True Epistat.Epistat Services.3rd Edition.Rich- DHHS publication no.(CDC)89-8411. ardson,Texas:True Epistat,1989. 16. CDC. State laws restricting minors' access to than age 19. 6.DBASE IV.Asthton-Tate Corporation.Torrance, tobacco.MMWR 1990;39(21):349-53. 3.14 Increase to 50 the number of California,1988. 17.CDC.Progress in chronic disease prevention.State States with plans to reduce tobacco use, 7. CDC. State tobacco-use prevention and control coalitions for prevention and control of tobacco use. plans.MMWR 1990;39(8):133-6. MMWR 1990;39(28):476-85. especially among youth. 8. Bal DG, Kizer KW, Felten PG, Malar HN, Nie- 18. US Department of Health and Human Services. 3.15 Eliminate or severely restrict all meyer D.Reducing tobacco consumption in California: Healthy people 2000: national health promotion and 5' development of a statewide anti-tobacco use campaign. disease prevention objectives. Washington, DC: US forms of tobacco product advertising and JAMA 1990;264:15704. Department of Health and Human Services, Public 9.US Department of Agriculture.Tobacco situation Health Service,1990;DHHS publication no.(PHS)90- promotion to which youth younger than and outlook report. Washington, DC: US Depart- 50212. the age of 18 years are likely to be ex- ment of Agriculture, Commodity Economics Divi- 19.Wroblewski M.Insurance incentives for not smok- posed. sion, Economic Research Service, 1990 publication ing.NY State J Med 1985;85:309. no. TS-213. 20. Novotny TE,Thomas WI. The Rocky Mountain 3.16 Increase to at least 75 percent 10.CDC.Behavioral Risk Factor Surveillance,1988. Tobacco-Free Challenge:year number two:evaluation the proportion of primary care and oral MMWR 1990;39:1-21. report 1990.Rockville,Maryland:US Department of 11.Marcus AC,Shopland DR,Crane LA,Lynn WR. Health and Human Services,Public Health Service, health care providers Who routinely ad- Prevalence of cigarette smoking in the United States: Office on Smoking and Health,1990. vise cessation and provide assistance and estimates from the 1985 Current Population Survey. 21. CDC. State-based chronic disease control: the followlz for all of their tobacco-using JNCI 1989;81:409-14. Rocky-Mountain Tobacco-Free Challenge. MMWR 11 g 12.Novotny TE,Romano RA,Davis RM, Mills SL. 1989;38:749-52. patients. The public health practice of tobacco control:lessons Cigarette Smoking Among Youth- 1989 MMWR.1991;40:712-715(1 table omit- NHIS households and were adjusted to the past month was lower among His- ted) provide national estimates. Confidence panic(11.7%)than among non-Hispanic IN 1988, an estimated 434 000 persons intervals(CIs)were calculated by using (16.1%)youth,the prevalence of smok- in the United States died as a result of the Software for Survey Data Analy- ing in the past week was similar in each cigarette smoking.'About three fourths sis.' Participants were asked the fol- group (9.3% and 11.8%, respectively). of adults who have ever been regular lowing questions about cigarette smok- Prevalence of smoking in the past month cigarette smokers reported trying their ing behavior: "Think about the last 30 and in the past week increased directly first cigarette before their 18th birth- days. On how many of these days did by age. day(National Institute on Drug Abuse you smoke?"and"Now,think carefully Among youth 17-18 years of age, the [NIDA], unpublished data), and about about the last SEVEN days. Did you prevalence of smoking during the pre- half of them had become regular smok- smoke cigarettes on any of THOSE vious week was substantially higher ers by that time' (NIDA, unpublished days?" among dropouts(43.3% [95% CI=plus data). This report, based on the Teen- Respondents who were still in school or minus 4.9%])than among school st- age Attitudes and Practices Survey or who had already graduated from high tenders/HS graduates (17.1% [95% (TAPS),presents the prevalence of self- school were classified as"school attend- CI=plus or minus 1.7%]).Among school reported smoking among U.S. adoles- ers/high school (HS) graduates." Re- attenders/HS graduates,the prevalence cents aged 12-18 years during 1989. spondents who were not attending school of smoking during the previous week In 1989, the TAPS focused on ado- at the time of the survey and who had was similar by gender (males: 17.5% leseents'knowledge,attitudes,and prat- not completed the 12th grade were clas- [95%CI=plus or minus 2.3%];females: tices regarding tobacco use.The sample sified as "dropouts."Among youth 17- 16.7%[95%CI= plus or minus 2.3%]). described in this report includes all youth 18 years of age, 2355(80.8%)were en- However, dropouts who were male aged 12-18 years who were living in rolled in school,489(16.8%)were drop- (51.7% [95% CI=plus or minus 6.6%]) households.Questionnaires were admin- outs,and 69(2.4%)had completed high were more likely to report having istered by computer-assisted telephone school and were not currently in school. smoked during the previous week than interviewing and mail'(for homes with- Overall, 15.7% of respondents re- were dropouts who were female(33.3% out telephones and for initial nonrespon- ported smoking on 1 or more days dur- [95%CI=plus or minus 6.5%]).Among dents).Adolescents were sampled from ingthe month,and 11.5%reported smok- school attenders/HS graduates, 19.3% households that had participated in the ing on 1 or more days during the week (95%CI=plus or minus 1.9%)of whites second half of the 1988 National Health before the survey. Patterns were sim- and 5.7%(95%CI=plus or minus 2.8%) Interview Survey(NNIS)and the first ilar by gender in all categories, except of blacks reported smoking during the half of the 1989 NHIS. During this pe- among persons 18 years of age.The prev- previous week.Similarly,dropouts who riod, the household participation rate alence of smoking was higher among were white (46.1% [95% CI=plus or was 95%.Data were obtained from 9965 white youth than among black youth. minus 5.2%])were more likely to report (82.4%) of 12097 adolescents in the Although the prevalence of smoking in (Continued on p 3111.) 3108 JAMA,December 11, 1991-Vol 266, No.22 From the CDC !, (Continued from p 3108.) for interventions that focus on both in- ing the 1990s,intensive collaborative ef- having smoked during the previous school and out-of-school youth, The na- forts will be necessary to reduce tobacco week than were dropouts who were tional health objectives for the year 2000 use among U.S. youth. black (17.1% [95% CI=plus or minus have established four relevant targets References 9.3%1)• for this problem: 1.CDC.Smoking-attributable mortality and years of potential life lost-United States, 1988. MMWR Reported by: CW Heath, MD, RD Corcoran, EdD, • establish tobacco-free environments 1991;40:62-3,69-71. American Cancer Society. SL Mills, MD, DR Shop- in all elementary Ii1lddle and S2COnd- 2.CDC.Reducing the health consequences of smoking: land,National Cancer Institute;SE Marcus,PhD, 25 years of progress-a report of the Surgeon Genera]. Na- tional Institute of Dental Research,National Institutes ary schools and include tobacco use pre- Rockville, Maryland:US Department of Health and of Health.JP Pierce,PhD,Univ of California at San vention programs in school curricula(ob- Human Services,Public Health Service,1989;DHHS Diego.Office on Smoking and Health and Div of Ado- publication no.(CDC)89-8411. lescent and School Health,National Center for Chronic jeetive 3.10); 3. Shah BV. Software for Survey Data Analysis Disease Prevention and Health Promotion; Div of • enact and enforce State laws ria- (SUDAAN) version 5.30 (software documentation). Health Interview Statistics, National Center for Research Triangle Park, North Carolina: Research Health Statistics,enc. tionwide prohibiting the.sale and dis- Triangle Institute,1989. tribution of tobacco products to youth 4. CDC. Tobacco use among high school students- . United' ed <19 ears (objective 3.13); 5.NationalInstitut on Drug Abuse.N.MMWR ONational House- this report are consistent with findings • implement state plans nationwide hold Survey on Drug Abuse: population estimates t0 reduce tobacco use eS eCiall 1990•Rockville,Maryland:US Department of Health from three other recent national sur- p Y amon g and Human Services,Public Health Service,Alcohol, veys that measure smoking by youth: youth (objective 3.14); and Drug Abuse,and Mental Health Administration,1991; rates of smokingare similar for males • eliminate or severely restrict all DHHS publication no.(ADM)91-1732. 6.Pirie PL,Murray DM,Luepker RV.Smoking prev- and females and higher for whites than forms of tobacco product advertising and alence in a cohort of adolescents,including absentees, blacks"(J.G.Bachman,L.D.Johnston, promotion to which youth--18 years of dropouts, and transfers. Am J Public Health .GsorP.M.O'MalleY University of age are like) to be exposed (objective 7GereJ. Influence of privacy on self-reported unpublished data,1990).In addition,the 3.15)." drug use by youths.In:Rouse BA,Kozel NJ,Richards LG,eds.Self-report methods of estimating drug use: findings from TAPS confirm previous To help achieve these and other meeting current challenges to validity. Rockville, reports of higher smoking rates among smoking-related objectives, the Public Maryland:US Department of Health and Human Ser- 6 vices,Public Health Service,Alcohol,Drug Abuse,and dropouts and suggest gender and racial Health Service has developed and im- Mental Health Administration,1985.(NIDA research differences in smoking prevalence among plemented several programs. For ex- monograph no.57). dropouts. Differences in overall preva- ample, the National Cancer Institute 8.Converse PE,Traugott nee 986;23Asses:1094-accuracy of polls and surveys.Science 1986;234:1094-8. lence estimates between surveys may and the American Cancer Society have 9.National Institute on Drug Abuse.NIDA capsules: be explained b the mode of data col- recent) ':gets about teenagers and drug abuse. Rockville, p y y initiated the American Stop Maryland:US Department of Health and Human Ser- lection (i.e., household interview vs. Smoking Intervention Study for Cancer vices,Public Health Service,Alcohol,Drug Abuse,and school-based, self-administered ques- Prevention (Project ASSIST) in 17 Mental Health Administration,1991. I 10.National Institute on Drug Abuse.National House- tionnaire), composition of the samples, states.This demonstration project is de- hold Survey on Drug Abuse:highlights 1990. Rock- varying response rates, and the word- signed to disseminate various interven- ville,Maryland:US Department of Health add Human Services,Public Health Service,Alcohol,Drug Abuse, ing of questions.' tions to prevent and stop tobacco use and Mental Health Administration,1991;DHHS pub- Cigarette use among U.S. youth ap- among adults and youth throughout the lication no.(ADM)1789-91. pears to have declined sharply in the nation. CDC provides states with tech- 11. Public Health Service. Healthy people 2000:na- tional health promotion and disease prevention late 1970s and stabilized in the 1980s,9,10 nical assistance to develop and conduct objectives-full report with commentary.Washington, especially among white youth.2 The find- targeted interventions to reduce to- DC:US Department of Health and Human Services, ins from TAPS underscore the need bacco consumption among Public Health Service, 1991; DHHS publication no. g p g youth. Dur- (Pxs)91-50212. Cigarette Smoking Among Reproductive-Aged Women- Behavioral Risk Factor Surveillance System, 1989 MMWR.1991;40:719-723(3 tables omit- ing to 12% among reproductive-aged participating state. To compare smok- ted) women (18-44 years of age).' This re- ing prevalences between states, WOMEN WHO SMOKE cigarettes are port summarizes data from the 1989 Be- weighted state-specific prevalences were at increased risk not only for chronic havioral Risk Factor Surveillance Sys- standardized for the distribution of the diseases (e.g., lung cancer and chronic tem(BRFSS)on the prevalence of smok- 1980 U.S. population by age,race, and obstructive pulmonary disease)but-if ing among reproductive-aged women. educational level. Smoking prevalences they use oral contraceptives-also for In 1989,health departments in 39 par- for subgroups (age, race, educational myocardial infarction.)In addition,cig- ticipating states and the District of Co- level,and pregnancy status)were stan- arette smoking during pregnancy in- lumbia used a standard questionnaire to dardized by adjusting for the other vari- creases the risk for low birth weight and conduct telephone interviews of adults ables. premature infants, miscarriage, still- aged>18 years.'Current smokers were In 1989,weighted crude prevalences birth, sudden infant death syndrome, defined as persons who had smoked at of cigarette smoking among reprodue- and infant mortality.'Because of these least 100 cigarettes and who reported tive-aged women varied from 17% in risks and other health problems associ- being a smoker at the time of the in- Utah to 32% in Kentucky and Rhode ated with cigarette smoking,one of the terview. Individual responses were Island (median: 26.5%). Standardized national health objectives for the year weighted to provide estimates repre- smoking prevalences ranged from 21% 2000 is to reduce the prevalence of smok- sentative of the adult population of each in Texas to 37%in Wisconsin. In gen- JAMA,December 11, 1991-Vol 266,No.22 From the CDC 3111 era],standardized smoking prevalences and Health,National Center for Chronic Disease Pre- than for women aged_-45 years because CDC. were highest in the midwestern states vention and Health Promotion, reproductive-aged women appear to be and lowest in the Rocky Mountain and more willing to attempt to quit smoking. midcentral states. CDC Editorial Note:In this report,the The 1989 BRFSS determined that the Older women and women with less state-to-state variations of smoking median prevalence of current smoking than a high school education were more prevalences among reproductive-aged was 26.5% among reproductive-aged likely to smoke.Pregnant women were women may reflect differences in socio- women in the states surveyed; accord- less likely than nonpregnant women to demographic characteristics(e.g., age, ingly,nearly all states will require con- smoke. Smoking prevalences did not race,and educational level)of state pop- certed efforts to reduce prevalence of vary substantially between white and ulations. However,because these vari- smoking among reproductive-aged black women,the only racial groups for ations persisted after standardization to women to 12% by the year 2000.3 Ef- which rates could be calculated because adjust for these differences, other fac- forts to reduce smoking initiation among the numbers of respondents of other ra- tors(e.g.,occupation,employment sta- adolescent girls and to target young cial/ethnic groups were too small to pro- tus,and family income)may affect state- women for smoking-cessation interven- vide stable estimates. specific smoking prevalences. These tions are important priorities to accom- Among reproductive-aged women variations may also reflect differences plish this objective.',' who smoked,84%smoked fewer than 25 in the intensity of cigarette advertising cigarettes per day. Women aged 35-44 and in the effectiveness of statewide References ears tended to be heavier smokers than " 1•Sharpiro S,Slone D,Rosenberg L,et al.Oral con- y smoking-control interventions. In ad- traceptive use in relation to myocardial infarction. younger women.Approximately 44%of dition,reasons for the lower prevalences Lancet 1979;1:743-7. all women who were current smokers of smokin amon certain ou s could 2.CDC.Reducing the health consequences of smoking: g g p 25 years of progress—a report of the Surgeon General. had attempted to quit smoking(i.e.,quit- include 1) declining smoking initiation Rockville, Maryland: US Department of Health and ting for at least 1 week)in the previous rates in younger cohorts of women (a Human Services,Public Health Service,1989;DHHS year.Women aged 35-44 ears were sub- publication no.(CDC)89-8411.Service. Heal y g Y trend observed previously for white and 3.Public Health Service.Healthy people 2000:national stantiall less likely than younger e 2 decreasing smok- health promotion and disease prevention objectives— full report, with commentary, Washington, DC: US women to have attempted quitting. ing-initiation and increasing Smoking- Department of Health and Human Services, Public cessation rates over time among women Health Service,1991;DHHS publication no.(PHS)91- 212, Reported by the following state BRFSS coordinators: with higher educational levels'; and 3) 4.Remington PL,Smith MY,Williamson DF,Anda L Eldridge, Alabama; J Contreras, Arizona; W the effect of higher smoking-cessation RF,Gentry EM,Hogelin GC.Design,characteristics, Wright, California;M Adams, Connecticut;M Rivo, rates for pregnant women.3 and usefulness of state-based behavioral risk factor District of Columbia; S Hoecherl, Florida;J Smith, The BRFSS findings regarding surveillance: 1981-1987. Public Health Rep 1988; Georgia; A Villafuerte, Hawaii;J Mitten, Idaho; B g g g 103:366-75. Steiner,Illinois;S Joseph,Indiana;S Schoon,Iowa;K amounts of smoking and attempts to quit 5.CDC.The health benefits of smoking cessation:a Bramblett,Kentucky;J Sheridan,Maine;A Weinstein, are COnSlSterit with previous S report of the Surgeon General. Rockville,Maryland:re OrtS.2' Maryland; R Letterman, Massachusetts; J Thrush, I> p US Department of Health and Human Services,Pub- Michigan;N Salem,Minnesota;J Jackson-Thompson, However,the proportion of women who he Health Service, 1989; DHHS publication no. Missouri;M McFarland, Montana;S Spanhake, Ne- attempted t0 quit smoking for at least (CDC)90-8416. braska;K Zaso,L Powers,New Hampshire;M Wat- p q g 6,Escobedo LG,Remington PL,Anda RF.Long-term son,New Mexico;J Marin,0 Munshi, New York;C 1 week in the year preceding the survey secular trends in initiation of cigarette smoking among Washington,North Carolina;M Maetzold,North Da- (44%)was substantially higher than that Hispanics in the United States. Public Health Rep kota;E Capwell,Ohio;N Hann,Oklahoma;J Grant- 1989;104:583-7. Worley,Oregon;C Becker,Pennsylvania;R Cabral, estimated in 1987 for the proportion of 7,pierce JP,Fiore MC,Novotny TE,Hatziandreu EJ, Rhode Island;M Mace,South Carolina;S Moritz,South all women in the general U.S. popula- Davis RM.Trends in cigarette smoking in the United Dakota;D Ridings,Tennessee;J Fellows,Texas;L tion had attempted t0 for at States:educational differences are increasing.JAMA 0ri who aem quit Post-Nilson, Utah;J Bowie,Virginia;K Tollestrup, Il tl 1989;261:56-60. Washington; R Barker, West Virginia; E Cautley, least 1 day(32%).5 Therefore,Smoking- S.Williamson DF,Serdula MD,Kendrick JS,Binkin Wisconsin.Office of Surveillance and Analysis,Div of cessation education for reproductive- NJ.Comparing the prevalence of smoking in pregnant Reproductive Health,Div of Chronic Disease Control and nonpregnant women, 1985 to 1986. JAMA and Community Intervention,and Office on Smoking aged women may be more successful 1989;261:70-4. Differences in Age of Smoking Initiation Between Blacks and Whites— United States MMWR.1991;40:754-757(2 tables omit- regular cigarette smoking by race(num- old were you when you first started ted) bers from racial groups other than white smoking cigarettes fairly regularly?";in IN 1988, an estimated 434175 prema- and black were too small to provide sep- 1988, persons were asked, "About how ture deaths in the United States were arate estimates)and sex,through anal- old were you when you first started attributed to cigarette smoking; for yses by birth cohort from 1910 through smoking cigarettes fairly regularly?" blacks,the rate of years of potential life 1959; the report is based on data from Those who said they had never smoked lost before age 65(YPLL)attributed to CDC's National Health Interview Sur- regularly were excluded. Responses smoking(2471.8 YPLL per 100 000 pop- veys(NHISs) for 1987 and 1988. from 38 906(44%)ever regular smokers ulation)was twice that for whites(1224.7 The NHIS interviews persons aged were used in this report.The data were YPLL per 100000 population).' In the =18 years selected from representative weighted to provide national estimates. United States,black adolescents are less national samples of the U.S. civilian, Ninety-five percent confidence intervals likely than white adolescents to smoke2,3; noninstitutionalized population.Approx- were calculated using SE SUDAAN.5 however, black adults are more likely imately 88 000 persons(44 000 each year) The overall proportion of persons who than white adults to begin smoking af- were interviewed during 1987 and 1988. became regular smokers before ages 16, ter adolescence." This report summa- In 1987, persons who had smoked at 18,21,25,and 30 years increased across rizes trends in the age at initiation of least 100 cigarettes were asked, "How successive birth cohorts; however, 3112 JAMA,December 11,1991—Vol 266,No.22 From the CDC .0 L ! among blacks, increases occurred only rette smoking.has decreased markedly ation of cigarette smoking by children before ages 21, 25, and 30. More than among black high school seniors(s;J.G. and youth so that no more than 15%have 80% of smokers born after 1930 began Bachman, L.D. Johnston, P.M. O'Mal- become regular smokers by age 20 years smoking regularly by age 21. ley, University of Michigan, unpub- (objective 3.5).To decrease initiation of The overall average age at which lished data, 1990)-possibly because smoking among younger age groups, smokers began smoking cigarettes reg- blacks begin smoking at older ages than the following measures should be con- ularly decreased from 19.7 years among whites. Although the findings from sidered:1)implementation of health ed- persons born from 1910 through 1919 to NHIS are consistent with this trend, ucation programs on tobacco use in 17.4 years among those born from 1950 current differences in adolescent smok- schools (objective 3.10); 2) establish- through 1959. ing by race suggest the prevalence of ment of tobacco-free environments in Among the successive birth cohorts smoking among black adolescents as schools (objective 3.10); 3) enactment in this study, the average age at smok- they mature will not attain the same and enforcement of laws prohibiting the ing initiation decreased 2.4 years for prevalence as that among whites of the sale and distribution of tobacco products whites and 1.3 years for blacks. The same age group. Additional efforts are to minors(objective 3.13);4)elimination average age at initiation decreased sub- needed to determine the factors that or restriction of tobacco product adver- stantially for white and black women affect cigarette smoking initiation by tising to which youth are likely to be ex- (5.4 and 4.6 years, respectively), de- race and sex. posed(objective 3.15);and 5)increasing creased slightly for white men (0.5 Monitoring trends in age at smoking to 50 the number of states with plans to years), and increased slightly for black initiation and in smoking prevalence of reduce tobacco use, especially among men (0.7 years). current adolescents as they mature may youth(objective 3.14).10 Reported by:HN Giebel,MD,Riverside General Has- enable their smoking behavior patterns pital, Riverside, California. SL Mills, MD, National in later adult life to be understood more References Cancer Institute;SE Marcus,PhD,National Institute clearly. In 1974 38.6% of whites and 1.CDC.Smoking-attributable mortality and years of of Dental Research,National Institutes of Health.Of- potential life lost-United States, 1988. MMWR fice on Smoking and Health, National Center for 47.1%of blacks aged 20-24 years were 1991;40:62-3,69-71. Chronic Disease Prevention and Health Promotion; current smokerss• however, by 1988 2. CDC. Cigarette smoking among youth-United Div of Health Interview Statistics,National Center for the proportions of whites and blacks in States,1989.MMWR 1991;40:712-5. Health Statistics;Surveillance Br,Div of Surveillance, e Il p 3. CDC. Tobacco use among high school students- Hazard Evaluations,and Field Studies, National In- this age group who Were current smok- United States,1990.MMWR 1991;40:617-9. stitute for Occupational Safety and Health,CDC, ers had decreased to 28.5°k and 24.8% 4.Escobedo LG,Anda RF,Smith PF,Remington PL, Mast EE. Sociodemographic characteristics of ciga- CDC Editorial Note: The findings in respectively (CDC, unpublished data), rette smoking initiation in the United States- this analysis are consistent with previ- with black smokers decreasing at a 1990;2 4:15 for smoking-prevention policy. JAMA 1990;264:1550-5. ous reports that indicate smokers in the higher rate (22.3 percentage points) 5.Shah BV. SESUDAAN:standard errors program United States are smoking regularly at than white smokers (10.1 percentage for computing of standardized rates from sample sur- e 7 vey data. Research Triangle Park, North Carolina: an earlier age ;in addition,the secular points). Although this trend suggests Research Triangle Institute,1981. patterns of age at which smoking begins smoking-related morbidity and mortal- 6.CDC.Reducing the health consequences ofsmoking: have changed substantially over time ity could decline among blacks, the 25 years of progress Rockville, Maryland: report of the Surgeon General. : US Department of Health and by both sex and race. greater likelihood of relapse among Human Services,Public Health Service,1989;DHHS One potential limitation of this analy- black smokers indicates thatsmoking- publication no.(CDC)89-8411. 7.Office on Smoking and Health. The health conse- sis is that respondents were asked to cessation efforts targeted toward black quences of smoking for women-a report of the Sur- recall an event (i.e., age at onset of smokers need to be intensified.' geon General.Rockville,Maryland:US Department of Health and Human Services,Public Health Service, regular smoking) that may have oc- The successive birth cohort data in Office of the Assistant Secretary for Health,1980. curred decades earlier. In addition, this report suggest that the average age 8.CDC.The health benefits of smoking cessation:a since mortalityis higher or smokers at which women begin smoking s report a the Surgeon General. Rockville,Maryland: fk g � kig icon- US Department of Health and Human Services,Pub- who begin smoking regularly at earlier tinuing to decline for both blacks and he Health Service, 1990; DHHS publication no. ages, the average age at initiation whites. Persons who begin smoking at (CDC190-8416. 9.Taioli E,Wynder EL. Effect of the age at which among persons born in the earlier co- younger ages are more likely to become smoking begins on frequency of smoking in adulthood. horts may be artificially inflated.'How- heavier smokers' and are at increased N Engl J Med 1991;325:968-9. 10.Public Health Service. Healthy people 2000:n- ever, the overall trend of decreasing risk for smoking-attributed illness or tional health promotion and disease prevention age at initiation is evident even among death.' objectives-full report, with commentary. Washing- those born since 1930. One of the national health objectives ton,DC:US Department of Health and Human Ser- those Public Health Service,1991;DHHS publication Since 1976, the prevalence of ciga- for the year 2000 is to reduce the initi- no.(PHS)91-50212. Cigarette Smoking Among Adults- United States, 1988 MMWR.1991;40:757-765(2 tables omit- points peryear.','To determine the prev- For 1988, the OHS included the fol- ted) alence of smoking among adults in the lowing questions on smoking behavior: IN 1964,the first Surgeon General's re- United States in 1988,the Occupational "Have you smoked at least 100 Giga- port on smoking focused on the health Health Supplement(OHS)of CDC's Na- rettes in your entire life?"and"Do you hazards associated with cigarette smok- tional Health Interview Survey collected smoke cigarettes now?"Among persons ing.'From 1965 through 1987,the over- information on cigarette smoking from who reported smoking at least 100 cig- all prevalence of cigarette smoking a representative sample of the U.S. ci- arettes, current smokers were defined among adults in the United States de- vilian, noninstitutionalized population as those who reported being a smoker at clined by approximately 0.5 percentage aged ,18 years. the time of the interview, and former JAMA,December 11, 1991-Vol 266,No.22 From the CDC 3113 I AP ri smokers,as those who were not current higher than that of blacks (32.4%[95% associated with difficulty in quittings) smokers.Both current and former smok CI=30.2%-34.6%]). The proportion of from marital discord may decrease the ers were classified as ever smokers.The Hispanics who had stopped smoking likelihood of quitting. proportion of persons who had stopped (44.9%[95%CI=41.7%-48.1%])was sim- Cigarette smoking is the single most smoking was defined as the number of ilar to that for non-Hispanics(45.9%[95% important preventable cause of death in former smokers divided by the number CI=45.1%-46.7%]). The proportions of the United States.' One of the national of ever smokers.Current smokers were adults with<a high school education who health objectives for the year 2000(objec- asked,"On the average,about how many had stopped smoking (41.1% [95% tive 3.4)is to reduce the prevalence of cig- cigarettes a day do you smoke?" Data CI=39.6%-42.7%]) and of adult high arette smoking among adults to no more were available on cigarette smoking sta- school graduates who had stopped smok- than 15%.8 To achieve this goal,the cur- tus for approximately 44 000 persons ing(41.3%[95%CI=40.0%-42.6%])were rent rate of decline must be doubled. aged ,18 years and were weighted to lower than those for persons with some Health-care providers and public provide national estimates.Ninety-five college education (47.7% [95% CI= health agencies must increase efforts to percent confidence intervals(CIS)were 46.10/c49.3%])and for college graduates prevent the initiation of smoking and,for calculated using SESUDAAN.1 (63.1%[95%CI=61.30/&-64.9%]). smokers,to support attempts to quit and Based on the survey, in 1988 an es- Overall, the mean number of ciga- maintain cessation. Persons with less timated 91.1 million (51.9%) adults in rettes smoked per day by current smok- than a high school education and in low Bo- the United States were ever smokers, ers in 1988 was 21.3. In general, the cioeconomic groups are at especially high and 49.4 million (28.1%) were current mean number of cigarettes smoked by risk for becoming smokers.1,9 In addition smokers. Current smokers included men was higher than the number smoked to directing interventions toward these 30.8%of all men(25.6 million)and 25.7% by women. Whites smoked more ciga- groups, smoking control and prevention of all women (23.7 million). In all age rettes per day than did blacks and per- efforts will require intensified public groups except 18-24-year-olds,the prev- sons of other races, and non-Hispanics health education,increased emphasis on alence of smoking was higher among smoked more cigarettes per day than school health education, and enactment men than women; smoking was most did Hispanics. In 1988, 27.0% (95% and enforcement of effective health- prevalent among persons 25-64 years of CI=26.0%-27.9%) of smokers smoked promoting policies and laws. age.The overall prevalence of smoking 25 or more cigarettes per day. References was higher among blacks (31.7%) than 1.CDC.Reducing the health consequences of smoking: whites(27.8%), and lowest among per- Reported by:Office on Smoking and Health,National 25 years of progress-a report ofthe Surgeon General, Center for Chronic Disease Prevention and Health 1989.Rockville,Maryland:US Department of Health sons of other races(23.8%).The overall Promotion; Div of Health Interview Statistics, Na- and Human Services, Public Health Service, 1989; prevalence also was higher among non- tional Center for Health Statistics; Surveillance Br, DHHS publication no.(CDC)89-8411. Div of Surveillance, Hazard Evaluations, and Field 2.CDC.Tobacco use by adults-United States,1987. Hispanics (28.4%) than Hispanics Studies,National Institute for Occupational Safety and MMWR 1989;38:685-7. (23.5%). Prevalence of smoking was Health,CDC. 3.Shah BV. SESUDAAN:standard errors program for computing of standardized rates from sample sur- highest among persons with less than a vey data. Research Triangle Park, North Carolina: high school education(34.0%)and with CDC Editorial Note: The findings in Research Triangle Institute,1981. only high school (32.0%). this report indicate that from 1987 to 4. NCHS, Schoenborn CA, Boyd GSmoking and � other tobacco use-United States,19898 7. Hyattsville, The prevalence of smoking was sig- 1988,the overall prevalence of smoking Maryland:US Department of Health and Human Ser- nificantly higher among separated and among adults,18 years of age declined vices,Public Health Service,1989;DHHS publication no. (PHS)89-1597. (vital and health statistics;series divorced persons (42.6% [95% from 28.8%1 to 28.1%-approximately 0.7 10,no. 169). CI=41.3%-44.0%])than among persons percentage points. In addition,in 1988, 5.Waldron 1,Lye D.Family roles and smoking.Am J Prev Med 1989;5:13641. in other marital categories: married the proportion of ever smokers who were 6. DiClemente CC, Prochaska JO. Processes and (27.4%[95%CI=26.7%-28.1%]),never former smokers was 45.8%, compared stages of self-change:coping and competence in smok- married (26.5% [95% CI= 25.2%- with 44.2%in 1987.' ing behavior change. In:Shiffman S,wins TA,eds. Coping and substance use.New York:Academic Press, 27.7%]), and widowed (19.5% [95% The higher rates of cigarette smoking Inc.,1985. CI=18.3%-20.6%]). among separated and divorced persons 7.CDC. The health benefits of smoking cessation:a report of the Surgeon General,1990.Rockville,Mary- In 1988, 41.8 million (45.8%) ever appear to reflect higher rates of smok- land:US Department of Health and Human services, smokers were former smokers.The pro- ing initiation before the usual age of Public Health Service, 1990; DHHS publication no. (CDC)90-8416. portion of men(49.0%[95%CI=47.8%- marriage.' In addition, separated and S.Public Health Service.Healthy people 2000:national 50.1%])who had stopped smoking was divorced persons were less likely to health promotion and disease prevention objectives- higher than that of women(42.0% 95% have quit smoking than married per- full report, with commentary. Washington, DC: US [ tl g P Department of Health and Human Services, Public CI=40.8%-43.1%]),and the proportion sons.'Social support provided in mar- Health Service,1991;DHHS publication no.(PHS)91- of whites (47.6% [95% CI=46.8%- riage may increase the probability of g,2CDC. Cigarette smoking among youth-United 48.4%])who had stopped smoking was cessation,'while stress(which has been States,1989.MMWR 1991;40:712-5. Annual Vital Statistics Summary Report MMWR. 1991,40:751 deaths by human immunodeficiency vi- and Technical Information Branch, CDC's National Center for Health Sta- rus infection,and infant mortality.Data NCHS,CDC,Room 1064,6525 Belcrest tistics(NCHS)has released provisional by state of occurrence are shown for Road,Hyattsville,MD 20782;telephone data on the number and rate of births, birth,marriage,divorce,death,and in- (301)436-8500. marriages,divorces,and deaths for 1990. fant death. Reference Monthly estimates and rates are included The report, Annual Summary of 1. NCHS. Annual summary of births, marriages, for each vital event. Births, Marriages, Divorces, and divorces,and deaths:United States,1990.Hyattsville, The report resents statistics on the Deaths: United States 1990 1 is avail- Maryland:US Department of Health and Human Ser- Il P vices,Public Health Service,CDC,1991.(Monthly vi- expectation oflife,major causes ofdeath, able free of charge from the Scientific tal statistics report;vol 39,no.13). 3114 JAMA,December 11, 1991-Vol 266,No.22 From the CDC Abstracts (1991;81:846.849)Shelia Hoar Zahm et al,Occupational Studies Section, NCI, ANN INTERN MED Philadelphia, Pa Executive Plaza N,Room 418,Rockville,MD 20892. Cigarette Smoking: Risk Factor for Premature Facial Wrinkling J GEN INTERN MED Philadelphia, Pa Objective:To determine if cigarette smoking is a risk factor for the development of premature facial wrinkling. Smoking Cessation Following Admission Design:Cross-sectional study. to a Coronary Care Unit Setting:Smoking cessation clinic and community. Patients:Convenience sample of 132 adult smokers and nonsmok- Objective: To determine the impact of an episode of serious ers in 1988. cardiovascular disease on smoking behavior and to identify factors Measurements: A questionnaire was administered to quantify associated with smoking cessation in this setting. cigarette smoking and to obtain information about possibly con- Design:Prospective observational study in which smokers admit- founding factors such as skin pigmentation,sun exposure,age,and ted to a coronary care unit(CCU)were followed for one year after sex.Wrinkling was assessed using photographs of the temple region, hospital discharge to determine subsequent smoking behavior. and a severity score based on predetermined criteria was assigned. Setting:Coronary care unit of a teaching hospital. A logistic regression model,which controlled for confounding van- Patients:Preadmission smoking status was assessed in all 828 pa- ables,was developed to assess the risk for premature wrinkling in tients admitted to the CCU during one year.The 310 smokers sur- response to pack-years of smoking. viving to hospital discharge were followed and their smoking behav- Main Results:The prevalence of premature wrinkling was inde- iors assessed by self-report at six and 12 months. pendently associated with sun exposure and pack-years of smoking. Intervention:None. After controlling for age, sex,and sun exposure,premature morin- Measurements and main results:Six months after discharge,32% klieg increased with increased pack-years of smoking. Heavy ciga- of survivors were not smoking;the rate of sustained cessation at one rette smokers(> 50 pack-years)were 4.7 times more likely to be year was 25%.Smokers with a new diagnosis of coronary heart dis- wrinkled than nonsmokers(95%CI,1.0 to 22.6;P value for trend = ease(CHD)made during hospitalization had the highest cessation 0.05).Sun exposure of more than 50 000 lifetime hours also increased rate(53%vs.31%,p=0.01).On multivariate analysis,smoking ces- the risk of being excessively wrinkled 3.1-fold(CI,1.2 to 7.1).When sation was more likely if patients were discharged with a diagnosis excessive sun exposure and cigarette smoking occurred together,the of CHD, had no prior history of CHD, were lighter smokers (<l risk for developing excessive wrinkling was multiplicative (preva- pack/day), and had congestive heart failure during hospitalization. lence ratio of 12.0;CI, 1.5 to 530). Among smokers admitted because of suspected myocardial infarc- Conclusion:Cigarette smoking is an independent risk factor for tion (MI), cessation was more likely if the diagnosis was CHD the development of premature wrinkling. than if it was noncoronary (37% vs. .19%, p < 0.05), but a diagnosis of MI led to no more smoking cessation than did coro- (1991;114:840-544) Donald P. Kadunce et al, Division of Dermatology, Room nary insufficiency. 48454,University of Utah Health Sciences Center,50 N Medical Dr,Salt Lake City, Conclusion:Hospitalization in a CCU is a stimulus to long-term UT 84132. p g" smoking cessation,especially for lighter smokers and those with a new diagnosis of CHD.Admission to a CCU may represent a time when smoking habits are particularly susceptible to intervention. AM J PUBLIC HEALTH Washington, DC Smoking cessation in this setting should improve patient outcomes because cessation reduces cardiovascular mortality, even when quitting occurs after the onset of CHD. Tobacco Smoking as a Risk Factor for Colon Polyps (1991,6:305311)Nancy A.Rigotti et al,General Internal Medicine Unit,BulSnch Background:Data from a cancer screening project among pattern 1,Massachusetts General Hospital,Boston,MA 02114. makers were used to evaluate the association between tobacco smoking and prevalence of colon polyps. Methods:From 1981-1983,549 White men were examined by flex- AM J OBSTET GYNECOL St Louis, Mo ible sigmoidoscopy and completed self-administered questionnaires including smoking histories. Results:One or more colon polyps were detected in 76 men.Stan- Improving Disclosure of Smoking by Pregnant Women dardized prevalence rates (SPR) for polyps increased by smoking category (never smoked = 0.094; ex-smokers = 0.118, current Smoking is a major modifiable risk factor in pregnancy,and low- smokers = 0.214)and by cigarettes per day,years of smoking,and cost interventions have been developed and tested in diverse popu- pack-years among both current and ex-smokers.Both adenomatous lations of pregnant smokers. Successful intervention depends on and hyperplastic polyps showed an association with smoking while identification, however, and nondisclosure can be a problem. This other types of polyps and polyps with unspecified histology did not. randomized study compared rates of disclosure with two response The risk associated with smoking was greater for polyps greater than formats—multiple choice, in which the patient is able to describe one centimeter in diameter.An interaction with occupational expo- herself as having"cut down,"and the usual history question,"Do you sures was suggested by a greater increase in the SPR for polyps smoke?,"in which she is forced to answer simply"yes"or"no."Each among current smokers employed as pattern makers for more than format was tested in both oral and written channels with a multieth- 10 years than among current smokers similarly employed for 10 years nic adult prenatal population(n= 1078)entering care in a multispe- or less. cialty group.Study results indicate that the multiple choice question Conclusions:Since at least some colon polyps are considered pre- unproved disclosure,regardless of channel(oral versus written),by cursor lesions to colon cancer,one of the most common cancers in the 40%.This effect was observed across racial and ethnic groups.Bio- United States,this report suggests that the possible link between chemical tests of urine samples from reported nonsmokers indicated colon polyps and smoking deserves further evaluation. smoking in only 3%.Eleven percent of the"nonsmokers"in the ex- perimental groups refused consent for the urine test,however,and Edited by Jeanette M.Smith,MD,Contributing Editor. many of these were probably smokers. Abstracts are selected on the basis of interest to our readers.Publication herein does (1991;165:409-413)Patricia Dolan Mullen et al,School of Public Health,Univer- rpt suggest an endorsement of content or a validation of conclusions. sity of Texas Health Science Center,Box 20186,Houston,TX 77225. JAMA,December 11, 1991—Vol 266,No.22 Abstracts 3121 }, Results. The prevalence of depression as measured by the BDI FERTIL STERIL Birmingham,Ala among smokers(n=232)and nonsmokers(n=472)was 24.1%and 15.3%, respectively, a significant difference (P<.001). Physicians ' identified depression at a significantly higher rate (75.0%) among The Effect of Cigarette Smoking on Ovarian Function depressed smokers than among depressed nonsmokers (48.6%) and Early Pregnancy Outcome Of (P<.0001).Smokers were 2.06 times as likely to be labeled depressed In Vitro Fertilization Treatment when controlling for the presence of a current depression,physician A retrospective analysis of the results of the first attempt at in knowledge of a depression history,and physician familiarity with the P Y P patient(P< .0001,95%CI = 1.44,2.94). vitro fertilization(IVF)treatment in 447 couples was performed by Conclusions.Smoking may serve as a cue for the clinician in the chart review.In 124 couples the female partners smoked cigarettes, recognition of depression.Further research is needed to determine and in 236 couples they did not. A total of 69 pregnancies(19.2%) how smoking or a related factor may be used by physicians to cor- were achieved,50 in nonsmokers(21.2%per cycle),and 19 in smok- rectly identify depression. ers(15.3%per cycle).There were no significant differences in these cycles between the two groups in peak estradiol level achieved,the (1991;33255-258)Michael L.Parchman,Department of Family Medicine,PO Box number of eggs retrieved, fertilization rate, or implantation.rate. 26901,Oklahoma City,OK 73190. The incidence of spontaneous abortion was higher in smokers(42.1%) than in nonsmokers (18.9%). Consequently, the delivery rate per cycle of IVF was significantly lower in the smoking group(11 of 124, PUBLIC HEALTH REP Hyattsville, Md 9.6%)than in the nonsmokers(40 of 236, 17.0%).There was no ef- fect when only the husband was a smoker.From these data,smok- ing appears to significantly reduce the chances of successful preg- Evaluation of`Gula para Dejar de Fumar,' nancy after IVF treatment. a Self-help Guide in Spanish to Quit Smoking (1991;55:780-783) H. Anthony Pattinson et al, University of Calgary Medical Because of the absence of culturally appropriate self-help smoking Clinic,3330 Hospital Dr NW,Calgary,Alberta,72N 4N1,Canada. cessation materials for Latinos, a new Spanish language cessation guide,"Guia para Dejar de Fumar,"was developed and evaluated. It was distributed as part of a community-wide intervention to de- crease the prevalence of smoking. The "Guia" is an attractive J NATL CANCER INST Bethesda, Md full-color booklet written in universal Spanish that uses simple text and numerous photographs. Motivation to quit smoking is empha- sized, and graphic demonstrations of the adverse health effects of Smoking-Attributable Cancer Mortality in 1991: smoking are included. A menu of quitting and maintenance tech- Is Lung Cancer Now the Leading Cause of Death piques is presented. Among Smokers in the United States? A total of 431 smokers were identified for evaluation at approxi- Findings from the new American Cancer Society prospective mately 3,6,and 12 months after receiving the"Guia."Self-reported quit rates declined from 21.1 percent at 2.5 months to 13.7 percent at study of 1.2 million men and women indicate that mortality risks 14 months;8.4 percent of the sample had a validated quit status by among smokers have increased substantially for most of the eight saliva cotinine test at 1 year.Persons older than 44 years were more major cancer sites causally associated with cigarette smoking.Lung likely to remain nonsmokers,but sex,education,acculturation score, cancer risk for male smokers doubled,while the risk for females in- and cigarettes smoked per day did not predict smoking cessation. creased more than fourfold.On the basis of the new American Can- The components of the"Guia"most mentioned by those who were cer Society relative risks,we project that cigarette smoking alone surveyed were the graphic photographs, the health emphasis, and will contribute to slightly more than 157 000 of the 514 000 total can- the overall format.The authors concluded that the"Guia"is an ap- cer deaths expected to occur in the United States in 1991. Overall, propriate self-help smoking cessation booklet for Spanish-speaking smoking directly contributes to 21.5%of all cancer deaths in women Latinos in the United States. but 45%of all cancer deaths in men. It would also appear that lung cancer has now displaced coronary heart disease as the single lead- (1991;106:564-570)Eliseo J.Pdrez-Stable et a1,400 Parnassus Ave,A-405,San ing cause of excess mortality among smokers in the United States. Francisco,CA 94143-0320. (1991;83:114&1148)Donald R.Shopland et al,Smoking and Tobacco Control Pro- gram,Executive Plaza N,Room 241,National Institutes of Health,9000 Rockville Pike,Bethesda,MD 20892• AM REV RESPIR DIS New York, NY Passive Smoking by Humans Sensitizes J FAM PRAC Norwalk, Conn Circulating Neutrophils The proinflammatory effects of passive inhalation of cigarette Recognition Of Depression smoke were investigated by exposing a total of 16 healthy, young In Patients Who Smoke nonsmokers(mean age 29± 1.4 yr, 11 women and five men)to ac- tively smoking individuals in a poorly-ventilated room. Neutrophil Background.Depression is a common illness in family practice and functions were measured before and after 3 h of exposure to cigarette is frequently missed by busy practitioners. Recent studies have smoke. Passive cigarette smoking was associated with increased suggested a relationship between smoking and depression in the leukocyte counts(mean increase 33%,p<0.005),chemotaxis(57%, general population. The purpose of this study is to determine p<0.001),and release of reactive oxidants(71%,p<0.005)by stim- whether a patient's smoking is related to the physician's recognition ulated neutrophils.These results were confirmed in a second study of a patient's depression. If so,smoking may serve as a cue used by designed to eliminate the possible complicating effects of serial physicians to recognize depression in their clinical decision-making venepuncture.Plasma concentrations of the proinflammatory cyto- process. kines interleukin-1(IL-1)a, IL-10,IL-6,and tumor necrosis factor Methods.Adult patients presenting to the University of Oklahomaalpha(TNFa)were not affected by passive smoking.These results Family Practice Residency Clinic were screened for depression us- indicate that inhalation of sidestream tobacco smoke promotes sys- ing the short form of the Beck Depression Inventory(BDI). After temic priming of neutrophils. These potentially proinflammatory each patient visit, upper level residents or fellows completed events may induce oxidant-mediated tissue damage and carcinogen- response cards on which they recorded their assessment of the like- esis in the lungs of passive smokers. lihood of a depression,their familiarity with the patient,and whether (1991;144:570-574)Ronald Anderson et al,Department of Immunology,PO Box they had any knowledge of a depression history. 2034,Pretoria 0001,South Africa. 3122 JAMA,December 11,1991—Vol 266,No.22 Abstracts Letters Kids'Concept of Cigarette Code attractive),19(86%)of the students said out aspirin treatment. This indicates a To the Editor.—Cigarette companies that the same advertisement suggests significant cost saving by the reduction have a voluntary code not to advertise that smoking makes the models look very of cesarean births alone. to children,' and they claim that their attractive. Two concerns may alter this assess- advertisements do not encourage chil- I asked my class how many people die ment. As mentioned in the meta-anal- dren to smoke.I The code says that mod- each year from smoking, AIDS, auto ysis, the indications for cesarean sec- els in advertisements shall appear to be accidents, drugs, and murder and how tion were not uniformly described. In over 25 years old, that the advertise- many adults smoke.Only seven(33%)of addition,although there were no aspirin- ments won't suggest that the models' the students knew that more than related complications in the studies,with attractiveness and good health are due 400000 people die from smoking every larger numbers,complications may be- to smoking, and that models shall not year. The average answer was 150 000 come evident,and the cost of complica- appear to have just participated "in a deaths.The students thought the num- tions would need to be added to the physical activity requiring stamina be- ber of deaths from AIDS was 100 000, analysis. yond normal recreation." The purpose four times the right answer, and for We hope that the results of the meta- of my science project was to find out if drugs, 100000, 20 times the right an- analysis will encourage larger studies. cigarette advertising reached children swer. The average answer for adult As more data become available, more in my age group(12 and 13 years)and smokers was 49%,almost twice the cor- detailed decision-making analyses can if the code protected children. rect number, 28%. be applied,perhaps including other out- My results show that cigarette ad- comes like low birth weight. Only then vertising reaches children and they think can the potential benefit of aspirin in the See also pp 3145, 3149, and 3154. that the advertisements make smoking prevention of pregnancy-induced hyper- look glamorous, healthy, and youthful. tension and its complications be more The results show that children may not fully assessed. I asked the 22 members of my seventh- be fully aware of the dangers of this Neal Clemenson, MD grade class what magazines they read habit. Therefore, because the tobacco Roberto Bisonni, MD most often and found that 70%read mag- industry fails to abide by their code and University of Oklahoma azines that advertise cigarettes. I then because cigarettes kill over 400 000 peo- Health Sciences Center cut out cigarette advertisements from ple per year, I think cigarette adver- Oklahoma City past issues of the top two magazines, tising should be prohibited. Sports Illustrated and People.To see if1. Imperiale TF,Petrulis AS.A meta-analysis of low- Deirdre Connolly dose aspirin for the prevention of pregnancy-induced cigarette advertising reached them, I Chenery Middle School hypertensive disease.JAMA. 1991;266261-265. took Camel and Marlboro advertise- Belmont, Mass 2. Schiff E,Peleg E,Goldenberg M,et al.The use of ments and cut out the brand names and aspirin to prevent pregnancy-induced hypertension and lower the ratio of thromboxane AZ to prostacyclin left Only the cartoon camel and cowboy. I would like to thank my sisters, Michelle and in relatively high-risk pregnancies.N Engl J Med. To see if the code worked,I cut out four Maura, my father, Gregory Connolly, and my 1989;321:351-356. teacher,Ms Walsh,for their help.Without them 3. Wallenburg HCS,Dekker GA,Makovitz JW,et al. advertisements with models who looked Low-dose aspirin prevents pregnancy-induced hyper- this project would not have been possible. tension and preeclampsia in young(Lucky Strike and Kool),attrac- p p' angiotensin-sensitive I. Tobacco Institute.Cigarette Advertising and Pro- primigravidae.Lancet. 1986;1:1-3. five (Salem), healthy (Salem), Or as if 4 McParland P,Pearce JM,Chamberlain G VP,et al. the had just played a sport(Winston). motion Code. Washington, DC: Tobacco Institute; Y j p y p lsso Doppler ultrasound and aspirin in recognition and pre- I made up a questionnaire based on State- 2. Tobacco Institute.Peer Pressure and Acting Grown vention of pregnancy-induced hypertension. Lancet. Up:The Big Reasons Why Youngsters Smoke.Wash- 1990;335:1552-1555. ments in the code. I made slides of the ington,DC:Tobacco Institute;1986. 5. Shiono PH,Fielden JG,McNellis D,Rhoads GG, six advertisements and showed them to Pearse WH.Recent trends in cesarean birth and trial of labor rates in the United States. JAMA. my class. They filled out the question- Low-Dose Aspirin to Prevent 1987;257;494-497. naire while watching my slide show. Pregnancy-induced H Everyone recognized the cartoon Hypertensive Disease camel to be an advertisement for Cam- Guidelines for Letters els,and 14(64%)recognized the cowboy To the Editor.—The recent meta-anal- sis of aspirin andpregnancy-induced Lettere will be published at the discretion of advertisement as representing Marl- Ythe editor as space permits and subject to boro.The class guessed that the females hypertension by Imperiale and Petru- Ming and abridgment. They should be in the Lucky Strike and Kool advertise- lis caught our interest, and we exam- typewritten double-spaced and submitted in ments were 19 years of age, and the ined this issue from the standpoint of duplicate.They should not exceed 500 words male in the Kool advertisement 23 years clinical decision making.Since the meta- of text.References,if any,should be held to of age. Sixteen students said that the analysis included studies that were not a minimum, preferably five or fewer. Let- group in the Winston advertisement had randomized,not blinded, or used dipy- ters discussing a recent JAMA article should just participated in softball or baseball, ridamole in addition to aspirin, we ex- be received within 1 month of the article's a strenuous physical activity.On a scale cluded data from those studies. Using publication. Letters must not duplicate of 0 to 3 (zero being very sickly and data from the remaining studies that other material published or submitted for three being healthy), 20(90%) of the addressed cesarean sections,'-''we found right�aion. hi signed statement taementility, end finan- students said that the Salem advertise- that 5.6%of the aspirin group required tial disclosure is essential for publication.It ment suggested that smoking makes the cesarean sections,compared with 23.9% is not feasible routinely to return unpub- models look healthy. On a scale of 0 to of the control group. With a maximum lished letters unless such is requested. Let- 3(zero being ugly and three being very aspirin cost of$4.60 and a cost differ- ters not meeting these guidelines are gener- ence between cesarean and vaginal birth ally not acknowledged.Also see Instructions Edited by Drummond Rennie,MD,Deputy Editor of$3014,5 the overall expected cost with for Authors. (west),and Bruce B.Dan,MD,Senior Editor. . aspirin is$173 per patient vs$720 with- 3126 JAMA,December 11,1991—Vol 266,No.22 Letters To the Editor.—"A Meta-analysis of effect on perinatal deaths, these were eclampsia,the impact on the fetal syn- Low-Dose Aspirin for Prevention of too few for analysis. The cause of each drome(intrauterine growth retardation Pregnancy-Induced Hypertensive Dis- death is important to detail, especially and perinatal mortality)is less clear. In ease,"by Imperiale and Petrulis,I and"A relative to the question of aspirin-asso- particular, although the trends are en- Multivariate Analysis of Risk Factors for ciated hemorrhagic risk. Reassurance couraging,there is not proof that peri- Preeclampsia,"by Eskenazi et al,2 add to was provided for low-dose aspirin safety natal mortality is influenced either ad- the growing knowledge"of the potential from a preliminary analysis of the Na- versely or beneficially. It is, however, to improve pregnancy outcome with daily tional Institute of Child Health and Hu- imprecise to say, as Imperiale and low-dose aspirin starting in the second tri- man Development study.' Petrulis do in the summary,that the tri- mester. The meta-analysis combines Franz Rosa, MD, MPH als show"no effect on fetal and neonatal studies for preventing pregnancy- Linda J. Miwa, PharmD death" when, as they must acknowl- induced hypertension with those for pre- Food and Drug Administration edge,the numbers are too small to mea- venting recurrent fetal growth retarda- Rockville, Md sure clinically valuable improvements in tion.An additional recent study by Uzan this relatively rare adverse outcome. et all provides more extensive data sup- The preceding letter is not an official statement Furthermore, the conclusion pre- porting prevention of fetal growth retar- of the Food and BYrug Administration.—ED. sented by Imperiale and Petrulis, that dation with aspirin.Thus,thromboxane- I. Imperials TF,Petrulis AS.A meta-analysis of low- low-dose aspirin is "highly efficacious doe aspirin for the prevention ofpregnancy-induced » induced placental insufficiency does not hypertesive diseasJAMA. 1992621- a. and safe, for preventing pregnancy-in- appear to be limited to pregnancies with 2. Eskenazi B,Fenster L,Sidney S.A multivariate duced hypertension,overstates the case. hypertension. analysis of risk factors for preeclampsia. JAMA. In doing so, they may make it more The published data suggest that preg- 3. He66:237ciott P g 3. Hertz-Piceiotto I, Hopenhayn-Rich C, Golub M, difficult to complete the necessary sci- nancy risk groups to be considered for Hooper K. The risks and benefits of taking aspirin entific evaluation of this potentially im- a Brtpregnancy. b BLl rolow-dose aspirin include(1)primiparae, n JRSiM Low-dose aspirin to improve Portant therapy. In short, safety has especially those who are black, and perinatal outcome. Clin Obstet Gynecol. 1991;34:251- yet to be proved. The studies so far (2)women with(a)previous pre eclamp 5. U. S,Beauf>is M,Bream G,Bazin B,Capitant C, have been too small to have the power sia,(b)a history of previous fetal growth Paris J. Prevention of fetal growth retardation with to demonstrate rare but serious side ef- retardation not explained by other low-dose aspirin:findings of the EPREDA trial.Lan- fects. Consequently,much larger trials causes,and(c)chronic maternal hyper- 6 tSibaai1;BM.1DDiiagnos s and management of chronic are being conducted such as the CLASP tension.s Questions remain as to whether hypertension in pregnancy. Obstet Gynecol. trial organized by the Medical Research 1991;78:451161. other groups or all pregnancies are at 7. Sibai BM,Mirro R,Chesney CM,Leger C.Low- Council of England,to which more than high enough risk of thromboxane-in- dose aspirin in pregnancy.Obstet Gynecol.1989;74:551- 6200 women have so far been entered. duced problems to warrant low-dose as- 556. Recruitment to another large trial or- pirin therapy. Whether low-dose aspi- To the Editor.—Imperiale and Petrulis ganized by the National Institute of rin is useful in preventing fetal growth report a meta-analysis of the use of low- Child Health and Development has retardation associated with postponed dose aspirin to prevent pregnancy-in- passed 2000 women. Neither trial has pregnancies in older women needs spe- duced hypertensive disease.'Despite a yet been stopped so that one must as- cial attention. review of the English-language litera- sume that unequivocal benefit has not Other major maternal factors need to ture, they are apparently unaware of a yet been demonstrated. When com- be considered,including cigarette smok- similar meta-analysis first published by pleted in the next year or two, these ing (associated with decreased pre- Collins in May 1988 and regularly up- studies should provide reliable informa- eclampsia and increased fetal growth re- dated subsequently.',' Unlike Imperi- tion about both clinical efficacy and tardation), alcohol intake, malnutrition ale and Petrulis,Collins excluded all non- safety.Their success,however,depends (eg, calcium deficiency, imbalance of randomized studies since such studies critically on the fact that clinicians are prostinoid precursors,obesity,undernu- may introduce a bias in the assessment not persuaded by reports such as that of trition), and family history. In analyzing of treatment. Similarly,because of con- Imperiale and Petrulis that suggest that outcomes,the distribution of weight-for- cerns about the selection biases that randomization is no longer necessary. dates has advantages over low birth might arise by reviewing only the pub- C. W. G. Redman, MB, BChir, FRCP weight or short gestational duration rates lished data, Collins included data from John Radcliffe Hospital or averages. The overall rates or aver- unpublished studies as well as seeking Oxford, United Kingdom ages may obscure disparate findings by information from investigators about rel- Marshall D. Lindheimer,MD lumping together fetal growth retarda- evant outcome measures that had not Chicago(Ill)Lying-In Hospital tion,postmaturity,prematurity, and in- been reported. Thus, 13 studies con- duced delivery (especially important tribute to his latest analysis including 1. Imperials TF,Petrulis AS.A meta-analysis of low- when hypertension is a complication). the results of the largest trial the re- dose aspirin for the prevention of pregnancy-induced hypertensive disease.JAMA. 1991;266:261-265. Follow-up data on head size and intelli- sults of which were presented publicly 2. Collins R,Wallenburg HCS.Pharmacological pre- gence will be important,since preventing in January 1990 and published recently.' vention and treatment of hypertensive disorders in fetal growth retardation may have long- We would commend databases such pregnancy.In:Chalmers I, ncyEnkn M,Childbirth. h. NC, � �' g eds.Effective Care in Pregnancy and Childbirth.New term developmental benefits. as the Oxford Database of Clinical Tri- York,NY:oxford University Press Inc;1989:512-533. Studies to date indicate more data als'in which u -to- a-anases date metl S. Collins R.Antiplatelet agents for IUGR and p �' eclampsia. In: Chalmers I, ed. Oxford Databasee of and perhaps better methods are needed are regularly presented and reviewed. Perinatal Trials.version 1.2,Disk Issue 5,February to define who should receive low-dose This is inherently a much more satisfac- 4. U Record 4000. 4. Uzan S,Beaufils M,Bream G,Bazin B,Capitant C, aspirin.More data are awaited from the tory way of monitoring trial results than Paris J. Prevention of fetal growth retardation with National Institute of Child Health and single ad hoc reports.The conclusions of low-dose aspirin:findings of the EPREDA trial.Lan- Human Development multi-institutional the more complete meta-analyses are oet•1991'337:1427-1431. randomized trial of 3000 pregnancies and similar to those of Imperiale and Petru- In Reply.—We appreciate the comments the British Council of Medical Research lis,namely,that while antiplatelet ther- by Drs Clemenson and Bisonni regard- trial of 7500 pregnancies. Although the apy with low-dose aspirin prevents the ing the potential cost savings of low- meta-analysis concluded there was no maternal syndrome of proteinuric pre- dose aspirin with respect to cesarean JAMA, December 11, 1991—Vol 266,No.22 Letters 3127 1 i section. We agree with their concerns strated(relative risk[RR]= 0.88;95% to the summary-level data). Such'data- j about the need to ascertain the indica- confidence interval [CI], 0.32 to 2.46). bases would greatly enhance the ability to tions for cesarean section and to study Evidently unnoticed by them was our examine treatment effects in different the effect of low-dose aspirin in greater acknowledgment of the need to study clinical subgroups of patients. numbers of subjects. greater numbers of patients to deter- Thomas F.Imperiale,MD The points made by Drs Rosa and mine whether low-dose aspirin affects Alice S.Petrulis,MD Miwa are well stated. We concur with perinatal mortality. Case Western Reserve University the need to examine the effect of low- We are puzzled by Redman and Lind- Cleveland,Ohio dose aspirin in well-characterized pa- heimer's apparent double standard con- tient subgroups. A more accurate and cerning publication of our analysis and 1. Collins R,wallenburg HCS.Pharmacological pre- vention way of estimating risk is the work of Collins,',' articular) since pennon and treatment of hypertensive disorders in q y g particularly pregnancy.In:Chalmers I,Enldn M,Keirse MJNC, needed and should consider the vari- the results are similar (despite aggre- eds.Effective care in Pregnancy and Childbirth.New ables that the mentiUnfortunate) tin dit setf trials). Why York,NY:Oxford University Press;1989:51 5' on. y� gating s oand p 2. Collins R.Antiplatelet agents for IUGR and pre- because of the summary-level data con- would publication of our work be any eclampsia. In: Chalmers I, ed. Oxford Database of tained in published studies, meta-anal- more likely to persuade clinicians"that Perinatal Trials.Version 1.2,Disk Issue 5,February ysis often cannot examine the effect of randomization is no longer necessary" 1991,Record 4000. individual variables on outcome. This than the work of Collins? We believe limitation may result in the failure to that clinicians can appreciate potential Prophylactic Aspirin Treatment: identify subgroups of patients that may (or likely) treatment efficacy and the The Merits of Timing respond differently to treatment than need for ongoing randomization simul- To the Editor.—Two recent articles sug- the overall results indicate. taneously. gested that the inhibition by aspirin of A careful reading of our article by Drs Finally,the Oxford Database of Clini- thromboxane synthesis and cyclooxyge- Redman and Lindheimer would have as- cal Trials was a fine idea and should be em- nase-dependent platelet aggegation pro- suaged many of their concerns, partic- ulated. However, those who assess the tects women against a first myocardial ularly those regarding our statements methods and analyses in the reports that infarction' and pregnancy-induced hy- about efficacy and safety of low-dose are selected must be vigilant and them- pertension.2 So far, little if any atten- aspirin.They neglected to note our qual- selves must use proper analyses to ensure tion has been paid in clinical trials to the ifiers. Moreover, with respect to our and preserve validity. This and future, timing of aspirin administration. Per- results,we stated that low-dose aspirin similar databases could be improved fur- haps some individuals may take aspirin had no effect on fetal and neonatal death ther by inclusion of the"raw"(ie,patient- at a convenient time,but when the drug because this is what the data demon- level)data from clinical trials(as opposed is inactive. Aspirin effect on lipoperoxides(LPs)in platelet-rich plasma(lett panel)and lymphocyte 0-adrenergic receptors(B-Rs)(right panel)may be predictably present or absent as a function of timing(circadian rhythm stage).Easily implemented and cost-effective timing of aspirin critically determines two of the drug's major effects. Last time point,labeled bedtime,corresponds to approximately 15 hours after awakening.Change in LPs and B-Rs are expressed as a percentage of the overall pretreatment average.MESOR indicates midline-estimating statistic of rhythm. Time of Therapy,h Awakening Bedtime Awakening Bedtime 0 (A) A+3h A+6h A+9h A+12h (A) A+31h A+6h A+9h A+12h ` k 25 20 - -20- V 0—20 15 mt c —40 S #," t i CD W 10 § L - L f� 'F."� �t 5 - -60- 0—60 —80 —5 MESOR=0.49 P=.012 Period Fitted=24 h 0°=00 MESOR=0.13 P=.003 Amplitude=0.35 Percent Rhythm=95 Amplitude=0.14 Percent Rhythm=98 Acrophase=-173°(A+11h,32 min) Acrophase=-3591(A-4 min) 3128 JAMA,December 11, 1991—Vol 266,No.22 Letters 10 Results shown in the Figure from a a six-treatment time-point approach emphasize the initial period of drug small sample of women suggest (with aimed at finding the optimal time for treatment (1 to 3 months), as that in statistical significance)not only that low targeting aspirin administration,will im- which agranulocytosis is most likely to doses of aspirin affect prostaglandin and prove, at minimal added cost, any ben- happen." We found only one case re- adrenergic pathways but also that such efits from aspirin in preventing adverse port of agranulocytosis in the second effects vary as a function of the circa- cardiovascular events. course of propylthiouracil(PTU)treat- than stage at which the aspirin is taken. Germaine Corn6lissen, PhD ment,4 and a few more cases on methi- Six clinically healthy women aged 20 to Franz Halberg, MD mazole.3,5 We report here four cases of 30 years volunteered to participate in a University of Minnesota agranulocytosis on second exposure to randomized pilot study that consisted of Minneapolis PTU. We think it is important to draw a reference stage (which lasted 2 days Pavel Prikryl, MD attention to the possibility of agranulo- starting after a 5-day adjustment to hos- Eva DankovA, RN cytosis in patients who were previously pital conditions)followed by a 7-day span Jarmila Siegelova, MD treated successfully with PTU. during which aspirin(100 mg/d)was ad- Jurf Dusek, MD ministered at one of six different circa- Masaryk University Study.—The records of all our pa- than stages:on awakening,3,6,9,or 12 Brno, Czechoslovakia tients who were treated with PTU for hours after awakening, or at bedtime. for the International Womb-to-Tomb thyrotoxicosis in the years 1985 through Chronome Study Group 1990 were reviewed. The diagnosis of During the reference stage and during 1. Manson JE,Stampfer MJ,Colditz GA,et al.A pro- thyrotoxicosis was made on the basis of the last 2 days of the low-dose aspirin spective study of aspirin use and primary prevention of physical findings elevated test span, venous blood samples were cardio-vascular disease in women. JAMA. 1991; symptoms, taken eve 4 hours for the determine- 266:521-527. thyroid function tests (total thyroxine ' 2. Imperiale TF,Petrulis A.A meta-analysis of low- [T4],triiodothyronine uptake[T3U],and tion of,among others,lipoperoxide con- dose aspirin for the prevention of pregnancy-induced xin index), and suppressed centration in platelet-rich plasma and 3. einbergA,Zagula-MallyZW,GhataJ rtensive disease.JAMA. HalbergF. thyroid-stimulating free hormone level.The the affinity of lymphocyte 02-adrener- Circadian rhythm in duration of salicylate excretion basal disease was defined b techne- gic receptors for 3H-dihydroalprenolol. referred to phase of excretory rhythms and routine. Y Proc Soc Exp Biol Med. 1967;124:826-83e. tium pertechnetate Tc 99m thyroid scan. The differences in mean value between 4. Markiewicz S, Semenowicz K. Time-dependent propylthiouracil(300 to 900 mg/d)was the low-dose aspirin test span and the chances in the pharmacokinetics of aspirin.Int J Clin reference stage computed for each sub- Pharmacol Biopharm.1979;17:409.411. started in all patients.As symptoms im- g P 5.Johansson BE,Norrving B,Widner H,Wu J,Hal- proved and thyroid function tests nor- ject were assigned to the circadian stage berg F. Stroke incidence:circadian and circaseptan of treatment administration and were (about weekly) variations in onset. In: Hayes DK, malized,the PTU dose was tapered. A Pauly JE,Reiter RJ,eds.Chronobiology:Its Role In trial to stop the drug was usually at- fitted by least squares with a 24-hour Clinical Medicine,General Biology,and Agriculture, cosine curve to assess the response Part A.New York,NY:Wiley-Liss;1990:427-436. tempted after 12 to 18 months.Patients (rhythm) as 6. Haus E,Cusulos M,Sackett-Lundeen L,Swoyer J. who relapsed were re-treated with PTU, Yt )to P rn.When taken on awak- Circadian variations in blood coagulation parameters, and definitive treatment was considered. ening or 3 hours thereafter, aspirin de- alpha-anatrypsin antigen and platelet aggregation and P P01� pressed li roxide concentrations(per- retention in clinically healthy subjects.Chronobiol Int. During treatment with antithyroid haps by inhibiting throm-boxane synthe- 1990;7:203-216. drugs,white blood cell count was mon- sis) and enhanced (3z-adrenergic recep itored every 2 weeks in the first 3 Pro- tors; these effects were much smaller Propylthiouracil-Induced months, and monthly thereafter. Pro- Agranulocytosis in Four Patients pylthiouracil was discontinued if the neu- when the drug was taken 6 or 9 hours Previously Treated With the Drug trophil count fell below 0.5 x 10°/L. after awakening and were not demon- To the Editor.—Agranulocytosis is a A strated 12 hours after awakening. granulocytosis was diagnosed by a neu- The pharmacokinetics and pharmaco- Fell-known major complication of an- trophil count of 0.2 x 101/L or less. In 10 dynamics of various agents change pre- tithyroid drug treatment. Most studies patients, methimazole was substituted dictably as a funtion of circadian (and Clinical Characteristics of Patients With Agranulocytosis* other)rhythm stage, as we have docu- mented broadly in various publications. Patient No. By 1967, Reinberg et all showed that characteristict 1 2 3* 4 the duration of salicylate excretion was Agey 65 62 46 55 longer (P=.002) when the drug (1000 First course mg)was administered at 7 AM vs 7 PM; Duration,mo 14 15 19 29 the availability of aspirin(1500-mg oral Maximal propylthiouracil dose),as gauged by the total serum sal- dose,mg 600 600 300 400 icylate concentration, was statistically Interval between courses,y 6 3 15 7 significantly better when the drug was Second course administered at 6 AM than at 6 PM or 10 Time until AGR,wk 3 4 4 6Maximal propyithiouraal PM.4 Rhythms of cardiac morbidity due dose,mg 300 300 300 300 to coronary occlusion(quantified by us WBC count at AGR, since 19725 and noted by others earlier') x 109/1- 0.8 0.9 2.0 1.2 may be related to rhythms of clotting PMN count at AGR, x 1 OVL 0 0 0.2 0.2 and fibrinolytic inhibitors(reported by Duration of neutropenia,d§ 7 12 11 14 us in 1981 and recently reviewed'). In Clinical manifestation Peritonsillar the morning,an increase in platelet ag- cellulitis Tonsilitis Tonsilitis Asymptomatic gregability,blood viscosity,and concen- Definitive treatments 1 131 1131 1131 ... tration of coagulant factors, and a de- Outcome Hypothyroidism Hypothyroidism Hypothyroidism Spontaneous remission crease in fibrinolytic activity could lead *All patients were women and had been diagnosed as having Graves disease. to a state of relative hypercoagulability. tAGR indicates agranulocytosis;WBC,white blood cell;and PMN,polymorphonuclear leukocyte. #This patient was also treated with propylthiouracil for a month,3 months after the first course. Apart from any mechanisms involved, §Neutropenia was indicated by a PMN cell count of less than 0.5 x 1091L. systematic if empirical research, using II 1 131 indicates radioactive iodine. JAMA,December 11, 1991—Vol 266,No.22 Letters 3129 10 1 for PTU because of minor side effects, Intem Med.1990;150'621-6e4• The more common morbidity of liver 4 Fibbe WE,Class FHJ,Van der star-Nkstra W,et mainly pruritus and rash. Their data al. Agranulocytosis induced by propylthiouracil:evi- disease,renal stones,and polyarthritis, were analyzed together with all other dence of a drug-dependent antibody reacting with which represent a major group of diffi- granulocytes,monocytes,and haematopoietic cells.Br culties for patients undergoing patients. J Haematol. 1986;64:363-373. pe- Statistical Analysis.—Student's t 5. Tamai H,Takaichi Y,Morita T,et al.Methimazole- junoileal bypass,.is infrequent in indi- y induced agranulocytosis in Japanese patients with viduals with gastric bypass. However, test was used whenever appropriate.A Graves'disease.Clin Endocrinol.1989;30:525-530. long-term results of this procedure have total of 178 patients were treated with been somewhat disappointing, and pa- PTU,55 of them for at least two courses. tients have tended a the weight Four patients developed agranulocyto- Obesity:The Role of regain g sis,all of them while being treated with Gastric Surgery they have lost..' the drug for the second time. The rate To the Editor.—The Grand Rounds at Second,although gastric surgery has of agranulocytosis was therefore 2.2% the Clinical Center of the National In- clearly superseded jejunoileal bypass of all patients, and 7.3% of patients stitutes of Health(NIH)by Dr Wilber and will produce significant weight loss, patients after 3 years remain n- treated with two courses or more. The entitled"Neural peptides,appetite reg- clinical characteristics of these four pa- ulation, and human obesity"' that ap- tially overweight and may experience tients are summarized in the Table. peared in JAMA failed to describe gas- symptomatic binary disease.Because of tric surgery for the management of se- these problems,the American Academy Comment.—Antithyroid drugs are of Clinical Nutrition has recommended in the treatment section obesity . widely used in the treatment of patients vere othat gastric procedures should be done with thyrotoxicosis. Perhaps the most This topic was recently discussed at a only in patients (1) who are 100% in dangerous side effect of these drugs is Consensus Panel on the Surgical Treat- excess of ideal weight; (2)who are 100 ment of Severe Obesity at the NIH at agranulocytosis. Our agranulocytosis pounds overweight; (3)who have seri- which it was concluded that gastric sur- rate is higher than the 0.3%to 1%re- ous associated risk factors, such as hy- ported, gem,is appropriate for patients who have possibly due to the small num- a body mass index o less than h with- pertension, diabetes, and cardiovascu- ber of patients in our series. Although lar disease;and(4)who have also been it was stated that the risk of agranulo- out comorbidity and that patients who repeatedly unsuccessful in losing weight cytosis in the second course is not in- have significant comorbidity and a body by nonsurgical means.' creased,','all four of our patients who mass index of 35 to 40 should be con- John F. Wilber, MD had agranulocytosis were treated with sidered as candidates for the procedure. University of Maryland at Baltimore at least two courses of PTU, and none Clearly,this has had the most effective School of Medicine had agranulocytosis in the first course. long-term efficacy in the management All four patients were women, had of severe obesity with the data now 1. Linner JH. Comparative effectiveness of gastric Graves'disease,and were older than 45 showing the average loss of excess bypass and gastroplasty:a clinical study.Arch Surg. years.Compared with patients without weight between 50%to 60%for as long 2.TTalosk Forceof the American Society for Clinical agranulocytosis, patients with agranu- as 10 years after surgical intervention. Nutrition.Guidelines for surgery for morbid obesity. locytosis had a significantly longer in- Gastric surgery should be considered Am J Clin Nutr. 1985;42:904-905. terval between courses of treatment in any review of treatment for the se- (91±53 vs 23±35 months,P<.005).The verely obese patient. Vertical banded Urine Drug Testing—Watch What only other case reported on the second gastroplasty or Roux-en-Y gastric by- You Eatl course of PTU had an interval of 13 pass provides significant satiety for these years,'but intervals with methimazole patients,enabling them to decrease their To the Editor.—Discussions in JAMA were shorter.5 It is possible that long caloric ingestion.In addition,symptoms have raised concerns about false- periods of abstinence from the drug are of the dumping syndrome help patients negative' and false-positive' results of necessary for agranulocytosis to develop who have a gastric bypass avoid the urine drug screening. Exposure to a po- on reexposure,possibly due to the loss ingestion of large quantities of simple tentially common source of an innocently of inhibiting factors,such as T-suppres- carbohydrates. positive screening test for opiates has in- sor cells and blocking antibodies. It would be valuable to understand creased my awareness of these issues and Although routine white blood cell count the neuroregulatory mechanisms that suggested to me that information about monitoring in patients with antithyroid provide this result,so that effective non- this source should be disseminated. drugs is controversial,',' it has been surgical therapy could be provided.Un- Report of a Case.—A 26ryear-old shown that as many as 78%of the patients til that time, surgery is currently the woman presented to us in response to an can be discovered while asymptomatic.' only long-term approach that has been advertisement seeking normal subjects This letter demonstrates that agran- proven to be efficacious for severely for research.Her medical history was sig- ulocytosis from PTU can occur on sec- obese patients. nificant only for chemocautery of a cervi- ond exposure.It may be prudent to mon- Harvey J. Sugerman, MD cal wart. She was receiving no medica- itor the white blood cell count closely Medical College of Virginia tions and denied use of alcohol or other when restarting antithyroid drugs.- Richmond drugs,Findings on physical examination Avinoam Shiran,MD,MSC 1. Wilber JF.Neuropeptides,appetite regulation,and were normal. Routine blood tests were Carmela Shechner,MD human obesity.JAMA.1991;266:257-259. conducted and revealed no significant ab- Gabriel Dickstein, MD normalities. Preliminary urine drug Haifa(Israel)Medical Center In Reply.—Because this article was de- screening was positive for opiates. Gas rived from a 20-minute NIH Grand chromatography—mass spectroscopy con- 1. Cooper PS.Antithyroid drugs.N Engl JMed.1984; Rounds that was limited in scope, all firmed the presence of morphine and co- 311:1353-13U. aspects of obesity management could deine. The subject convincingly denied 2. Cooper PS,Goldminz D,Levin AA,et al.Agranu- locytosis not be considered. I reC0 never- any known ingestion ofopiateS.Review Of associated with antithyroid drugs.Ann In- lm�a tern.Med.1983;98:26-29. theless, that gastric surgery is an im- oral intake revealed that she had eaten a 3. Tajiri J, Noguchi S, Murakami T, Murakami N. Octant option in management of se- lemon seed muffin 5 hours prior t0 Antithyroid drug-induced agranulocytosis:the useful- p p g Poppy ness of routine white blood cell count monitoring.Arch verely obese individuals. screening.A friend provided single-blind 3130 JAMA,December 11,1991—Vol 266,No.22 Letters confithiation of her ingestion. A second Using Helium for Insufflation eating that the observed disorder was urine sample was obtained 6 days after in- During Laparoscopy 'purely respiratory. gestion and was negative for opiates.The To the Editor.—Laparoscopic surgery Further investigation on the use of subject then ate another muffin and pro- requires the creation of a pneumoperi- helium for pneumoperitoneum is war- vided a urine sample 5 hours later.It was toneum to facilitate visualization and op- ranted. We believe that it may be an again positive for opiates. The subject erative manipulation. Carbon dioxide excellent alternative to CO2,especially was not excluded from the research pro- traditionally has been used for this pur- in those patients with preexistent car- tocol. pose because it is colorless, inexpen- diopulmonary disease or with decreased A literature search revealed that the sive,and not flammable.Laboratory in- respiratory reserve. ability of poppy seeds to produce posi- vestigation has shown, however, that Fred S. Bongard, MD tive urine tests for morphine and co- CO2 pneumoperitoneum can produce sig- Nana Pianim, MD deine is well established.'A survey of nificant hypercapnia and respiratory ac- Se-Yuan Liu, MD colleagues in the departments of psy- idosis.t Although these complications are Maurice Lippmann chiatry and internal medicine revealed rarely a problem in younger patients Ian Davis, MD that only a few knew of this potential older patients with preexisting pulmo Stanley Klein, MD source of a positive drug screen. Addi- Harbor-UCLA Medical Center P g nary disease and those with increased Torrance, Calif tional surveying and direct sampling con- respiratory dead space may have diffi- firmed that lemon poppy seed muffins culty eliminating the CO2 burden in spite 1. Leighton TA,Bongard FS,Liu S-Y,et al.Compar- are extremely popular locally and are of mechanically increased minute ven- ative cardiopulmonary effects of helium and carbon di- truly delicious,and there are some con- tilation. m 2 oxide pneumoperitoneum.Surg Forum.In press. u 2. Wittgen CM, Andrus CH, Fitzgerald SD, et al. sumers who consider themselves ad- Helium has been investigated as an Analysis of hemodynamic and ventilatory effects of dieted."The director of our drug anal- insufflating gas in a porcine model in our laparoscopic cholecystectomy. Arch Surg. S1S laboratorywas aware that 0 1991;126:997-1001. ypoppy laboratory., Simulation of clinical con- 3. Liu S-Y,Leighton TA,Davis I,et al.Prospective seeds could result in the appearance of ditions demonstrated the absence of hy- analysis of cardiopulmonary responses to laparoscopic 6opiates in urine but was skeptical that a pereapnia and the resultant respiratory cholecystectomy.J Laparoend Surg.In press. single muffin was sufficient. I decided acidosis observed in animals that re- to break a rule of science and experi- ceived CO2.We have now begun a clin- ment on myself.A urine sample obtained ical protocol, approved by our institu- prior to muffin ingestion was negative tional review board,and wish to inform for opiates. A sample obtained 1 hour JAMA readers of the successful use of CORRECTION after ingestion of a single muffin was helium as an agent for pneumoperito- Error in Figure.—An error occurred in the positive. A second muffin was then in- neum during laparoscopy. Original Contribution entitled"Changes Over gested(like potato chips,it's hard to eat Report of a Case.—A 33-year-old Time in the Knowledge Base of Practicing just one),and a sample obtained 4 hours woman with symptomatic cholelithiasis, Internists,"published in the August 28 issue later was strongly positive.Gas chroma- after receiving appropriate informed of THE JOURNAL(1991;266:1103-1107).In Fig tography—mass spectroscopy confirmed g3 on page 1105,the diamond and circle sym- the presence of morphine. My only life- consent,underwent laparo-scopic chole- bols for 1983 were transposed. The figure time ingestion of opiates consisted of cystectomy during which helium was should have appeared as below: 150µg of fentanyl citrate and two 60-mg used as the insufflating gas to achieve a doses of codeine during removal of a Pneumoperitoneum. Intraperitoneal lipoma 8 months ago. pressure was maintained at 15 mm Hg. via a manually operated valve mecha- nism. Constant respiratory minute vol- 66 ness of this source of a positive urine ume was maintained throughout. 64 drug screen,it is likely that extensive, Prior to insufflation,the arterial CO2 62 random urine testing for illicit drugs (PaCO2)level was 33 mm Hg and the pH 60 will result in potentially severe nega- concentration was 7.42. After 45 min- 58 tive consequences for at least a few in- . utes of pneumoperitoneum, the PaCO2 256 nocent individuals.I believe that exclu- level was 33 mm Hg and the pH con- 854. sion of this subject from our protocol centration remained at 7.42.At the con- )52 (which paid well)would have constituted elusion of the 120-minute procedure the a) employment discrimination resulting PaCO2 level was 30 mm Hg and the pH 250- 48 from inappropriate use of a laboratory concentration was 7.41.Cardiac output, test.Both positive and negative results systemic blood pressure, and arterial 46 on such tests must be handled with cau- bicarbonate concentration remained con- 44 tion, and their use for screening pur- stant. No arrhythmias were detected. 42. 0 poses must be seriously questioned. _ The patient did not complain of postop- 4019.73 1975 1977 1979 1981 1983 James L. Abelson, MD,PhD erative shoulder pain and was discharged Year of Initial Certification University of Michigan on the first day after the procedure. Ann Arbor Comment.—In a prior clinical study, 1. Orentlicher D.Drug testing of physicians.JAMA. which used CO2 for insufflation,we found Fig 3.—Mean score on examination by year of 1990;264:1039-1040. that Pac02 levels rose an average of certification and subspecialty practice.Squares in- t. Woolley tiv Drug testing of physicians:the danger 10 mm H while H concentrations de- dicate cardiol and gastroenterology;of false positives.JAMA.1990;264:3148. gr Il cardiology g logy; circles, 3. Hayes LW,Krasselt WG,Mueggler PA.Concen- clined an average of O.1 during the course other subspecianies;and diamonds,general inter- trations of morphine and codeine in serum and urine after ingestion of poppy seeds. Clin Chem. Of laparoscopic cholecystectomy.'Blcal- nal medicine. 1987;33:806-808. bonate levels remained constant, indi- JAMA,December 11, 1991—Vol 266,No.22 Letters 3131 af" a Original Contributions Transdermal Nicotine for Smoking Cessation Six-Month Results From Two Multicenter Controlled Clinical Trials Transdermal Nicotine Study Group Objective.—To evaluate the efficacy of a new transdermal nicotine system ability, and numerous educational pro- for smoking cessation. grams and antismolang campaigns. Design.—Two 6-week,randomized,double-blind,placebo-controlled,paral- While the prevalence of smoking in the lel group trials were conducted. Successful abstainers from both trials enrolled United States is-decreasing, current in a third'trial for blinded downtitration from medications (6 weeks) and subse- trends suggest that 22% of adults will quent off-drug follow-up (12 weeks). still smoke in the year 2000.' Setting.—Nine outpatient clinics specializing in the treatment of smoking Cigarette smoking is a complex ad- Cessation. diction,with both behavioral and phar- Patients.—Healthy volunteers who smoked one or more packs of cigarettes macologic components.','The addictive daily and wanted to participate in a smoking cessation program. intervention.—Patients were randomly assigned to a transdermal nicotine TRANSDERMAL NICOTINE STUDY GROUP system delivering nicotine at rates of 21,14,or 7 mg(in trial 1 only)over 24 hours Members of the Transdermal Nicotine StudyGroupare or to placebo. Group counseling sessions were provided to all participants. as follows:oral Health Research,IndiapoliInsts:Arte,Indiana uhi- versitySchoo/o/Dentistry,Indianapolis:Arden Christen, Main Outcome Measure.—Rates of continuous smoking abstinence were DDS(chief investigator),Bradley Beiswanger,DDS,Me- determined during6 weeks of full-dose treatment, a 6-week weaning period men Mau,Ms,and Cheryl Walker;Psychiatry Depart- 9 P mens,University of Minnesota,Minneapolis:Dorothy Hat- (through week 12),and a 3-month follow-up receiving no therapy(through week sukami,PhD(chief investigator),Sharon Alien,PhD,Mar- 24). Abstinence was defined by patient diary reports of no smoking during the 9uerite Huber,NP,and Joni Jensen;Pulmonary and Crit- ical Care Medicine Section,University ofNebraska Medical designated periods, confirmed by expired-breath carbon monoxide levels of Center,Omaha:Stephen Rennard,MD(chief investiga- 8 ppm or lower. tor),David Daughton,MS,Ronald Cheney,PA-C,Kath- leen Hatleiid,PA-C,and Austin Thompson,MD;Oregon Results.—The centers enrolled 935 patients.Cessation rates during the last Research Institute,Eugene:Edward Lichtenstein, PhD 4 weeks of the two 6-week trials(pooled data)were 61%,48%,and 27%for 21- (chief investigator);Anthony Biglan,PhD,and Linda Ochs, MS;Pacific Medical Research Services,Redwood City, and 14-mg transdermal nicotine and placebo, respectively (P-.001 for each Calif-Scott Heatley,PhD(chief investigator);Pulmonary active treatment vs placebo). Six-month abstinence rates for 21-mg transder- Drug Evaluation Program,Wheat Ridge,Colo:Lawrence Repsher,MD(chief investigator),William Schones,Dara mal nicotine and placebo were 26%and 12%, respectively(P,.001).All trans- Stillman,CRTT,Cheryl Casey,CRTT,Bonnie Poole,RRT, dermal nicotine doses significantly decreased the severity of nicotine withdrawal and Jennifer Leitch,RN;Center for Research in Disease Prevention,Stanfordty SchoolPalosymptoms and significantly reduced.ci cigarette use by Patients who did not Stop Alto,Calif:Stephen Fortmann,MD(chief investigator), smoking. Compliance was excellent, and no serious systemic adverse effects Joel Killen,PhD,Mark Hansen,and L.Rasenick Douss; Burl- were reported. Department of Psychiatry, University o/Vermont,Burl- POfington:John Hughes, MD(chief investigator), William Conclusions.—Transdermal nicotine systems show considerable promise Valliere,Laura Solomon,PhD,and William Wadland,MD; as an aid to smokingCessation. Tobacco Research Center,West Virginia University School of Medicine,Morgantown:Elbert Glover,PhD(chief in- (JAMA.1991;266:3133-3138) vestigator),and Penny Glover,MEd;Alza Corp,Palo Alto, Calif.Donna Causey,MD,Mark Knowles,PhD,Katherine Voss-Roberts, MS,Della Prather,MA, Nancy Trunnell, and Diane Moos;and Marion Merrell Dow Inc,Kansas Reprint requests to Pulmonary and Critical Care SOME 50 million Americans regularly City, Mo:Clyde Rolf, MD, Robert Nowak, PhD, Phillip Medicine Section, Department of Internal Medicine, smoke cigarettes despite serious health Ackerman,MS,and Neil Malone,MA.Statistician:Mark Sarees University of Nebraska Medical Center,600 S 42nd St, g p Knowles,PhD,Alza Corp,Palo Alto,Calif. Omaha,NE 68198-2465(Dr Kennard). consequences, declining social accept- JAMA,December 11, 1991—Vol 266, No.22 Smoking Cessation—Transdermal Nicotine Study Group 3133 ff t substance in tobacco products is nico- tine.2 Upon abstinence from smoking, rapid appearance of nicotine withdrawal 21-mg TN symptoms may prevent even highly mo- Trial 1 14-mg TN tivated smokers from quitting.0 5 Centers t Nicotine replacement therapy with (N=513) 7-mg TN nicotine polacrilex gum,when used with behavioral support, aids smoking ces- Placebo cation.','However, special chewing re- Phase 1 Phase 2 quirements,gastrointestinal side effects, (Downtitration) (Off-Drug Follow-up) and social prohibitions limit compliance. Our multicenter,randomized trial eval- 1 2 3 4 5 6 7 8 9 10 11 12131415 161718 uated the efficacy and safety of a new 21-mg TN Time,wk transdermal nicotine system(Nicoderm, Trial 2 Marion Merrell Dow Inc, Kansas City, 4 Centers 14-mg TN Mo, and Alza Corp, Palo Alto, Calif)as (N=422) Trial 3 an aid to smoking cessation.' Placebo Continuation Study I I I I I I (N=415) METHODS 1 2 3 4 5 6 Transdermal Nicotine Time,wk The transdermal nicotine system uses (Primary Efficacy) rate-control membrane technology to de- liver nicotine for 24 hours. Three doses of transdermal nicotine, providing 21, Fig 1.—Study designs.TN indicates transdermal nicotine.Asterisk,Only successful quitters from trials land 14, or 7 mg of nicotine over 24 hours, 2 could enroll in trial 3.Dagger,Quitters who remained abstinent in phase 1 of trial 3 were followed up for were studied; they provide average an additional 3 months. steady-state plasma nicotine concentra- tions of 17, 12, and 6 ng/mL, respec- Table 1.—Downtitration Schedules for Phase 1 of Trial 3 tively.1 By comparison,cigarette smok- ing produces average plasma nicotine Treatment by study week concentrations of 20 to 50 ng/mL,11 and Original Treatment 7 and 8 9 and 10 11 and 12 hourly use of nicotine polacrilex ther- Transdermal nicotine apy(2 mg)provides average plasma con- 21 mg/24 h 14 mg/24 h 7 mg/24 h Placebo centrations of about 8 to 12 ng/mL.11,12 14 mg/24 h 7 mg/24 h 7 mg/24 h Placebo Placebo systems contained nicotine in 7 mg/24 h Placebo Placebo Placebo the drug reservoir to mimic the odor of Placebo Placebo Placebo Placebo active systems but delivered less than 1 mg of nicotine in 24 hours, a dose not expected to affect outcome. A 24-hour dosing regimen was chosen based on a from further analysis because of major tients were evaluated at biweekly or preliminary study' that showed im- protocol violations.Because concordance monthly visits for an additional 12 weeks proved early quit rates with no increase between married subjects could bias re- while not receiving treatment. The tri- in side effects with 24-hour administra- sults, married partners were assigned als thus provided a 6-month smoking tion. Systems were applied daily to a (randomly)to identical study regimens. cessation history for each patient. Ab- clean,dry skin site on the upper torso or Only one partner chosen randomly,how- stinence at the end of weaning and after to the upper,outer arm on a 7-day cycle. ever,was included for analysis.Patients 6 months required verbal reports of ab- successfully abstaining from smoking in solutely no cigarette use and an expired Study Overview trials 1 and 2 were enrolled in trial 3 for breath carbon monoxide measurement Two parallel, 6-week, multicenter, blinded downtitration from their med- of 8 ppm or lower at all clinic visits dur- double-blind trials(trials 1 and 2)and an ications over a 6-week period and fur- ing these periods. 18-week continuation trial(trial 3)were ther follow-up for an additional 12 weeks conducted for this study report(Fig 1). receiving no treatment. Patients Two trials were conducted to provide Clinic visits were scheduled weekly Patients eligible for enrollment had a confirmatory independent results and during trials 1 and 2. Patients were to history of smoking a minimum of one to achieve a sample size large enough to stop smoking at the start of the trial but pack of cigarettes daily for at least 1 year; show a difference in outcome at 6 kept a daily diary to record any ciga- had made at least one attempt to quit; months, given the high relapse rates rette use. Smoking abstinence was de- had a baseline expired breath carbon noted with any smoking cessation ther- fined by patient diary reports of abso- monoxide level of 10 ppm or greater; apy. All study sites obtained institu- lutely no smoking confirmed by expired were generally healthy as determined tional review board approval.In trial 1, breath carbon monoxide levels of 8 ppm by medical history, physical examina- five centers enrolled 513 patients who or lower during each of the last 4 weeks tion,and routine laboratory tests;and, were randomized to one of four treat- of the 6-week trials. if female and with childbearing poten- ments: 21-, 14-, or 7-mg transdermal In phase 1 of trial 3, the successful tial,were using an acceptable method of nicotine or placebo.Trial 2 enrolled 422 abstainers from trials 1 and 2 under- birth control. All patients gave signed patients at four centers and was iden- went blinded dose reduction from their informed consent.Patients were not en- tical to trial 1 except that the 7-mg trans- previous study medications at biweekly rolled if they used smokeless tobacco, dermal nicotine dose was omitted. A clinic visits, as shown in Table 1. pipes,cigars,or nicotine gum;had med- total of nine subjects(1%)were excluded In phase 2 of trial 3, abstinent pa- ical conditions deemed inappropriate for 3134 JAMA,December 11, 1991—Vol 266, No.22 Smoking Cessation—Transdermal Nicotine Study Group Table 2.-Mean Demographic Results for the Pooled Database(Trials 1 and 2) ! Treatment Group Transdermal Nicotine Placebo All Subjects Treatment 21 mg/d(N=262) 14 mg/d(N=275) 7mg/d(N=127) (N=271) (N=935) Difference* Women,No.(%) 156 (60) 163 (59) 73 (57) 172 (63) 564 (60) NS Age,y Mean±SD 43.1±10.4 42.5±10.6 40.7±9.8 43.2±9.9 42.6±10.3 Range 22-66 22-68 23-65 2165 2168 NS Weight,kg Mean±SD 73.1±16.3 72.9±16.0 71.3±15.2 71.9±15.7 72.4±15.9 Range 43-155 43-127 45-131 44-122 43-155 NS Fagerstrom score Mean±SD 7.2±1.7 7.0±1.7 7.2±1.7 7.1±1.7 7.1±1.7 Range 2-11 2-11 3-11 2-11 2-11 NS Time smoking,y Mean±SD 24.9±10.4 24.0±10.4 22.3±9.1 24.2±9.9 24.1±10.1 Range 1-50 1-50 6-48 2-55 1-55 NS Cigarettes per day Mean±SD 31.1±10.5 31.0±10.3 29.8±8.4 30.5±10.6 30.7±10.2 Range 12-90 15-80 20-50 20.80 12.90 NS No.of previous quit attempts Mean±SD 4.4±4.2 4.1±4.4 4.1±5.4 3.8±3.3 4.1±4.2 Range 1-30 1-40 1-50 1-20 1-50 NS Baseline carbon monoxide,ppm Mean±SD 35.0±14.6 33.8±13.3 34.9±12.3 37.0±14.2 35.2±13.8 Range 10-95 5-74 13-68 5-89 5-95 NS *NS indicates not significant. study entry; or gave a history of con- lessness.Severity of each symptom was Cary, NC). An analysis of variance comitant alcohol or drug dependency. rated as 0 (none), 1 (slight), 2 (mild), model, taking baseline values into ac- During screening,patients completed a 3(moderate),or 4(severe).A mean com- count, was used to assess the smoking Fagerstrom nicotine dependency ques- bined withdrawal symptom score was reduction data, with main effects for tionnaire13 and rated their motivation to calculated for each patient as the aver- treatment and center; a Student-New- stop smoking (no minimum score was age of the individual symptom scores; man-Keuls multiple comparisons test required for inclusion)using a scale of 0 the nicotine craving symptom was also was employed to compare results among (ambivalent)to 11(maximal).Initial ran- analyzed separately. For patients who treatments. Analysis of variance was domization procedures ensured that completed trials 1 and 2 but did not stop used to compare withdrawal symptom members of the same household received smoking, smoking reduction was as- scores at each study week. Two-tailed the same treatment and preselected one sessed from daily diary reports of cig- t tests were used for.pairwise compar- member for inclusion in efficacy analyses. arette usage. isons between treatments. The sample Behavioral Support Program Statistical Methods sizes chosen for trials 1 and 2 were suf- ficient to show,with 90%power, a sig- Patients at each center attended semi- Demographic and smoking history nificant difference at the P=.05 between standardized group support sessions variables were compared across study two groups with smoking cessation rates weekly during trials 1 and 2 and biweekly centers and treatments to assess base- of 15% and 35% (the average success during phase 1 of trial 3. Each group line comparability of the treatment rates for nicotine gum in this setting). could contain a maximum of 25 patients, groups. Gender was analyzed using X2 The combined patient population from but most had five to 15 patients. The tests;the other variables were analyzed trials 1 and 2 was used in trial 3 to pro- group sessions lasted from 45 to 60 min- using an analysis of variance model. vide a sufficient sample size to assess utes and included 1 to 2 minutes per All patients were included in outcome treatment differences assuming a 50% patient for review of individual progress, evaluations except for the excluded relapse rate between the end of treat- followed by a discussion of applicable members of couples (49 patients) and ment and the 6-month follow-up. behavior modification techniques. Dis- nine patients with major protocol in- Pooled Data cussion topics were chosen from the Quit fractions. Patients who prematurely and Win Participant Manual,"which withdrew from the trials,regardless of Data from trials 1 and 2 were pooled was given to all patients at enrollment. reason,were classified as treatment fail- since the conduct of the trials was iden- ures.Pairwise comparisons of cessation tical and there were only minor demo- Withdrawal Symptoms and rates between treatment groups were graphic differences between the patient Smoking Reduction made using a two-way X1 test. populations. Quit rates were signifi- Patients recorded the severity of nic- A logistic regression analysis was used cantly different for the two trials, but otine withdrawal symptoms daily dur- to model the effects of the transdermal no treatment-by-trial interaction was Be- ing trials 1 and 2. The following with- nicotine dose, together with demo- tected at any of the efficacy time points. drawal symptoms specified by the Di- graphic and smoking history covariates, Analyses of the pooled data excluded agnostic and Statistical Manual of Men- on smoking cessation rates. The model the 7-mg transdermal nicotine group be- tal Disorders, Revised Third Edition included a treatment-by-center interac- cause this treatment was administered (DSM-III-R)were scored:nicotine crav- tion.The analysis was performed using only in trial 1. Survival time for each ing,irritability,frustration,anger,anx- the CATMOD and LOGISTIC proce- patient was calculated (beginning at iety,difficulty concentrating,and rest- dures in PC SAS Version 6.04(SAS Inc, week 3)as the number of days until the JAMA,December 11, 1991-Vol 266,No.22 Smoking Cessation-Transdermal Nicotine Study Group 3135 patient first smoked or until the patient Table 3.—Smoking Cessation Rates withdrew from the study,whichever was Smoking Cessation by Treatment Group,No.(%)of subjects smaller. Transdermal Nicotine RESULTS Time 21 mg/24 h 14 mg/24 h 7 m04 h Placebo Patients Demographic and smoking historyTrial1 data for the 935 enrolled patients did No.of subjects 121 121 121 124 not differ significantly among treatment End of 6 wk 86 (71)" 65 (54)• 58 (48)t 41 (33) groups (Table 2). Almost all subjects Tdal2 were white.Overall,250 patients(27%) No.of subjects 128 133 129 withdrew from the study prematurely, End of 6 wk 65 (51)' 58 (44)* 27 (21) 151 patients(23%)receiving active treat- Pooled Database* ment and 99 patients (37%) receiving No.of subjects 249 254 253 placebo (P<.001). Reasons for with- End of 6 wk 151 (61)* 123 (48)` 68 (27) drawal included lack of efficacy (34%); End of 12 wk 98 (39)" 69 (27)§ 40 (16) personal reasons,such as losing a job or End of 24 wk 65 (26)• 46 (18) 31 (12) moving from the area(18%);poor com- pliance or unavailability for follow-up •P:5.001 compared with placebo. (34%);and unrelated illness(6%). Nine tP<.05 compared with placebo. *Smoking cessation rates at the end of 12 and 24 weeks are from the continuation trial.The 12-week time point percent of patients were withdrawn by was the end of downtitration;24 weeks was the end of a 3-month follow-up with no treatment. the investigators for adverse effects.Sig- §P<.01 compared with placebo. nificantly fewer patients receiving ac- tive treatment than placebo withdrew due to lack of efficacy (P<.05); with- drawals due to the other reasons oc- curred with similar frequency between 100 treatment groups. Of the 367 patients eligible for trial 3, 90- • Placebo(n=253) 355 (97%) enrolled. A total of 152 pa- o 21-mg TN(n=249) tients underwent downtitration from 21- 80 a 14-mg TN(n=254) mg transdermal nicotine and 134 from 14-mg transdermal nicotine,and 69 were o 70 already receiving placebo. Ninety-five a patients (26%) withdrew prematurely o 60 during downtitration,and an additional E 59(16%)withdrew during the off-treat- o 50 ment follow-up.A similar percentage of z patients was discontinued from each - 40 treatment group. °' Quit Rates a 30 The 6-week smoking cessation rates 20 for 21-and 14-mg transdermal nicotine in both trials 1 and 2 were significantly 10 Primary higher than for placebo(P--.001,Table Efficacy Downtitration Off-Drug Follow-up 3).In trial 1,7-mg transdermal nicotine 0 also provided quit rates superior to those 0 2 4 6 8 10 12 14 16 18 20 22 24 of placebo(P<.02).The increase in ces- sation rates with transdermal nicotine End of Study Week was quite consistent across the nine cen- ters, with no significant treatment-by- center interactions in the analyses. At Fig 2.—Time to smoking for the pooled database.The first 2 study weeks were not considered in the anal- the 6-month time point (week 24), ab- yses of efficacy.TN indicates transdermal nicotine. stinence rates for the 21-mg transder- mal nicotine group remained signifi- cantly higher than those for the placebo the dose-response relationship. How- Withdrawal Symptom group(P<.001),while those for the 14- ever,regardless of treatment,patients Severity Scores mg transdermal nicotine group were who smoked more cigarettes at baseline marginally higher(P=.066).At all time were less likely to quit, as were those Combined nicotine withdrawal symp- points,the 21-mg transdermal nicotine with the lowest self-rated motivation to tom scores and nicotine craving scores group did significantly better than the quit. were significantly lower for 21-and 14- 14-mg transdermal nicotine group Relapse rates during downtitration mg transdermal nicotine than for pla- (P=.031). were similar for all treatment groups, cebo at every study week(P<.001,Fig None of the demographic or smoking indicating that the dose-reduction steps 3).Withdrawal symptoms were also sig- history variables(including Fagerstrom were effective. Times to smoking for nificanty lower for 7-mg transdermal nic- score and baseline cigarette use)inter- the patients in the pooled database are otine than for placebo(P<.02).Twenty- acted with the treatment response or shown in Fig 2. one-milligram transdermal nicotine was 3136 JAMA,December 11,1991—Vol 266,No.22 Smoking Cessation—Transdermal Nicotine Study Group Compliance kg on active treatment and 2.6±1.5 kg Compliance was assessed by compar- on placebo after 6 weeks. Weight gain a ing the number of used and unused sys- was similar among the treatment groups. 2 (Moderate)3 terns returned each week with the num- For abstinent patients,mean heart rate m ber of elapsed days between visits.Over decreased from baseline values by 0.5, 3 in 90%of patients in trials 1 and 2 receiv- 0,2,and 6 beats per minute for 21-, 14-, (Mild)2 and 7-m transdermal nicotine and 1 - E ing both active and placebo treatment g P a a a used systems daily at least 90%of the cebo,respectively;the decrease was sig- - ; time. Over 90% of the patients rated nificant only for the placebo group 0 rn (Slight) 1 transdermal nicotine he convenient and (P<.05). Transdermal nicotine con- comfortable to use. ment was not associated with any con- sistent changes in electrocardiograms 041 1 1 1 or routine hematology and chemistry 1 2 3 4 5 6 S8fety blood tests. Study week Transdermal nicotine was well toler- ated systemically and topically, and no c (Moderate)3 serious systemic adverse events oc- The efficacy of transdermal nicotine Z curred. About 50% of patients in each as an aid to smoking cessation was ap- z treatment group(including the placebo parent despite the positive influence15,16 c (Mild)2 group)reported adverse effects. Many of the group counseling provided to all of these symptoms (eg, headache, in- patients. Six-week smoking cessation -> somnia,dizziness)may have been due to rates for all doses of transdermal nico- (Slight)1 withdrawal rather than the study med- tine were significantly higher than for ications. placebo, with 61% of patients who re- Eleven (4%), 15 (5%), 1 (1%), and ceived 21-mg transdermal nicotine 0 6(2%)of the patients receiving 21-, 14-, achieving abstinence. As with all cur- 1 2 3 4 5 6 and 7-mg transdermal nicotine and pla- rent treatment modalities for smoking Study week cebo, respectively, withdrew prema- cessation, relapse rates during follow- Symptom Scale Treatmentturely from the trials due to adverse up were substantial for all treatment 4=Severe Placebo effects. Only six (0.9%) of the 664 pa- groups. Still, one in four patients orig- 3=Moderate 21-mg TN/24 h tients(0.9%)receiving transdermal nic- inally treated with 21-mg transdermal 2=Mild v 9-ti 14-mg TN/24 h otine(four receiving 14-mg and two re- nicotine remained a nonsmoker at 1=Slight o-e-•o7_mg TN/24 h ceiving 21-mg transdermal nicotine) 6 months compared with one in nine for o=None withdrew during the first week of treat- the placebo group. ment;their symptoms included nausea, All doses of transdermal nicotine sig- vomiting, dizziness, headache, insom- nificantly alleviated nicotine withdrawal Fig 3.—Mean combined withdrawal symptom and ma, and/or general malaise. The only symptoms, including craving. Reduc- craving for nicotine scores for the pooled database dose-related adverse effects of trans- tions in withdrawal symptom severity (trials f and 2)for weeks 1 through 6(N=877).TN dermal nicotine therapy were generally were similar for all transdermal nico- indicates transdermal nicotine,Asterisk,Seven in- mild to moderate sleep disturbances(eg, tine doses, supporting the hypothesis y patients daily.These individual symptom scores ys- y nicotine withdrawal symptoms were rated binsomnia and abnormal dreams), dthat relatively small amounts of nico- ypaie were averaged to provide a mean combined with- pepsia,various myalgias and body aches, tine are effective in moderating many drawal symptom score for each study week. and increased cough. Conversely, som- withdrawal Symptoms."Craving for nic- nolence and headache were reported otine, however, responded better to more frequently by placebo-treated pa- higher transdermal nicotine doses. superior to the other active doses in tients. The incidence of almost all side Although transdermal nicotine pro- reducing nicotine craving during the first effects decreased considerably after the vides smokers with nicotine,plasma lev- 3 weeks of abstinence(P<.05).Craving first few days of treatment. els of nicotine from transdermal nico- for nicotine did not increase significantly Almost 50% of the 664 patients re- tine are lower than from a usual smok- when patients were weaned from active ceiving active treatment and 13% re- ing regimen.' In addition, other con- treatment. ceiving placebo reported transient itch- taminants associated with cigarette ing or burning at skin sites,particularly smoke are not present.While a prepon- Smoking Reduction following system application or removal, derance of data demonstrates the car- The 242 patients who completed tri- but only 14% had definite or severe cinogenic and other harmful effects of als 1 and 2 but did not stop smoking erythema noted at skin sites at least tobacco smoke,1'nicotine itself is not a significantly reduced the number of once during the trials. Sixteen patients carcinogen.19 cigarettes they smoked compared receiving active treatment(2.4%)were Transdermal nicotine is likely to have with baseline. On average, patients suspected to have developed contact sen- a lower abuse potential than smoking. receiving 21-, 14-, or 7-mg transder- sitization;14 of them were women.Sen- Transdermal nicotine differs from smok- mal nicotine who did not stop smok- sitization was confirmed in 11 of the 16 ing in four characteristics considered to ing reduced their cigarette use 66%, patients by rechallenge and was unre- contribute to abuse: much slower ab- 67%, and 19% more, respectively, lated to the dose of transdermal nicotine sorption, smaller fluctuations in blood than did patients in the placebo received. Follow-up information was levels, lower blood levels of nicotine, group. Reductions for the 21- and available for seven sensitized patients and less frequent use(ie, once day).21 14-mg transdermal nicotine treatment who resumed smoking;none developed In conclusion,this study suggests that groups were significantly higher than systemic symptoms. transdermal nicotine replacement ther- for the 7-mg transdermal nicotine and Body weight increases for patients apy can be an important adjunct to smok- placebo groups (P<.05). who stopped smoking averaged 2.0±1.9 ing cessation. Promising results have JAMA, December 11, 1991—Vol 266, No.22 Smoking Cessation—Transdermal Nicotine Study Group 3137 also been reported for other transder- and social acceptability of a transdermal consultancies and honoraria for educational activi- mal nicotine products currently under dosage form. ties.Authors employed by Marion Merrell Dow Inc. i (Drs Rolf and Nowak and Messrs Ackerman and i development.2� While differences lri Malone) and those employed by Alza Corp (Drs transdermal nicotine products will likely This investigation was supported by a grant from Causey and Knowles and Mss Voss-Roberts, appear, given their different physical Alza Corp. Drs Christen, Hatsukami, Rennard, Prather,Trunnell,and Moos)own shares of com- Lichtenstein, Heatley, Repsher, Fortmann, pany stock.Dr Biglan's spouse owns stock in Alza and drug-releasing properties, they Killen,Hughes,and Glover and Mr Daughton have' Corp. should share the improved compliance received fees from Marion Merrell Dow Inc for References Pierce JP,Fiore MC,Novotny TE, 9. GorslineJ,Gupta SK,Dye D,Rolf CN.Nicotine 17. Hughes JR,Gust SW, Keenan RM, Fenwick Hatziandreu EJ, Davis RM. Trends in cigarette dose relationship for Nicoderm(nicotine transder- JW.Effect of dose on nicotine's reinforcing,with- smoking in the United States:projections to the mal system) at steady state. Pharm Res. 1991; drawal-suppression and self-reported effects. J year 2000.JAMA. 1989;261:61-65. 10(suppl):5299. Pharmacol Exp Ther. 1990;252:1175-1183. 2. Report of the Surgeon General.The Health Con- 10. Russell MA,Raw M,Jarvis W.Clinical use of 18. Hoffmann D, Hecht SS,Wynder EL.Tumor sequences of Smoking:Nicotine Addiction.Wash- nicotine chewing-gum.BMJ. 1980;280:1599-1602. promoters and cocarcinogens in tobacco carcino- ington,DC:US Dept of Health and Human Ser- 11. Benowitz NL,Jacob P III,Savanapridi C.De- genesis.Ent^ironHealth Perspect.1983;50:247-257. vices;1989.Publication CDC 88-8406. terminants of nicotine intake while chewing nico- 19. Benowitz NL.Pharmacologic aspects of ciga- 3. West R, Schneider N. Craving for cigarettes. tine polacrilex gum. Clin Pharmacol Ther. 1987; rette smoking and nicotine addiction. N Engl J Br J Addict. 1987;82:407415. 41:467473. Med. 1988;319:1318-1330. 4. West RJ,Hajek P,Belcher M.Severity of with- 12. McNabb ME,Ebert RV,McCusker K.Plasma 20. Hughes J.Dependence potential and abuse li- drawal symptoms as a predictor of outcome of an nicotine levels produced by chewing nicotine gum. ability of nicotine replacement therapies. Biomed attempt to quit smoking.Psychol Med.1989;19:981- JAMA. 1982;248:865-868. Pharmacother. 1989;43:11-17. 985. 13. Fagerstrom KO,Schneider NG.Measuring nic- 21. Abelin T, Buehler A, Muller P, Vesanen K, 5. Gritz ER, Carr CR, Marcus AC. The tobacco otine dependence:a review of the Fagerstrom tol- Imhof PR.Controlled trial of transdermal nicotine withdrawal syndrome in unaided quitters. Br J erance questionnaire. J Behan Med. 1989;12:159- patch in tobacco withdrawal.Lancet. 1989;1:7-10. Addict. 1991;86:57-69. 182. 22. Hurt RD,Laugher GG,Offord KP,Kottke TE, 6. Fagerstrom K0. Efficacy of nicotine chewing 14. Minnesota Heart Health Program. Quit and Dale LC.Nicotine-replacement therapy with use of gum:a review.In:Pomerleau OF,Pomerleau CS, Win Participant Manual: Consumer's Guide to a transdermal nicotine patch:a randomized double- Fagerstrom KO,HenningfieldJE,HughesJR,eds. Smoking Cessation. Minneapolis, Minn: Regents blind placebo-controlled trial.Mayo Clin Proc.1990; Nicotine Replacement:ACritical Evaluation.New of the University of Minnesota;1989. 65:1529-1537. York,NY:Alan R Liss Inc;1988:109-128. 15. Kottke TE,Battista RN,DeFriese GH,Brekke 23. Buchkremer G,Bents H,Horstmann M,Opitz 7. Lam WL, Sze PC, Sacks HS, Chalmers TC. ML.Attributes of successful smoking cessation in- K, Tolle R. Combination of behavioral smoking Meta-analysis of randomized controlled trials of nic- terventions in medical practice:a meta-analysis of cessation with transdermal nicotine substitution. otine chewing gum.Lancet. 1987;2:27-29. 39 controlled trials.JAMA. 1988;259:2882-2889. Addict Behan. 1989;14:229-238. 8. Daughton DM,Heatley SA,Prendergast JJ,et 16. Schwartz JL.Retnew and Evaluation of Smok- 24. Tonnesen P,Norregaard J,Simonsen K,Sawe al. Effect of transdermal nicotine delivery as an ing Cessation Methods:The UnitedStates and Can- U. A double-blind trial of a 16-hour transdermal adjunct to low-intervention smoking cessation ther- ado.Bethesda,Md:US Dept of Health and Human nicotine patch in smoking cessation.N Engl J Med. apy.Arch Intern Med. 1991;151:749-752. Services;1987.Publication NIH 87-2940. 1991;325:311-315. 3138 JAMA,December 11, 1991-Vol 266,No.22 Smoking Cessation-Transdermal Nicotine Study Group 1 Predictors of Physicians' Smoking Cessation Advice Erica Frank, MD, MPH; Marilyn A. Winkleby, PhD; David G. Altman, PhD; Beverly Rockhill, MA; Stephen P. Fortmann, MD Objectives.—To determine the percentage of smokers reporting that a phy- some organizations(such as the National sician had ever advised them to smoke less or to stop smoking, and the effect Cancer Institute of the National Insti- of time, demographics, medical history, and cigarette dependence on the like- tutes of Health,Bethesda,Md,and the lihood that respondents would state that a physician had ever advised them to American Academy ofFan lyPhysicians, stop smoking. Kansas City,Mo)have created kits and Design and Setting.—Data were collected from the Stanford Five-City training programs and have mounted Project, a communitywide health education intervention program. The two large public education campaigns. treatment and three control cities were located in northern and central Califor- nia.As there was no significant difference between treatment and control cities See also pp 3172 and 3183. regarding cessation advice, data were pooled for these analyses. Participants.—There were five cross-sectional, population-based Five-City Encouragement of physician partici- Project surveys (conducted in 1979-1980, 1981-1982, 1983-1984, 1985-1986, pation is well justified.While some stud- and 1989-1990);these surveys randomly sampled households and included all ies have found that physician counseling residents aged 12 to 74 years. does not increase long-term smoking ces- Main Outcome Measures.—Improved smoking advice rates overtime in all sation rates,'most studies have shown towns was an a priori hypothesis. that physicians' encouragement of pa- Results.—Of the 2710 current smokers,48.8%stated that their physicians tient', smoking cessation attempts in- had ever advised them to smoke less or stop smoking.Respondents were more creases the stop smoking.3,1 that patients will . Furthermore, physi- likely to have been so advised if they smoked more cigarettes per day, were cians will increase their counseling ef- surveyed later in the decade,had more office visits in the last year,or were older. forts when trained to do so.','Despite In 1979-1980, 44.1% of smokers stated that they had ever been advised to the documented efficacy ofphysician ad- smoke less or to quit by a physician,vs 49.8%of smokers in 1989-1990(P<.07). vice,a 1975 study'found that only 35% Only 3.6%of 1672 ex-smokers stated that their physicians had helped them quit. of men who smoked and 38%of women Conclusion.—These findings suggest that physicians still need to increase who smoked had ever been told to quit by smoking cessation counseling to all patients,particularly adolescents and other a physician,and a 1980/1983 study found young smokers, minorities, and those without cigarette-related disease. that only 42%of male smokers and 46% (JAMA. 1991;266.3139-3144) of female smokers had ever received physician advice to quit smoking.' The objectives of this study were to determine (1) the percentage of 2710 current smokers reporting that they had From the Stanford Center for Research in Disease IT HAS BEEN estimated that during ever received physician advice to smoke Prevention and the Department of Medicine,Stanford the 1990s, approximately 3 million to- less or to stop smoking(henceforth re- University School of Medicine,Palo Alto,Calif. bacco-related deaths will occur world- (erred to as smoking cessation advice); Presented at Prevention'91,the annual meeting of g the American College of Preventive Medicine,March wide annually.'Throughout the 1980s, (2)the effect ofrespondents'demograph- 1s,1991. many health organizations have called ics, medical history, and cigarette de- Reprint requests to Stanford Center for Research in Disease Prevention,Stanford University School of Med- on physicians to diminish this pandemic pendence on respondents'likelihood to icine, 1000 welch Rd, Palo Alto,CA 94304-1885(Dr by counseling patients to stop smoking. be offered smoking cessation advice;(3) Fortmann). To help physicians with this counseling, changes in smoking cessation advice JAMA,December 11, 1991—Vol 266,No.22 Smoking Cessation Advice—Frank et al 3139 rates from 1979-1980 to 1989-1990; (4) Table 1.—Prevalence of Smokers in Study Popula- the percentage of 1672 ex-smokers cit- tion* ing their physician as an influence on smoking 100 +Males their decision to quit;and(5)the valid- Characteristics Prevalence,i o 90 ity of respondents'reports that they had Survey years(No.of subjects) Y 80 Females received smoking cessation advice from 1979-1980(2487) 30.4 0 h 1981-1982(2330) 27.7 E 70 physicians. 1983-1984(1846) 27.4 We examined these objectives by eval- 1985-1986(2360) 20.2 N 60 - uating data collected between 1979 and Total�1 �05) :° a 50 19W from the Stanford Five-City Project Age,y ° 40 - (FCP),a communitywide health educa- 12-17 9.5 m 18-24 23.0 0 30 i tion program designed to change favor- 25-49 28,5 a 20 ably the prevalence and sequelae of car- 50-74 24.7 diovascular disease risk factors.A more G der 23 s w 10 complete description of the FCP design M 26.4 p has been published elsewhere.'The FCP origpanic white 26.4 in 1979- 1981- 1983- 19887 5- 19 - 1989- has thus far demonstrated improvement Hispanic 22.1 1980 1682 1984 1986 1988 1990 in general cardiovascular disease knowl- Education(subjects>25 y) Years edge, smoking cessation rates, blood <High school 35.0High school graduate 32.8 pressure, resting pulse rate, coronary Some college 28.4 heart disease risk scores,and total mor- Household income($/y)'College17'0 9 Fi 1.—Smoking cessation advice rates from 1979 tality risk scores in treatment cities over <10 000 27.1 to 1990,reported for males and females. control cities.' 10 000 to 19 999 30.3 20 000 to 29 999 27.2 30 000 to 39 999 24.2 tories was collected by asking "has a METHODS >40 000 20.5 doctor ever told you that you have had The FCP included two treatment cit- a up erowbar 34.1 a heart attack(or myocardial infarction)? ies(total population, 122 800),two tori- white collar 21.4 A stroke?" and "has a health profes- trol cities for comparison(total popula- Homemaker 23.1 sional(doctor,nurse,etc)ever told you tion, 197 500), and one additional con- *smoking prevalence is the percentage of the num- that you have high blood pressure? If trol city for monitoring morbidity and ber of subjects that smoke. yes, are you under treatment for it?" mortality rates only. Representative cross-sectional population surveys were RESULTS conducted from May 1979 to April 1980, than were participants. We eliminated Smoking Prevalence May 1981 to July 1982, June 1983 to from all analyses (except Table 1) 154 The prevalence of smoking among re- June 1984,April 1985 to June 1986,and current smokers who did not answer spondents was 25.1%(2864 of 11428 re- April 1989 to May 1990; these surveys questions regarding smoking cessation spondents)for all five surveys combined randomly sampled households and in- advice offered by physicians. All re- (Table 1). Smoking prevalence was cluded all household residents aged 12 ported P values are for two-tailed tests. 202% for the 1986 survey, somewhat to 74 years. In addition, the original Only the 2255 non-Hispanic white smok- lower than the 1986 smoking prevalence cross-sectional sample was restudied bi- ers and the 273 Hispanic smokers are re- rate of 24.5% for the state of Califor- ennially as a cohort.By 1990,five cross- ported in the stratifications by ethnicity nia.11(The year 1986 was chosen for com- sectional and five cohort surveys had (due to the small samples of`other"races parison because it was midway through been completed.Cross-sectional data are in each survey). All other stratifications our survey and preceded recent changes used in the current study,with the ex- include all smokers regardless of ethnic- in California's smoking legislation.)As ception of the examination of recall bias, ity.Since respondents'self-identified eth- elsewhere in California and the rest of which uses cohort data.Treatment-and nicity was not collected in the 1985-1986 the country,12 smoking peaked in the control-city data are pooled because survey,we coded ethnicity using a Span- middle adult years,and males were more there was no substantial difference in ish surname program developed by the likely to smoke than were females reports of receiving smoking cessation US Census Bureau,10 internally validated (P<.002).Also similar to both state and advice from physicians between treat- the method by examining data from the national statistics, the data showed an ment and control cities. In the 1979- 1979-1980 and 1989-1990 surveys,and in- increased prevalence of quitters in older 1980 survey,708 smokers were included cluded only non-Hispanic whites and His- age groups, and a higher prevalence of in our analyses; in 1981-1982, 620; in panics in the models. quitters among males than among fe- 1983-1984,473;in 1985-1986,453;and in Participants were classified as ciga- males. Ex-smokers comprised 23.4% 1989-1990,456(a total of 2710 smokers). rette smokers if they reported ever (n=2674) of the total population. Information was collected on site at smoking cigarettes on a daily basis and permanent survey centers using prima- had smoked one or more cigarettes in Temporal and Demographic Factors rily self-administered questionnaires(al- the last week. Any participant whose There was some net improvement in though some questions were adminis- plasma thiocyanate level exceeded 100 respondents'likelihood to be given smok- tered by trained staff who were usually µmol/L and expired-air carbon monox- ing cessation advice overthe 11-yearstudy unknown to the respondents).There was ide level exceeded 8 ppm was also clas- period,although in females this improve- an overall response rate for the cross- sified as a smoker. Participants were ment leveled off,and in males it regressed sectional surveys of 62.8%. An abbre- classified as ex-smokers ifthey reported (Fig 1). There were strong demographic viated telephone or in-person question- ever smoking on a daily basis, had not differences between those who reported naire completed by nonrespondents re- smoked in the last week,and had plasma receiving smoking cessation advice and vealed that they were less likely to be thiocyanate and expired-air carbon mon- those who did not(Table 2). English-speaking, were somewhat less oxide levels less than those mentioned There was a strong relationship be- educated,and were more likely to smoke above.Information on past medical his- tween a respondent's age and the like- 3140 JAMA,December 11, 1991—Vol 266,No.22 Smoking Cessation Advice—Frank et al Table £.-Smokers' Likelihood to Report Ever Having Been Told to Quit Smoking by a Physician and the Effect of Demographics,Medical Risk Factors, %of Smokers and Cigarette Use* Ever Told to %of Smokers Quit If Physician Ever Told to Quit Seen in Past Year %of Smokers Ever Told to Characteristics M F Total M F Total %of Smokers Quit If Physician Ever Told to Quit Seen in Past Year Hospitalized(all causes)in past year Characteristics M F Total M F Total Yes 54.8 55.3 Survey years(No.of subjects) No 48.1 51.6 1979-1980(708) 43.2 44.9 44.1 47.8 47.1 47.4 P <.05 NS 1981-1982(620) 39.5 49.4 45.0 42.2 51.8 48.1 Oral contraceptive use(for 1983-1984(473) 53.3 56.1 54.8 61.3 59.2 60.1 women younger than 35 y) Yes 44.1 43.6 1985-1986(453) 53.6 54.1 53.9 55.3 56.2 55.8 No 41.6 43.2 1989-1990(456) 44.4 55.4 49.8 45.4 59.7 53.0 P NS NS All smokers(2710) 46.1 51.2 48.7 49.9 53.8 52.1 History of myocardial Pt NS <.01 <.07 NS <.02 NS infarction or cerebrovascular Age y accident 12-17 13.4 13.9 Yes 84.4 87.7 18-24 26.7 29,5 No 47.2 50.3 25-49 48.3 51.6 P <.001 <.001 50-74 64,9 89.2 History of hypertension Yes 65.3 68.3 P 5,001 5.001 No 44.0 46.9 Ethnic origin P <.001 <.001 Non-Hispanic white 51.1 54.1 Hispanic 32.6 36.6 Currently treated hypertension Yes 78.3 79.0 P 5001 5.001 o 55.5 58.8 Education(subjects>_25 y) NN5.001 5.001 <High school 53.8 58.8 High school graduate 54.3 59.3 Cigarettes,Nodd Some college 52.4 55.2 0.9 28.7 30.810.20 50,9 54,7 z College 54.8 57.3 21-40 63.8 69.1 P NS NS 2!41 63.5 63.0 Household income($/y) p 5.001 5.001 <20 000 47.6 52.0 20 000-39 999 51.5 55.6 Years smoked >_40 000 56.3 57.1 0-4 21.3 23.5 5-9 34.8 38.3 P 5.02 Ni- Occupation 10-14 48.6 49.8 Blue collar 45.8 49.4 >15 58.8 63.1 White collar 51.1 53.2 P <.001 <.001 Homemaker 56.4 59.1 No.of attempts to quit 0 36.2 41.6 P 5.01 <,06 1 3 48.4 50.2 No.of office visits in past year 0 35.0 4-6 56.4 58.7 at 52.1 >7 54.7 59.4 1-3 47.2 P <.001 <.001 4-6 58.3 *P values are calculated for X°that included all cells.NS indicates not significant. >_7 60.5 tP value is for X'comparison of 1979-1980 and 1989.1990 data. P <.001 lihood that he or she would be given 3.3;P<.001), and in the rate of having much more likely to report having had smoking cessation advice, with the el- seen a physician at least once in the last smoking cessation advice offered by phy- dest group being nearly five times as year (female, 87.7%; male, 72.8%; sicians than were Hispanics,and this trend likely as the youngest to be advised. P<.001), may partly explain females' persisted when the data were stratified Nonetheless,respondents aged 12 to 17 higher smoking cessation advice rates by respondent's having seen a physician years did have many potential oppor- (P<.009).Interestingly,men with a his- in the last year. While other minority tunities to receive smoking cessation ad- tory of myocardial infarction or stroke groups also seemed less likely than did vice. These adolescents had smoked a seemed somewhat more likely to be ad- non-Hispanic white subjects toreport phy- mean of 2.0 years, 78.7% of them had vised against smoking than did women sician advice,the number of subjects was smoked for at least a year, and they with cardiovascular disease(88.7%of 62 small(mean number of subjects per sur- were more likely than any other age men with cardiovascular diseases were vey for all"other"smokers,45). group to have seen a physician in the advised; 79.3% of 53 women with car- Although there appeared to be a trend last year(87.8%vs --83.6%). diovascular diseases were advised; P, by educational level, once we limited Differences by gender in mean num- not significant [NS]), the effect of age(by including only those ber of office visits (female, 5.4; male, Non-Hispanic white subjects were respondents aged 25 years or older), JAMA, December 11, 1991-Vol 266,No,22 Smoking Cessation Advice-Frank et al 3141 ' educational level did not affect respon- smoked fewer cigarettes (those who dents' likelihood to be given smoking smoked 0 to 9 cigarettes per day had 4.6 cessation advice. Differences between visits per year; 10 to 20 cigarettes, 4.3 100 -- income categories for respondents'like- visits;21 to 40 cigarettes,4.3 visits;more e�.a lihood to report being given smoking than 40 cigarettes, 3.8 visits; P, NS). 80 cessation advice showed a statistically Similarly,the number of years a patient m significant trend.Regarding occupation, smoked and number of attempts to quit n 60 although homemakers'higher smoking were correlated positively with a his- 8 cessation advice rates may be partially tory of physician advice to cut down or a 40 explicable by their higher mean office quit;relationships between smoking ces- visits(6.0;P<.001),white-collar work- sation advice rates and all three smoking- 20 ers had fewer mean office visits than did related factors leveled off at higher 9'4 3.6 1.6 1.2 blue-collar workers(3.3 vs 4.0,respec- values. 41 M On Own Family Physician's Sell-help Group Lively;P<.07)but received smoking ces- or Help sation advice more often. Ex-Smokers Friends Consistent with other studies,T3 quit- Medical Factors ting on one's own was the most fre- With the exception of women using quently cited method of smoking cessa- Fig 2.—Contributors to smoking cessation(respon- oral contraceptives, smokers with his- tion(Fig 2).Only 3.6%of 1672 ex-smok- dents were permitted to check as many influences tories of relevant medical problems ers stated that their physicians helped as applied). were more likely to report receiving them quit smoking(this ex-smoker ques- smoking cessation advice than were tion was available in the 1979-1980,1985- denied ever receiving smoking cessa- smokers without histories of medical 1986, and 1989-1990 surveys), and only tion advice when requestioned in 1988- problems. This finding persisted after 6.717c of 1096 ex-smokers stated that their 1989.The high recall rates of this group controlling for age,gender,and number physician or another health professional may not be representative,however,as of office visits. was the most influential person in their only 23.6%of the 755 smokers enrolled The number of office visits was an decision to quit (not shown; this ques- at baseline remained in the cohort important predictor of respondents'like- tion was available in the 1979-1980 and throughout the decade. lihood to report being given smoking 1985-1986 surveys). cessation advice. This was reflected in COMMENT both the separate analysis of the effect Regression Analyses Half of all smokers surveyed reported of office visits and in the effect of strat- Forward logistic regression analyses that they had never been advised to quit ifying for one or more office visits on were conducted to rank those variables smoking.Respondents were more likely other variables.Although number of of- that independently predicted respon- to have been so advised if they smoked Tice visits was an important predictor of dents'receiving smoking cessation ad- more cigarettes per day, were older, advice, 39.5% of those smokers with vice. Time of survey(years from base- were surveyed later in the decade, or seven or more office visits in the last line), age, gender, ethnicity, office vis- had more office visits in the last year. year reported never receiving smoking its per year,and cigarettes per day were Less than 5%of 1672 ex-smokers stated cessation advice. forced into the model. Offered for in- that their physicians had helped them We examined whether oral contra- elusion in the model (at P<.05) were quit. ceptive users' lower rates of smoking occupation, number of years a patient cessation advice, when compared with smoked, number of attempts to quit, Potential Bias other women younger than 35 years, and the first-order interactions of time It is often difficult to assess the reli- improved over time.We found that they and gender, and time and ethnicity. In ability and validity of patient reports. did; smoking cessation advice rates the population free of cardiovascular dis- Perhaps all patients are equally and among oral contraceptive users were ease,cigarettes per day(P<.0002),age highly likely to be advised to quit smok- 32.4%in 1979-1980,38.7%in 1981-1982, (P<.0002), time (P<.0005), and office ing,and we are simply testing the prob- 50.0%in 1983-1984,55.6%in 1985-1986, visits per year(P<.02)were significant ability that individuals will recall and and 58.3%in 1989-1990(n, 118;P,NS). independent predictors of having re- consistently admit that a physician has Rates for women younger than 35 years ceived smoking cessation advice.Among told them to stop smoking. Nonethe- who were not using oral contraceptives those respondents with a history of hy- less, we found that, 9 years after first showed no trend (41.5% in 1979-1980, pertension, myocardial infarction; or testing, less than 10% of respondents 33.9%in 1981-1982,52.5%in 1983-1984, stroke,only cigarettes per day(P<,007), who originally stated that they had been 45.8%in 1985-1986, and 41.0%in 1989- and time(P<.03)were significant pre- advised to quit reversed their answers 1990, [n, 471; P, NS). dictors of respondents'likelihood to re- and stated that they had never been ceive smoking cessation advice. advised to quit.This suggests that these Smoking-Related Factors respondents have a high rate of recall There was a strong positive relation- Recall Bias and a consistent willingness to admit ship between heavier smoking and like- To test the importance of recall bias, receiving smoking cessation advice. lihood to have been advised by a phy- we evaluated the cohort sample to de- An independent confirmation of the sician to quit;nearly two thirds of those termine how many subjects who stated validity of these recall data is found in respondents who smoked more than one in 1979-1980 that they had been told by an ancillary study conducted with 318 pack per day reported being given smok- a physician to stop smoking reversed physicians from two of the four cities ing cessation advice. This was true de- their answers and stated in 1988-1989 from which this study's sample was spite the fact that those respondents that they had never been told to stop. taken.14 There was a strong correspon- who smoked more cigarettes per day From the cohort sample,we found that dente between physicians who reported were less likely to have visited a phy- of 55 smokers who stated in 1979-1980 offering smoking cessation advice and sician in the year before than those who that they had been advised to quit,9.1% smokers who reported smoking cessa- 3142 JAMA,December 11,1991—Vol 266,No.22 Smoking Cessation Advice—Frank et al tion advice offered by physicians.Fifty- lihood of smoking cessation advice be- 55.9%of women who smoked and were one percent of physicians stated that ing offered to non-Hispanic whites vs taking oral contraceptives stated that they advised symptom-free smokers vs Hispanics also requires attention, al- they had never been counseled to stop our patient-reported advice rates of 52% though explanations for these differences smoking.Denial or forgetting may play for those who had seen a physician in are also unclear. It is clear, however, a role in these patients who said that the last year.Furthermore,88%of phy- that as cigarette manufacturers increas- they were never told to quit. Nonethe- sicians stated that they advised all or ingly target Hispanics and other ethnic less,physicians should aim to be so clear, almost all patients with chronic smoking- minorities,physicians should help these memorable, and repetitive in their at- associated disease to quit, and 88% of groups combat tobacco companies' ef- tempts to provide smoking cessation ad- those patients with a history of myo- forts to promote smoking." vice for these groups that it is nearly cardial infarction or stroke who had seen Regarding cigarette-related factors, impossible for patients to forget or deny a physician in the last year reported there was a strong relationship between counseling. having been so advised. However,con- respondents'likelihood to be given smok- clusions from these and all other data ing cessation advice and their degree of Past Studies and Changes in Advice should be tempered by our response rate addiction, as seen in number of ciga- Over Time of 62.8%. rettes smoked per day, years having Our analysis showed many consisten- smoked,and number of unsuccessful at- cies with the other large study by Anda Discussion of Findings tempts to quit.Although previous stud- et all of characteristics of smokers who There are many possible explanations ies have hypothesized that the reason reported ever being advised to quit for the nearly fivefold increase from the heavier smokers receive more smoking smoking.'The study by Anda et al also youngest age group to the eldest in the cessation advice is because they have found that whites, heavier and longer- likelihood to have been advised to quit more illnesses and therefore have more term smokers, those with more office smoking;physicians may not take smok- physician contacts,'we found that lighter visits, and those with a history of my- ing histories from younger patients;ad- smokers had a higher mean number of ocardial infarction or stroke were more olescents might not admit to their phy- physician contacts per year than did likely to receive smoking cessation ad- sicians that they smoke;physicians may heavier smokers. Like cigarettes per vice, and that oral contraceptive users not wish to embarrass adolescents with day, years having smoked and number were not more likely to be advised to counseling in front of their parents;phy- of attempts to quit may reflect an indi- quit smoking. The study by Anda et al sicians may not view smoking among vidual's inability to quit. Nevertheless, also found an increase over time in re- otherwise healthy younger patients as even if degree of addiction were an in- spondents' likelihood to be advised to important;and older people have more dependent predictor of respondents who quit smoking, from 42%of adult smok- smoking-related diseases and have had received smoking cessation advice,there ers who had seen a physician in 1980/ more opportunities to be advised to quit remains the question of causality.Were 1981 to 46%in 1982/1983.Another con- smoking.Another explanation is the high people who made more attempts to quit firmation that smoking cessation advice correlation(r=.7)between age and num- trying to stop because their physicians had increased over time is a 1975 na- ber of years patients have smoked;phy- were especially concerned,or did those tional survey of smokers that found that sicians were more likely to counsel trying to quit ask their physicians for only 38% of women who smoked and longer-term smokers. Age may simply help and advice?Did higher.pack-years 35%of men who smoked had ever been serve as a (highly convenient) marker increase physicians'likelihood to advise advised to quit by a physician.' Addi- for the number of years patients have individuals to quit smoking, or did it tionally, the aforementioned survey of smoked. This does not alter the recom- simplyprovide agreatercumulative pos- FCP physicians" found that younger mendation that physicians should be- sibility of potential opportunities to ad- physicians were more likely than were come more vigilant about counseling vise against smoking?The explanation older physicians to determine patients' younger patients, performing primary could be simply that heavier smokers smoking status, to counsel patients to and secondary prevention rather than provided more cues: they had more quit,and to be aggressive in their coun- waiting for disease to become manifest. smoking-related disease, their clothes seling,an additional corroboration show- In our population,as well as in the gen- had a stronger odor of cigarettes, or ing that smoking cessation advice has eral population, it is in the adolescent they were more likely to have been seen increased over time. It is worth reas- years that cigarette smoking and addic- smoking a cigarette before they came serting, however, that the improve- tion begin,15 and younger patients are into the office. ments in smoking cessation advice rates the ones for whom counseling has the Given the finding that the most ad- seen in this study in the early 1980s greatest potential benefit.It is also well dieted smokers received the most coun- were not sustained through the remain- documented that stopping smoking at seling advice,it was not surprising that der of the decade. any age can have positive health effects.16 those respondents who were likely to CONCLUSIONS Since even the long-term smoker can have been the most addicted of all smok- benefit from cessation,an advice rate of ers—those with a history of myocardial We conclude that there are three pos- 65%found in those patients aged 50 to infarction,stroke,or hypertension who sible explanations for respondents'stat- 74 years is still low. still persisted in smoking—were,in fact, ing that they had never been counseled The advice differences by gender and the most likely to have been counseled. to quit smoking: (1) they had, in fact, by ethnicity are also of interest. The Nonetheless,although smokers with car- never been counseled (either because improvement and maintenance of fe- diovascular diseases were more likely the physician did not ask about their males' smoking cessation advice rates to have received advice than were those smoking habits, they denied smoking, may be partly attributable to increased smokers who did not have cardiovascu- or the physician did not respond to the physician sensitivity to women's lung lar diseases, 15.6% of smokers with a available information); (2) they denied cancer rates,but it is difficult either to history of myocardial infarction or the counseling; or (3) they forgot the ignore or to explain the relative lack of stroke,34.7%of hypertensive smokers counseling. However, even presuming change in males'smoking cessation ad- who smoked, 21.7%of smokers receiv- that recall decreases over time,we still vice rates over time.The increased like- ing antihypertensive medications, and found that only 52.1% of smokers who JAMA,December 11, 1991—Vol 266,No.22 Smoking Cessation Advice—Frank et al 3143 had seen a physician in the last year arette-related illness and potential ill- sation counseling. Smokers in this stated that they had ever been advised ness. But secondary prevention mea- study had a mean of 4.4 office visits per, to quit smoking,only 60.5%of smokers sures such as cessation counseling may year; each visit presented an opportu- with seven or more office visits in the be the greatest help to those with the nity for physicians to tell patients that past year stated that they had ever re- smallest current burden of disease- they would like to help them quit. ceived advice,and only 3.6%of ex-smok- those youngest and least addicted. Because half of the smokers in this ers stated that their physicians had It is clear from the high correlation study stated that no physician had ever helped them quit. between reported smoking cessation told them to quit smoking,and because While our data show some improve- advice and the number of office visits of physicians' potential for influencing ment in physicians' counseling efforts per year that increased exposure to smokers'resolve and ability to quit,it is from 1979 to 1990, they also suggest a medical care increased smokers' incumbent upon physicians to increase recent decline in these improvements. chances of being advised to quit smok- routine cessation counseling for all pa- There is good reason to believe that such ing by a physician.Since frequency and tients who smoke. advice can help patients stop smoking,1,4 type of office visits are both malleable and there are resources available to help parameters,there are two possible ap- This work was supported by Public Health physicians provide this advice.","It is proaches to improving counseling rates. Service grant 1R01-HL-219M to John W. Far- possible that physicians are waiting for One is to create smoking cessation- quhar, MD, and an institutional National Re- search Service Award, 5T32-HL-07034 (E.F.) patient cues: for obvious or heavy cig- specific office visits with physicians or from the National Heart, Lung,and Blood Insti- arette use,cardiovascular disease,cere- other health care providers, an option tute, Bethesda, Md. brovascular disease, hypertension, or that could become easier if compensa- The authors wish to thank John W. Farquhar, frequent office visits.And it is certainly tion becomes more available for preven- MD,Darius Jatulis, MS, Randall F.White, MD, el- a ro riate that physicians counsel tive services. The second is to make Abby a King,PhD,Nathan Maccoby,PhD, for PP A P Y ena Kraemer,PhD,and the late Ulrich Frank for those with the greatest burden of cig- each office visit an opportunity for ces- their invaluable assistance on this manuscript. References 1. World Health Organization.Consultative Group Government Printing Office;1976. quhar JW. Attitudes and practices of physicians on Statistical Aspects of Tobacco-Related Mortal- 7. Anda RF, Remington PL, Sienko DG, Davis regarding hypertension and smoking:the Stanford ity.The future worldwide health effects of current RM.Are physicians advising smokers to quit:the Five-City Project.Prev Med. 1985;14:70-80. smoking patterns.Presented at the Seventh World patient's perspective.JAMA.1987;257:19161919. 15. US Dept of Health and Human Services.Smok- 'Conference on Tobacco and Health;April 3,1990; 8. Farquhar JW,Fortmann SP,Maccoby N,et al. ing, Tobacco, and Health. Washington, DC: US Perth,Western Australia. The Stanford Five-City Project:design and meth- Dept of Health and Human Services;1989:5-6.US 2. Cummings SR,Richard RJ,Duncan CL,et al. ods.Am J Epidemiol. 1985;122:323-334. Dept of Health and Human Services publication Training physicians about smoking cessation:a con- 9. Farquhar JW, Fortmann SP, Flora JA,et al. CDC 87-8397. trolled trial in private practices.J Gen Intern Med. Effects of communitywide education on cardiovas- 16. US Dept of Health and Human Services. The 1989;4:482489. cular disease risk factors:The Stanford Five-City Health Benefits of Smoking Cessation.Washing- 3. Glynn TJ,Manley MW,Pechacek TF.Physician- Project.JAMA. 1990;264:359-365. ton,DC:US Dept of Health and Human Services; initiated smoking cessation program:the National 10. Passel JS,Word DL.Constructing the list of 1990.US Dept of Health and Human Services pub- Cancer Institute trials.In:Engstrom PF,Rimer B, Spanish surnames for the 1980 census:an applica- lication CDC 90-8416. Mortenson LE,eds.Advances in Cancer Control. tion of Bayes'theorem. Presented at the annual 17. Cooper R, Simmons BE. Cigarette smoking New York,NY:John Wiley&Sons Inc;1990:11-26. meeting of the Population Association of America; and ill health among black Americans.N Y State J 4. Ockene JK,Msteller J,Goldberg R,et al.In- April 11, 1980;Denver,Colo. Med. 1985;85:344349. creasing the efficacy of physician-delivered smok- 11. Centers for Disease Control. Behavioral risk 18. US Dept of Health and Human Services.Clin- ing interventions:a randomized clinical trial.J Gen factor surveillance,1986.MMWR.1987;36:252-254. ical Opportunities for Smoking Intervention: A Intern Med. 1991;6:1-8. 12. Centers for Disease Control. Behavioral risk Guide for the Busy Physician. Washington, DC: 5. Strecher VJ,O'Malley MS,Villagra VG,et al. factor surveillance, 1988.MMWR. 1990;39:58-68. US Dept of Health and Human Services; 1986. Can residents be trained to counsel patients about 13. Fiore MC,Novotny TE,Pierce JP,et al.Meth- National Institutes of Health publication 86-2178. quitting smoking?J Gen Intern Med. 1991;6:9-17. ods used to quit smoking in the United States:do 19. The American Academy of Family Physicians. 6. US Dept of Health, Education, and Welfare. cessation programs help?JAMA.1990;263:2760-2765. The AAFP Stop Smoking Kit. Kansas City, Mo: Adult Use of Tobacco,1975.Washington,DC:US 14. Fortmann SP, Sallis JF, Magnus PM, Far- American Academy of Family Physicians;1987. 3144 JAMA,December 11,1991-Vol 266,No.22 Smoking Cessation Advice-Frank at al I Brand Logo Children Aged 3 to 6 Years Mickey Mouse and Old Joe the Camel Paul M. Fischer, MD; Meyer P. Schwartz, MD; John W. Richards, Jr, MD; Adam O. Goldstein, MD; Tina H. Rojas Objective.—Little is known about the influence of advertising on very young targeted to adolescents and that adver- children. We,therefore, measured product logo recognition by subjects aged 3 tising does not increase the use of to- to 6 years. bacco products. These companies claim Design.—Children were instructed to match logos with one of 12 products that advertising and promotion are de- pictured on a game board.Twenty-two logos were tested, including those rep- signed instead to produce brand itch- resenting children's products,adult products,and those for two popular cigarette ing adults who already smokea. In Camel and Marlboro). 1988,, the tobacco industry spent$3.27 brands ( ) billion on cigarette advertising and pro- Setting.—Preschools in Augusta and Atlanta, Ga. motions, making cigarettes the second Participants.—A convenience sample of 229 children attending preschool. most heavily advertised product in the Results.—The children demonstrated high rates of logo recognition. When United States." This level of advertis- analyzed by product category,the level of recognition of cigarette logos was in- ing cannot be justified on the basis of termediate between children's and adult products.The recognition rates of The brand switching alone,since only 100 of Disney Channel logo and Old Joe(the cartoon character promoting Camel cig- current smokers change brands within arettes) were highest in their respective product categories. Recognition rates agiven year. increased with age.Approximately 30%of 3-year-old children correctly matched Old Joe with a picture of a cigarette compared with 91.3%of 6-year-old children. See also pp 3149, 3154, and 3185. Conclusion.—Very young children see, understand, and remember adver- tising.Given the serious health consequences of smoking,the exposure of chil- There has been considerable specu- dren to environmental tobacco advertising may represent an important health lation but little published research on risk and should be studied further. the impact of advertising on very young (JAMA. 1991;266:3145-3148) children. Typical survey methods are generally unreliable when conducting re- search involving this age group since SEVERAL types of research have been cent study of teenagers, the self-per- children have limited verbal skills and, used in the past decade to examine the ceived influence of cigarette advertis- therefore, cannot articulate concepts association between cigarette advertis- ing had the strongest and most consis- that they may understand. There is, ing and rates of underage smoking. tent effect on the initiation of smoking however,someevidence that veryyoung These studies have included economet- among a group of variables that included children understand advertising.Dono- ric modeling,"the measurement of ad- parental smoking,sibling smoking,peer hue et al 14 used nonverbal measures(ie, vertisement recognition," and experi- influence, and intention to smoke.' picture games)to study young children's mental studies of adolescents'reactions Collectively, these studies provide knowledge of television commercials. to advertisement imagery.", In one re- compelling evidence that cigarette ad- They concluded that by 3 years of age, vertisements are seen by adolescents children understand both the intent of and that they respond to the advertise- television commercials and the sophis- From the Department of Family Medicine,The Med- merits' intent. Some health experts, ticated concept of audience segmenta- ical College of Georgia. Augusta (Drs Fischer, therefore now believe that ci arettead- Schwartz. Richards,and Ms Rojas);and the Depart- g tion (le, that advertisements are tar - Schwartz of Family Medicine,University of North Carolina, vertlsing is causally linked to smoking geted to specific groups). Chapel Hill(Dr Goldstein). behavior. -" We studied the recognition level of 22 Reprint requests to Department of Family Medicine, In contrast, the tobacco industry ar- brand logos by children aged:3 to 6 years. The Medical College of Georgia,Augusta,GA 30912 (Dr Fischer). gues that cigarette advertising is not Included were logos from two of the JAMA,December 11, 1991—Vol 266, No.22 Brand Logo Recognition—Fischer et al 3145 most heavily advertised cigarette Table 1.-Logos Tested,*Correct Product Response,and Recognition Rates for 229 Subjects Aged 3 to 6 brands: Camel and Marlboro. Years METHODS Correct Product Category Logo Product Response Recognition Rate,i The study design was based on the Children's brands Disney Channel Mickey Mouse 917 well-accepted market research concept "McDonald's" Hamburger 81.7 of advertisement recognition." Reco - "Burger King" Hamburger 79.9 !� g "Domino's Pizza" Pizza 78.2 nition of an advertisement indicates that "coca cola" Glass of cola 7s.o it has been both seen and remembered. "Pepsi" GlassDole 56. "Nike" Athletice 56 shoe .88 Twenty-two brand logos were col- "Waft Disney" Mickey Mouse 48.9 lected from a variety of printed sources "Kellogg's" Bowl of cereal 38.0 including advertisements and product "Cheerios" Bowl of cereal 25.3 Cigarette brands Old Joe Cigarette 51.1 packaging. These included the logos of "Marlboro"and red roof Cigarette 32.8 10 products that are, in part, targeted Marlboro man Cigarette 27.9 os representing two Camel and pyramids Cigarette 271 to children,five to $� P g "Camel" Cigarette 18.0 cigarette brands, and seven logos of Adult brands "Chevrolet" Automobile 54.1 products primarily targeted to adults "Ford" Automobile 52.8 (Table). No logos had specific image or 'CBS"le Computer z3 1 word clues that might indicate what kind "NSC" Television 21.0 of product the brand represented(eg,the "Kodak" Camera 17.9 "Marlboro man"was not smoking).One of "IBM'• Computer 16.2 Surgeon the current Sur General's warnin Surgeon General's g gs warning Cigarette 10.0 ("Quitting Smoking Now Greatly Re- duces Serious Risks to Your Health")as *Quotation marks on the logo indicate that the brand name is part of the test item. it appears on printed advertisements was also included as a test item. Subscores were calculated for each prod- (43.7%) watched 2 to 4 hours, and 21 Recognition was measured by having uct type(ie,children's brands,cigarette (9.2%)watched 4 or more hours per day. the subjects match 22 logo cards to one brands, or adult brands), which were The parents were also asked to rate 'of 12 products pictured on a game board then transformed into a recognition rate how often their child requested specific (Table). Product position on the board score based on the percentage of correct product brands, measured with a four- was made by random assignment and matches for that type of product and for item Likert scale.Thirty-seven parents was not varied between subjects. the products overall. These were then (16.2%) reported almost always, 106 Subjects were recruited from 10 pre- compared with data from the parental (46.3%) reported often, 79 (34.5%) re- schools in Augusta and Atlanta,Ga.The questionnaire.Analyses using the X2 test ported infrequently, and seven (3.1%) schools were selected in an attempt to were used to test for independence be- reported never. Unexpectedly, X'anal- balance the sample for race and socio- tween categorical variables.The McNe- ysis revealed no association of more fre- economic variables.Each subject's par- mar Test was used to test for the sig- quent brand requests among older chil- ent signed a parental consent form and nificance of change in correct responses dren. Fifty-two percent of 3-year-old completed a short questionnaire about between two logos (The Disney Chan- children,73%of 4-year-old children,6317C the child's age,gender,race,number of nel and Old Joe).16 A correlation analy- of 5-year-old children,and 65%of 6-year- hours of television watched each day, sis was used to test for association of old children often or almost always re- frequency with which the child requested recognition rates by age. A multifactor quested specific brands (P=.10). specific brands, number of years of pa- analysis of variance was used to test the The mean logo recognition rates for rental education, and the use of ciga- significance of the survey variables and all subjects are shown in the Table. As rettes in the subject's home. the recognition of logos. would be expected, children had high On the following day, each child was RESULTS recognition of the children's brand logos individually tested in a quiet, separate ranging from 91.7% for The Disney area of his or her classroom. It was ex- Two hundred twenty-nine children Channel to 25.3%for Cheerios.Random plained to the child that he or she would were recruited. Subjects ranged in age guessing alone would produce a recog- play a game matching cards with prod- from 3 to 6 years. Seventy-nine were 3 nition rate of 8.3%(ie,one of 12 items). ucts. Each of the 12 products on the years of age(34.5%),67 were 4 years of Old Joe, the cartoon character pro- game board was then named. A dem- age (29.3%), 60 were 5 years of age moting Camel cigarettes,had the high- onstration of matching was done with a (26.2%), and 23 were 6 years of age est recognition rate among the tested sample logo card.The child was then given (10.0%). One hundred twenty-three cigarette logos. More than half of the a test card to match.Afterthe child placed (53.7%)were boys. One hundred sixty- subjects correctly matched this figure the card on the board(whether correct or six (72.5%) were white and 63 (27.5%) with a picture of a cigarette. The other incorrect), the child was told, "That's were black.Of the subjects'parents,67 cigarette logos were correctly recognized good."No other feedback or encourage- (29.3%) had less than 12 years of edu- at rates of 18.0%to 32.8%.The logos for ment was given. Following each match, cation, 123(53.7%)had 12 to 16 years of adult products were recognized by 16.2% that card was removed from the board education,and 39(17.0%)had more than to 54.1% of subjects, with automobile and the child was presented with the next 16 years of education. Many subjects brand logos having the highest recog- card.The cards were randomized for each (34.1%)came from homes where at least nition rates. subject to prevent bias due to the order one person smoked. Logo recognition was highly associ- of presentation. Each subject's parent was asked to ated with the subject's age. Figure 1 Responses were graded as either cor- report the number of hours of television shows mean recognition rates by age for rect(grade of 1) or incorrect (grade of watched per day by his or her child.One each of the three product categories. 0),and a score was derived by summing hundred eight (47.2%) of the subjects This association of increased_recogni- the binary values assigned to each logo. watched up to 2 hours per day, 100 tion with older age was significant for 3146 JAMA,December 11,1991-Vol 266,No.22 Brand Logo Recognition-Fischer et al was positively associated with the recognition of adult product logos too- (F(2,198)=6.14; P=.002), but not with 90 6y children's products or cigarettes. The 80 5y recognition of cigarette logos was in- dependent of the use of cigarettes in R= 70 4y 6y the subjects' homes. ¢ m 6y Only 23 (10.0%) of the subjects o ci 60 3 matched the Surgeon General's warn- 2� 50 y 5y 5y ing with the picture of a cigarette. The ID 40 recognition rate for the warning in sub- cc a 30 4y 4y jects aged 3, 4, and 5 years was only 8.4%, a rate nearly equal to that ex- 20 3y 3y pected for random guessing. 10 COMMENT 04- Children are referred to by market- Children's Brands Cigarette Brands' Adult Brands ing researchers as"consumers in train- ing."As stated by McNeal,""All of the skills,knowledge,and behavior patterns Fig 1.—Logo recognition rates by years of age for children's brands,cigarette brands,and adult brands. that together we call consumer behav- ior are purposely taught to our children right along with toilet training,toddling P=NS and talking." 100 P<.001 Research has identified three distinct 90 P«.001 P«.001 children's markets.First are the markets 80 directly under the control of children.Chil- dren aged 5 to 12 years spend$4.2 billion 70 of their own money each year."By age 6 6 60 years, half of all children regularly go o C) shopping by themselves.' I 50- The second market is for products in a) 40- which children influence household pur- ¢a 30 chasing decisions.Children influence the 20 spending of$131 billion each year, in- 10 eluding $82 billion for food and bever- ages, $17 billion for leisure activities 0 4 ts,and$13 billion for apparel 3 4 5 6 (Business Week.September 9,1991:94). Age,y The third market is for products that children will consume when they become adolescents and adults.Market research- ers believe that brand awareness cre- ated in childhood can be the basis for ® product preference later in life."It has been shown that children prefer the The Disney Channel Old Joe the Camel brands that they see advertised."This effect has been shown to even influence their preference of products that they Fig 2.—Logo recognition rates for The Disney Channel and Old Joe by subject age. are too young to use, such as lipstick and diet soft drinks."This potential in- children's products (r=.51; P<.0001), not significant in subjects aged 6 years. fluence has raised concern about the ex- cigarette brands(r=.52;P<.0001),and In that age group,both the silhouette of posure of children to cigarette adver- adult brands (r=.50; P<.0001). Mickey Mouse and the face of Old Joe tising. For instance, it has been sug- Cigarette logo recognition rates were nearly equally well recognized and gested that children receive positive ranged from 11.4% (Camel) to 30.4% correctly matched by almost all children. messages about smoking when they view (Old Joe) for 3-year-old subjects. This Neither race nor gender was asso- cigarette advertisements and that this rate increased to between 43.5%(Camel) ciated with the recognition scores of may influence later decisions t o smoke.-'0 and 91.3%(Old Joe)for children 6 years any of the three product categories. The children in this study demon- of age. A multifactor analysis of variance strated high recognition rates of brand Figure 2 compares the recognition that included each of the other survey logos for products that are targeted to rates for The Disney Channel and Old variables showed that the hours of both children and adults. It is no sur- Joe by subject age.These two were the television watched were positively as- prise that most children can properly most highly recognized logos in their sociated with the recognition of chil- match the McDonald's arches to a ham- respective product categories. While dren's product logos (F(2,198)=10.1; burger. It is also not surprising that The Disney Channel's logo recognition P<.0001) and adult logos there is high recognition of the Chev- was higher for subjects aged 3,4,and 5 (F(2,198)=4.41; P=.01), but not with rolet and Ford logos. Automobiles are years,this difference in recognition was cigarette logos. Parental education heavily advertised on television, and JAMA, December 11, 1991—Vol 266,No.22 Brand Logo Recognition—Fischer et al 3147 many children are exposed to these rent study. First, the subjects were a It is obviously impossible to predict brands through personal family use. In convenience sample and, therefore, did how the exposure of children to envi-. contrast, the high recognition rate of not include children cared.for at home. ronmental tobacco advertising might in- cigarette logos may be counterintuitive. While the sample does represent children fluence their later smoking behavior. After all,cigarette advertising no longer from families with a range of ethnic and While cigarette companies claim that appears on television and very young socioeconomic backgrounds,it is unknown they do not intend to market to chil- children cannot read. Yet by the age of how the sampling method might have in- dren, their intentions are irrelevant if 6 years,Old Joe is as well recognized as fluenced the study results.The study de- advertising affects what children know. Mickey Mouse. sign relied on the ability of children to R J Reynolds Tobacco Company is as Children's knowledge of cigarette match cards picturingbrand logos to prod- effective as The Disney Channel in reach- brand logos is most likely the result of ucts on a game board. This is a novel ing 6-year-old children. Given this fact their exposure to "environmental to- design that has face validity, but which and the known health consequences of bacco advertising."Camel and Marlboro would be difficult to validate in any other smoking, cigarette advertising may be brand advertising is ubiquitous,appear- way. Finally, there is some obvious ar- an important health risk for children. ing in movies,on billboards,promotional bitrariness to the selection of specific logos displays at youth-oriented events, on as test items.The choice ofchildren's prod- television during sporting events, and ucts was based on published data describ- This work was supported in part by a grant from on "line extenders," such as T-shirts, ing children's markets."Heavily adver- Doctors ought to Care,Houston,Tex,and grant posters,and caps.21,"In addition to this tised cigarette brands were tested.Adult PBR-55 from the American Cancer Society, At- lanta, t- aid advertising, Camel and Marlboro brands were chosen from among those �" ' 11 g $� har We are indebted to Alan BIum, MD, Martha brand logos appear on video arcade products that are primarily targeted to Anne Tudor, Don Shopland, Dean M. Krugman, games,children's toys,and candy prod- consumers who are older than the study PhD, Nan Richards,MPH,and Karen Shipp,for ucts (Washington Post. March 13, subjects. Of all the brands tested, only prepublication review of the manuscript,and Jen- nifer Dushku for review of the data analysis.We 1990;sectA:8). the manufacturers of cigarettes specifi- would also like to thank the administration at the There are several limitations to the cur- cally claim not to advertise to children. 10 preschools where this research was conducted. References 1. Levit EM,Coate D,Grossman M.The effects of 8. Armstrong BK,de Klerk NH,Shean RE,Dunn ulus measurement and assessment research:a re- government regulation on teenage smoking.J law DA, Dolin PJ. Influence of education and adver- view of advertising testing methods. Curr Issues Econ.1981;24:545-569. tising on the uptake of smoking by children.Med J Res Advertising. 1983:135-165. 2. Seldon BJ,Dordoodian K.A simultaneous model Aust. 1990;152:117-124. 16. Siegel S, Castellan, NJ. Nonparametric Sta- of cigarette advertising:effects on demand and in- 9. Warner KE.A ban on the promotion of tobacco tistics for the Behavioral Sciences. 2nd ed. New dustry response to public policy. Rev Econ Stat. products.N Engl J Med. 1987;316:745-747. York,NY:McGraw-Hill Book Co Inc;1988:75-87. 1989;71:673-677. 10. Davis RM.Current trends in cigarette adver- 17. McNeal JU.Children as Consumers.Lexing- 3. Goldstein A0,Fischer PM,Richards JW,Cre- tising and marketing.N Engl J Med.1987;316:725- ton,Mass:Lexington Books;1987:5, 12,47,179. ten D. Relationship between high school student 732. 18. Goldberg ME,Gorn GJ,Gibson W. TV mes- smoking and recognition ofcigarette advertisement. 11. Centers for Disease Control.Cigarette adver- sages for snack and breakfast foods:do they influ- J Pediatr. 1987;110:488491. tising-US, 1988.MMWR. 1990;39:261-265. ence children's preference?J Consumer Res.1978; 4. Chapman S,Fitzgerald B.Brand preference and 12. Surgeon General.Reducing the Health Conse- 5:73-81. advertising recall in adolescent smokers:some im- quence of Smoking:25 Years of Progress:A Report 19. Gorn GJ,Florsheim R.The effect of commer- plications for health promotion.Am JPublic Health. of the Surgeon General.Rockville,Maryland:Office cials for adult products on children. J Consumer 1982;72:491494. onSmokingand Health:1989.US DeptofHealthand Res. 1985;11:962-967. 5. Aitken PP,Leathar DS,O'Hagan FS,Squair SI. Human Services publication CDC 89-8411. 20. Charlton A.Children's advertisement aware- Children's awareness to cigarette advertisements and 13. Tye JB,Warner KE,Glantz SA.Tobacco ad- ness related to their views on smoking. Health. brand imagery.Br J Addict. 1987;82:615-622. vertising and consumption: evidence of a casual Educ J. 1986;45:75-78. 6. Aitken PP, Leathar DS,Scott AL,Squair SI. relationship. J Public Health Policy. 1987;8:492- 21. Richards JW,Fischer PM.Smokescreen:how Cigarette brand preference of teenagers and adults. 508. tobacco companies market to children.World Srnok- Health Promotion. 1988;2:219-226 14. Donohue TR, Henke LL, Donohue WA. Do ing Health. 1990;15:12-14. 7. Aitken PP, Eadie DR. Reinforcing effects of kids know what TV commercials intend?J Adver- 22. Blum A. The Marlboro Grand Prix: circum- cigarette advertising on underage smoking. Br J tising Res. 1980;20:51-57. vention of the television ban on tobacco advertis- Addict. 1990;85:399412. 15. Leckenby JD,Plummer JT.Advertising stim- ing.N Engl J Med. 1991;324:913-917. 3148 JAMA,December 11, 1991-Vol 266, No.22 Brand Logo Recognition-Fischer et al i RJR Nabisco's Cartoon Camel Promotes Camel Cigarettes to Children Joseph R. DiFranza, MD; John W. Richards, Jr, MD; Paul M. Paulman, MD; Nancy Wolf-Gillespie, MA; Christopher Fletcher, MD; Robert D. Jaffe, MD; David Murray, PhD Objectives.—To determine if RJR Nabisco's cartoon-theme advertising is have not bought is Camel. In seven sur- more effective in promoting Camel cigarettes to children or to adults.To deter- veys,involving 3400 smokers in the sev- mine if children see, remember, and are influenced by cigarette advertising. enththrough l2thgrades,conducted be- Design.—Use of four standard marketing measures to compare the effects tween 1976 and 1988 in Georgia, Loui- of Camel's Old Joe cartoon advertising on children and adults. siana,and Minnesota,Camel was given Subjects.—High school students, grades 9 through 12,from five regions of as the preferred brand by less than 0.5% the United States, and adults, aged 21 years and over, from Massachusetts. (SaundraMacD.Hunter,PhD,Weihang Bao,PhD, Larry S.Webber,PhD,and Outcome Measures.—Recognition of Camel's Old Joe cartoon character, Gerald S. Berenson, MD, unpublished product and brand name recall,brand preference,appeal of advertising themes. data, 1991; D.M., unpublished data, Results.—Children were more likely to report prior exposure to the Old Joe 1991).14.18.19 In 1986, Camels were most cartoon character (97.7% vs 72.2%; P<.0001). Children were better able to popular with smokers over the age of 65 identify the type of product being advertised (97.5%vs 67:0%; P<.0001) and years,of whom 4.4%chose Camels,and the Camel cigarette brand name (93.6% vs 57.7°/x; P<.0001). Children also least popular among those 17 to 24 years found the Camel cigarette advertisements more appealing(P<.0001).Camel's of age, of whom only 2.7% preferred share of the illegal children's cigarette market segment has increased from 0.5% Camels.20 to 32.8%, representing sales estimated at $476 million per year. it In 1988,RJR Nabisco launched the Conclusion.—Old Joe Camel cartoon advertisements are far more success- smooth character advertising cam- paign,ful at marketingCamel cigarettes to children than to adults.This finding Is con- featuring Old Joe,acartoon camel g g modeled after James Bond and Don sistent with tobacco industry documents that indicate that a major function of to- Johnson of"Miami Vice."21 Many indus- bacco advertising is to promote and maintain tobacco addiction among children. try analysts believe that the goal of this (JAMA 1991;266:3149-3153) campaign is to reposition Camel to com- pete with Philip Morris'Marlboro brand WITH the number of US smokers de- alty among adult smokers.5~8 However, for the illegal children's market segment. clining by about 1 million each year,the industry advertising expenditures can- To determine the relative impact of Cam- tobacco industry's viability is critically not be economically justified on this ba- els Old Joe cartoon advertising on chil- dependent on its ability to recruit re- sis alone.'This study was therefore un- dren and adults,we used four standard placement smokers.'Since children and dertaken to determine the relative im- marketing measures. teenagers constitute 90% of all new pact of tobacco advertising on children 1. Recognition.We compared the pro- smokers their importance to the indus- and adults. portions of teenagers and adults aged 21 try is obvious! Many experts are con- There is abundant evidence that to- years and over who recognize Camel's vinced that the industry is actively pro- bacco.advertising influences children's Old Joe cartoon character. moting nicotine addiction among youth.',' images of smoking."In Britain,the pro- 2. Recall.We compared the ability of portion of children who gave gooks teenagers and adults to recall from a tough"as a reason for smoking declined masked Old Joe advertisement the type See also pp 3145, 3154,and 3185. after tough images were banned from of product being advertised and the cigarette advertisements."Children as brand name. young as the age of 6 years can reliably 3. Appeal. We compared how inter- Spokespersons for the tobacco indus- recall tobacco advertisements12 and esting and appealing a series of Old Joe try assert that they do not advertise to match personality sketches with the cartoon character advertisements were people under 21 years of age, the sole brands using that imagery.10 In fact, to teenagers and adults. purpose of their advertising being to cigarette advertising establishes such 4. Brand preference. We compared promote brand switching and brand loy- imagery among children who are cog- brand preferences of teenaged smokers nitively too immature to understand the prior to the Old Joe cartoon character From the of Massachusetts rtment Med cal School,Family rFitchburg n(or DI purpose of advertising."Subsequently, campaign with those 3 years into the Franza);Department of Family Medicine,Medical Col- children who are most attuned to ciga- campaign to determine if the campaign lege of Georgia,Augusta(Dr Richards);Department of rette advertising have the most positive had been more effective with children Family Practice,University of Nebraska Medical Cen- or with adults,and to determine if Camel ter,Omaha(Dr Paulman and Ms Wolf-Gillespie);De- attitudes toward smoking, whether or partment of Family,Community and Emergency Med- not they already smoke."Children who had been repositioned as a children's icine. University of New Mexico School of Medicine, are more aware of, or who approve of, brand. Albuquerque(Dr Fletcher);Department of Family Prac- tice, cigarette advertisements are more like, , - METHODS l University of Washington,Seattle(Dr Jaffe);and g 10111316 y $Ub'eCtS Department of Epidemiology,University of Minnesota, to smoke, and those who do smoke l Minneapolis(Dr Murray). buy the most heavily advertised Since adolescent brand preferences Reprint requests to University of Massachusetts brands.`-` may v from onegeographic location Medical School, Department of Family Practice, 47 y Ashby State Rd,Fitchburg,MA 01420(Dr DiFranza). Historically, one brand that children to another (Saundra MacD. Hunter, JAMA,December 11, 1991—Vol 266,No.22 Promotion of Camel Cigarettes to Children—DiFranza et al 3149 Fig 1.—Masked Old Joe Camel cartoon advertisement. V Fig 2.—A portion of an eight-page Camel advertising supplement. Midnight at the Nib When Joe throws a party at his favorite watering hole, only smooth characters need apply. §.. PhD,Weihang Bao,PhD,Larry S.Web- considered to be smokers if they had Old Joe cartoon character campaign ber,PhD,and Gerald S.Berenson,MD, smoked one or more cigarettes during early in 1988(Saundra MacD. Hunter, unpublished data, 1991; D.M., unpub- the previous week.Previously validated PhD,Weihang Bao,PhD,Larry S.Web- lished data,1991),11,11,11 we selected chil- questions were used to determine chil- ber,PhD,and Gerald S.Berenson,MD, dren from Georgia,Massachusetts,Ne- dren's intentions regarding smoking in unpublished data, 1991; D.M., unpub- braska, New Mexico, and Washington, the next month and year22 and their at- lished data, 1991).19.18,19 representing five regions.One school in titudes toward the advertised social ben- Tests of significance were made using each state was selected based on its ad- efits of smoking.21,1l the Two-tailed Student's t Test for con- ministration's willingness to participate. Subjects rated the ads as"cool or stu- tinuous data and the X2 and Fisher's Schools with a smoking prevention pro- pid" and `interesting or boring." Sub- Exact Test for discrete data.A P value gram focused on tobacco advertising jects were asked if they thought Old Joe of less than .05 was used to define sta- were excluded. was"cool" and if they would like to be tistical significance. A target of 60 students in each grade, friends with him.Each positive response The study was conducted during the 9 through 12,from each school was set. to these four questions was scored as a 1990-1991 school year. In large schools, classes were selected one,a negative response as a zero.The RESULTS to obtain a sample representative of all "appeal score"was the arithmetic sum levels of academic ability.Students were. of the responses to these four questions, A total of 1060 students and 491 sub- told that the study concerned advertis- with the lowest possible score per re- jects from the Registry of Motor Vehi- ing and were invited to participate anon- spondent being a zero and the highest a cles were asked to participate. Usable ymously. four. surveys were obtained from 1055 stu- Since adult brand preferences are dents(99%)and 415 license renewal ap- available from national surveys, adult Procedure plicants(84.5%). Seventy drivers were subjects were recruited only at the Mas- Subjects were first shown the masked under 21 years of age,leaving 345 adults sachusetts site. All drivers, regardless ad and asked if they had seen the Old aged 21 years or older.Students ranged of age,who were renewing their licenses Joe character before. They were then in age from 12 to 19 years(mean, 15.99 at the Registry of Motor Vehicles on the asked to identify the product being ad- years) and adults from 21 to 87 years days of the study during the 1990-1991 vertised and the brand name of the prod- (mean, 40.47 years). Females repre- school year were asked to participate. uct.Subjects who could not answer these sented 51.0%of the students and 54.8% Since licenses must be renewed in per- questions were required to respond of the adults. son,this is a heterogeneous population. "Don't know"so they would not be able Children were much more likely than to write in the correct answer when the adults to recognize Camel's Old Joe car- Materials unmasked advertisements were shown. toon character (97.7% vs 72.2%; Seven Camel Old Joe cartoon char- The subjects were then shown,one at a P<.0001)(Table).It is not plausible that atter advertisements were obtained time,the six unmasked advertisements the children were simply saying they from popular magazines during the 3 and asked to rate how the advertise- had seen Old Joe when they had not, years prior to the study. One ad was ments and the Old Joe cartoon charac- since they also demonstrated a greater masked to hide all clues(except Old Joe) ter appealed to them.Subjects then com- familiarity with the advertisement on as to the product and brand being ad- pleted the remainder of the survey in- the two objective measures. vertised (Fig 1). strument. When shown the masked advertise- The survey instrument collected de- Adolescent brand preference data ment,the children were much more suc- mographic information and information from this study were compared with the cessful than the adults in identifying the on past and present use of tobacco, in- data obtained by seven surveys com- product being advertised (97.5% vs cluding brand preference.Children were pleted prior to the kickoff of Camel's 67.0%;P<.0001)and the Camel brand 3150 JAMA, December 11, 1991—Vol 266,No.22 Promotion of Camel Cigarettes to Children—DiFranza et al Comparison of Student and Adult Responses to Camel's Old Joe Cartoon Character Advertisements Georgia Massachusetts Nebraska New Mexico Washington Total Total Students Students Students Students Students Students* Adultst No.of subjectst 212 224 232 210 177 1055 345 Have seen Old Joe,% 98.1 99.6 96.6 95.7 98.9 97.7§ 72.2§ Know product,% 95.3 100 97.8 95.7 98.9 97.5§ 67.0§ Know brand,% 92.9 97.3 91.8 90.0 96.6 93.6§ 57.7§ Think ads look cool,% 62.4 54.1 57.4 61.2 55.1 58.0§ 39.9§ Ads are interesting,% 83.4 73.9 77.3 62.3 69.9 73.6§ 55.1§ Uke Joe as friend,% 46.2 31.1 33.9 31.4 32.6 35.0§ 14.4§ Think Joe is cool,% 51.0 38.6 44.1 40.9 40.0 43.0§ 25.7§ Mean appeal scorell 2.4 2.0 2.1 1.9 2.0 2.1§ 1.4§ Smoke Camel(%)11 29!76 12/55 13/52 23/43 9/35 86/261 8/92 (38.2) (21.8) (25.0) (53.5) (25.7) (33.0)§ (8.7)§ *Age range,12 to 19 years. tAge range,21 to 87 years. *This is the total number of subjects in each category;due to incomplete questionnaires,respondents for some questions may be fewer. §P<.0001. IlSee text for explanation. $Percentage of smokers who identify Camel as their favorite brand. name(93.6%vs 57.7%;P<.0001).Even may be more familiar with the Old Joe tion of smokers under 18 years of age when the analysis was limited to those Camel campaign than adults in general. who choose Camels has risen from 0.5% subjects who were familiar with the Old Camel cigarettes are now most popular to 32.8%. Given that children under 18 Joe cartoon character,children were still with children and progressively less pop- years account for 3.3% of all cigarette more likely than adults to remember ular with older smokers. sales,f and given a national market share the product(98.6%vs 89.6%;P<.0001) About equal proportions of adults of 4.4% for Camel,' we compute that and the Camel brand name (95.0% vs (28.2%) and children (29.0%) reported Camel's adult market share is actually 79.1%;P<.0001).This confirms that Old some current cigarette use, making it 3.4%. Given a current average price of Joe cartoon advertisements are more unlikely that this factor influenced any 153.3 cents per pack,'the illegal sale of effective at communicating product and of the above findings. Although there Camel cigarettes to children under 18 brand name information to children than were some statistically significant dif- years of age is estimated to have risen to adults. ferences in the responses of children from $6 million per year prior to the Because Massachusetts adults may from different regions, these were not cartoon advertisements to$476 million not be representative of adults in the the focus of this study(Table). per year now, accounting for one guar- other four states where children were When compared with nonsmokers, ter of all Camel sales. surveyed, the above analyses were re- children who were currently smoking From both a legal and moral perspec- peated comparing only Massachusetts gave higher approval ratings to the ad- tive,it is important to determine if the children and adults. In all cases the dif- vertisements (mean approval score of tobacco industry is actively promoting ferences between adults and children 2.8 for smokers vs 1.8 for nonsmokers; nicotine addiction among youngsters. were significant and of even greater P<.0001). Approving attitudes toward However,from a public health perspec- magnitude(P<.0001),excluding the pos- cigarette advertisements seem to pre- tive it is irrelevant whether the effects sibility that the above findings were due cede actual smoking. Among the non- of tobacco advertising on children are to a lighter level of advertising expo- smoking children,those who either were intentional. If tobacco advertising is a sure in the Massachusetts area. ambivalent about their future smoking proximate cause of disease, it must be On all four measures, the children intentions or expressed a definite in- addressed accordingly. In the following found the Camel cartoon advertisements tention to smoke were more approving discussion we will examine the evidence more appealing than did the adults.Chil- of the advertisements than those chil- produced by this study, the marketing dren were more likely to think the ad- dren who intended not to smoke(mean practices of the tobacco industry as a vertisements looked "cool" (58.0% vs approval scores of 2.6 and 1.8, respec- whole as revealed in industry docu- 39.9%'0;P<.0001)or"interesting"(73.6% tively; P<.001). ments,and the marketing practices used vs 55.1%; P<.0001). More of the chil- Children were more likely to smoke if by RJR Nabisco, in particular, to pro- dren thought Old Joe was"cool"(43.0% they believed that smoking is pleasur- mote Camel cigarettes. The quotations vs 25.7%; P<.0001) and wanted to be able(relative risk(RRI, 6.6;P<.0001) cited below are from tobacco industry friends with him (35.0% vs 14.4%; and that it makes a person more popular personnel and from documents obtained P<.0001). (RR,2.0;P<.0001),and attractive(RR, during litigation over Canada's ban of The brand preference data revealed a 2.5; P<.0001), all common themes in tobacco advertising. dramatic reversal in the market seg- cigarette advertising.Among nonsmok- Our data show that children are much ment pattern that existed prior to Cam- ing children, those who believed that more familiar with Camel's Old Joe car- el's Old Joe cartoon character campaign. smoking would make them more attrac- toon character than are adults.This may Camel was given as the preferred brand tive were eight times more likely to ex- be because children have more expo- by 32.8%of children up to the age of 18 press an intention to smoke in the next sure to these advertisements,or because years who smoked,23.1%of Massachu- year(P<.001). the advertisements are inherently more setts adult smokers aged 19 and 20 years, COMMENT appealing to youngsters. The tobacco and 8.7% of those 21 years of age and industry has long followed a policy of over.The figures for the Massachusetts Our data demonstrate that in just 3 preferentially placing selected advertise- adults were significantly higher than the years Camel's Old Joe cartoon charac- ments where children are most likely to national market share for Camel,4.4%,' ter had an astounding influence on chil- see them.',',' For example, print ad- suggesting that Massachusetts adults dren's smoking behavior. The propor- vertisements are placed in magazines JAMA,December 11, 1991-Vol 266,No.22 Promotion of Camel Cigarettes to Children-DiFranza et al 3151 "specifically designed to reach young because they believe the benefits out- imagery of Export `A' against young people.""Paid cigarette brand promo- weigh the risks. To the insecure child, starter smokers."49 The average age for tions appear in dozens of teen movies.30 the benefits are the"psychological ben- starter smokers is 13 years.50 Camels are featured in the Walt Disney efits" promised in tobacco advertise- The industry also researches the best movies Who Framed Roger Rabbit?and ments:confidence, an improved image, ways of keeping children from quitting Honey I Shrunk the Kids. and popularity.2•4," Children who be- once they are "hooked on smoking.1136 The industry targets poster adver- lieve that smoking will make them more The purpose of one tobacco industry tisements for"key youth locations/meet- popular or more attractive are up to 4.7 study was to assess the feasibility of ing places in the proximity of theaters, times more likely to smoke.23,1A marketing low-tar brands to teens as an records[sic]stores,video arcades,etc."29 Previous research makes it clear that alternative to quitting.m The study found It is common to see Old Joe poster ad- children derive some of their positive that for boys, "[t]he single most com- vertisements in malls,an obvious gath- images of smoking from advertis- monly voiced reason for quitting among ering spot for young teens. Billboards, ing.11.13.34 Children who are aware of to- those who had done so. . . was sports."M T-shirts, baseball caps, posters, candy bacco advertising, and those who ap- The tobacco industry's sponsorship of cigarettes, and the sponsorship of tele- prove of it, are also more likely to be sporting events, such as the Camel Su- vised sporting events and entertainment smokers.10,11,13-16 Children's favorable at- percross motorcycle race,should be seen events such as the Camel"Mud and Mon- titudes toward smoking and advertising in relation to its need to discourage teen- ster"series are all used to promote Cam- precede actual tobacco use and corre- age boys from quitting. Similarly, its els. All are effective marketing tech- late with the child's intention to smoke, emphasis on slimness serves as a con- niques for reaching children.1,21,11-'4 suggesting that the images children de- stant reinforcement of teenage girls' The fact that children are much more rive from advertising encourage them fears of gaining weight as a result of attracted to the themes used in the Old to smoke.42 Our data confirm these ear- quitting. Joe cartoon character advertisements lier findings. Among nonsmoking chil- Our study provides further evidences' may also explain why they are more dren, those who were more approving that tobacco advertising promotes and familiar with them.The themes used in of the Old Joe advertisements were more maintains nicotine addiction among chil- tobacco advertising that is targeted at likely either to be ambivalent about their dren and adolescents. A total ban of children are the result of extensive re- smoking intentions or to express a def- tobacco advertising and promotions,as search on children conducted by the to- inite intention to smoke. Nonsmoking part of an effort to protect children from bacco industry to `learn everything children who believed that smoking the dangers of tobacco,-2,m can be based there was to learn about how smoking would make them more popular were on sound scientific reasoning. begins."l-'11 Their research identifies the eight times more likely to express an This project was supported by grants from the major psychological vulnerabilities of intention to smoke in the future. University of Massachusetts Medical Center,the children,which can then be exploited by Since a child's intention to smoke is Massachusetts chapter of the American Cancer advertising to foster and maintain nic- considered to be a good predictor of fu- Society,and Doctors Ought to Care. We would like to thank the participating schools otine addiction. ture smoking behavior,'it seems rea- and the following for their contributions to this The marketing plan for "Export A" sonable to conclude that a belief in the study:the Massachusetts Registry of Motor Vehi- cigarettes describes their "psychologi- psychological benefits of smoking, de- cles, Bruce Churchill, MD, Della C. de Baca, cal benefits": "Export smokers will be rived from advertising, precedes, and Sharon DiFranza, Saundra MacD. Hunter, Ca- perceived as . . . characterized b their contributes to the adoption of smoking. Melinda CounRabocil MD,Mary andSheHeal and the Ca- pe Y � p g• nadian Council on Smoking and Health. self-confidence, strength of character There are other lines of evidence in- References and individuality which makes them pop- dicating that tobacco advertising in1.- ular and admired by their peers."39 creases the number of children who use in Pierce JP,Fiore MC,Novotny St et al.Trends Consider a child's vulnerabilityto r tobacco.In countries where advertising ti cigarette smoking 0.the United States61-65. c- p� g tions to the year 2000.JAMA. 1989;261:61-65. pressure. According to one industry has been totally banned or severely re- 2. Kandel DB, Logan JA. Patterns of drug use study, "The goading and taunting that stricted, the percentage of young peo- from adolescence to young adulthood,I:periods of exists at the age of 11 or 12 to get non- ple who smoke has decreased more rap- risk for initiation,continued use,and discontinua- smokers to start smokingis virtually idly than in countries where tobacco pro- tion. e J . Public old Health.gets1984;74:660-666.owith Y Y P 3. Tye JB.RI Reynolds targets teens with sophis- gone from the peer group circles by 16 motion has been less restricted.""After ticated marketing campaign. Tobacco Youth Rep. or 17."31,36 If peer influence is virtually a 24-year decline in smokeless tobacco 1987;2(1):1-16. gone by the age of 16 years, who is the sales,an aggressive youth-oriented mar- 4. Borsch B. How Madison Avenue seduces chil- dren. r Management.March intended target group for RJR-MacDon- ketin campaign has been followed by 5. Voluntary Initiatives of¢Responsible Ind us- ald's Tempo brand, described as indi- what has been termed"an epidemic"of try.Washington,DC:The Tobacco Institute;1983. viduals who are"[e]xtremely influenced smokeless tobacco use among children, 6. Cigarette Advertising Code. Washington, DC: by their peer group"?'(RJR-MacDon- with the average age for new users be- The Tobacco Institute;1964:1-8. 7. Code of Cigarette Sampling Practices. Wash- ald is a wholely owned subsidiary of RJR ing 10 years.45,46 ington,DC:The Tobacco Institute;1971:1-4. Nabisco.) The recommended strategy Many of the tobacco industry docu- 8. RJ Reynolds Tobacco Company advertisement. for promoting this brand is the"[m]ajor ments cited above provide abundant ev- Time.April 9, 1984:91. usage of imagery which portrays the idence that one purpose of tobacco ad- 9. Tye JB,Warner KE, Glantz SA. Tobacco ad- vertising social appeal of peer group ac- vertisin is to addict children to tobacco. relations and consumption: evidence of a causal p pp Pe � P g relationship. J Public Health Policy. 1987;8:492- ceptance."40 In one Camel advertise- In the words of one advertising consult- 508. ment,a cowboy(a Marlboro smoker?)is ant, "Where I worked we were trying 10. Aitken PP,Leathar DS,O'Hagan FJ,Squair being denied admission to a party be- very hard to influence kids who were 14 SI. Children's awareness of cigarette advertise- cause"only smooth characters[ie,Camel to start to smoke.1191 Two marketing 622. mentsand brand imagery.BrJAddict.1987;82:615 smokers]need apply"(Fig 2).It appears strategy documents for Export A also 11. Charlton A.Children's advertisement-aware- that Camel advertisements are also tar- reveal that it is the youngest children ness related to their views on smoking. Health geted at individuals who are influenced they are after.48,4'"Whose behavior are Educ J. 1986;45(2)75-78. ,,� 12. Aitken PP, Leathar DS, O'Hagan FJ. Chil- by their peer group. we trying to affect?:new users. The dren's perceptions of advertisements for cigarettes. Children use tobacco, quite simply, goal is "[o]ptimizing product and user Soc Sci Med. 1985;21:785-797. 3152 JAMA,December 11', 1991-Vol 266,No.22 Promotion of Camel Cigarettes to Children-DiFranza et al IN. Aitken PP,Leathar DS,Squair SI.Children's 26. DiFranza JR, Tye JB. Who profits from to- bacco v The Attorney General of Canada). awareness of cigarette brand sponsorship of sports bacco sales to children?JAMA.1990;263:2784-2787. 41. Chassin L,Presson CC,Sherman SJ,Olshaysky and games in the UK.Health Educ Res.1986;1:203- 27. The Tax Burden of Tobacco.Washington,DC: RW.Self-images and cigarette smoking in adoles- 211. The Tobacco Institute;1990;25:109. cence. Personality Soc Psychol Bull. 1981;7:670- 14. Goldstein AO,Fischer PM,Richards JW,Cre- 28. Davis RM.Current trends in cigarette adver- 676. tin BA.Relationship between high school student tising and marketing.N Engl J Med.1987;316:725- 42. Aitken PP,Leathar DS,Squair SI.Children's smoking and recognition of cigarette advertise- 732. opinions on whether or not cigarette advertise- ments.J Pediatr. 1987;110:488-491. 29. RJR-MacDonald Document 286,p 1980-1(RJR- ments should be banned.Health Educ J. 1986. 15. Chapman S, Fitzgerald B. Brand preference MacDonald Inc&Imperial Tobacco v The Attor- 43. Teenage Smoking,Immediate and Long-term and advertising recall in adolescent smokers:some ney General of Canada). Patterns. Washington, DC: National Institute of implications for health promotion. Am J Public 30. Tye JB.Cigarette ads in kid's movies.Tobacco Education; 1979. US Dept of Health, Education, Health. 1982;72:491-494. Youth Rep. 1989;4(1):1-2. and Welfare,contract 400.79-0010. .16. Alexander HM,Callcott R,Dobson AJ,et al. 31. Blum A.Medicine vs Madison Avenue:fighting 44. Toxic Substances Board. Health or Tobacco: Cigarette smoking and drug use in schoolchildren, smoke with smoke.JAMA. 1980;244:739-740. An End to Tobacco Advertising and Promotion. IV:factors associated with changes in smoking be- 32. Klein JD,Forehand B,OliveriJ,Patterson CI, Wellington,New Zealand:Dept of Health;1989. havior.Int J Epidemiol. 1983;12:59-66. Kupersmidt JB,Strecher V.Candy cigarettes:do 45. US Dept of Health and Human Services,Pub- 17. McNeill AD,Jarvis MJ,West RJ.Brand pref- they encourage children's smoking?Pediatrics.In lic Health Service. The Health Consequences of erences among schoolchildren who smoke.Lancet. press. Using Smokeless Tobacco:A Report of the Advi- 1985;2:271-272. 33. Blum A. The Marlboro Grand Prix: circum- sory Committee to the Surgeon General.Bethesda, 18. Hunter SM,Webber LS,Berenson GS.Ciga- vention of the television ban on tobacco advertis- Md:National Institutes of Health;1986. NIH 86- rette smoking and tobacco usage behavior in chil- ing.N Engl J Med. 1991;324:913-917. 2874. dren and adolescents:Bogalusa Heart Study.Prev 34. Ledwith F. Does tobacco sports sponsorship 46. Connolly GN,Winn DM,Hecht SS,Henning- Med. 1980;9:701-712. on television act as advertising to children?Health field JE,Walker B,Hoffman D.The reemergence 19. Hunter SM,Croft JB,Burke GL,Parker FC, Educ J. 1984;43:85-88. ofsmokeless tobacco.NEngl J Med.1986;314:1020- Webber LS,Berenson GS. Longitudinal patterns 35. Report for Imperial Tobacco Ltd. Montreal, 1027. of cigarette smoking and smokeless tobacco use in Quebec:Kwechansky Marketing Research Inc;1977. 47. Report to Congress Pursuant to the Cigarette ,youth: the Bogalusa Heart Study. Am J Public Project 16(RJR-MacDonald Inc&Imperial To- Smoking Act 1978:An Action Oriented Research Health. 1986;76:193-195. bacco v The Attorney General of Canada). Program for Discovering and Creating the Best 20. Centers for Disease Control. Cigarette brand 36. Report for Imperial Tobacco Ltd.Kweehansky Possible Image for Viceroy Cigarettes, Ted Bates use among adult smokers-United States, 1986. Marketing Research Inc; 1982. Project PLUS/ Advertising,1975.Washington,DC:Federal Trade MMWR. 1990;39:665-673. MINUS(RJR-MacDonald Inc&Imperial Tobacco Commission;1975.Document AD11345. 21. Bird L.Joe Smooth for President. Adweek's v The Attorney General of Canada). 48. RJR-MacDonald.Export A Brand,Long Term Marketing Week.May 20, 1991:20-22. 37. Youth 1987.The Creative Research Group Ltd; Strategy(RJR-MacDonald Inc&Imperial Tobacco 22. Barton J,Chassin L,Presson CC.Social image 1987.Prepared for RJR-MacDonald(RJR-MacDon- v The Attorney General of Canada). factors as motivators of smoking initiation in early ald Inc&Imperial Tobacco v The Attorney Gen- 49. RJR-MacDonald.Export Family Strategy Doc- and middle adolescence. Child Dev. 1982;53:1499- era of Canada). ument(RJR-MacDonald Inc&Imperial Tobacco 1511. 38. Report to RJR-MacDonald Inc:Third Family v The Attorney General of Canada). 23. McAlpine KE,DiFranza J.Social attitudes of Qualitative Concept Test. Toronto, Ontario: Co- 50. Johnston LD,O'Malley PM,Bachman JG.Use children to cigarette smoking.Am Assoc Adv Sci. gemic Marketing; 1981 (RJR-MacDonald Inc & of Licit and Illicit Drugs by America's High School 1988;88:215.Abstract 725. Imperial Tobacco v The Attorney General of Can- Students:1975-198 . Rockville, Md: National In- 24. DiFranza JR,Murphy PJ, Ellefsen K. Coun- ada). stitute on Drug Abuse;1985. teradvertising:an image oriented school smoking 39. RJR-MacDonald.Export Family Strategy Doc- 51. Mintz M. Marketing tobacco to children. The prevention program.Presented at the annual meet- ument,1982(RJR-MacDonald Inc&Imperial To- Nation.May 6, 1991:cover,591,592,594,596. ing of the Society of Teachers of Family Medicine; bacco v The Attorney General of Canada). 52. Media advertising for tobacco products.JAMA. May 8, 1986;San Diego,Calif. 40. J.Walter Thompson.Report to RJR-MacDon- 1986;255:1033. 25. PM keeps cig lead.Advertising Age.December ald:Third Family Creative Direction Recommen- 53. Blasi V,Monaghan HP.The First Amendment 3, 1990:56. dation,1984.(RJR-MacDonald Inc&Imperial To- and cigarette advertising.JAMA.1986;256:502-509. JAMA, December 11, 1991-Vol 266, No.22 Promotion of Camel Cigarettes to Children-DiFranza et al 3153 I Does Tobacco Advertising Target Young People to Start Smoking? Evidence From California John P. Pierce, PhD; Elizabeth Gilpin, MS; David M. Burns, MD; Elizabeth Whalen, MA; Bradley Rosbrook,,MS; Donald Shopland; Michael Johnson, PhD Objective.—To evaluate whether tobacco advertising encourages teenag- vertising targets adolescents or en- ers younger than 18 years to start smoking. courages them to smoke or to use Design.—Comparison of 1990 California telephone survey data with data other tobacco products.' from a 1986 national telephone survey (both used a random-digit dialing sys- The effects of tobacco advertising have tem); 95% confidence intervals were calculated. To test our hypothesis, we become increasingly important and rel- considered whether the perception of advertising was related to age, whether evant issues, as can be seen by recent pub- the pattern of market share across age and sex groups followed the pattern of events t Canada.In that country,used lic health advocates successfully used perceived advertising,and whether changes in market share paralleled changes the argument that advertising promotes in advertising as perceived by the youngest age group. smoking by minors to convince the na- Participants.—There were 24296 adults and 5040 teenagers. tional legislature to ban all tobacco ad- Results.—The most advertised brands of cigarettes were Marlboro, accord- vertising. The tobacco industry ap- ing to 33.6%of adults and 41.8%of teenagers,and Camel,according to 13.7% pealed,and in the first round of appeal, of adults and 28.5% of teenagers—named most often by 12- to 13-year-olds the judge overturned the law(although (34.2%).The brands that were purchased most often were Marlboro and Camel. the law remains in effect pending the Together these were the brands of choice of 79.9%of males and 85%of females next round of appeal). This judicial de- aged 12 through 17 years. Marlboro's market share increased in youths and cision was based on acceptance of the tobacco industry's argument on the pur- young adults up to age 24 years and then decreased gradually with age;Cam- pose of its advertising." This decision el's market share decreased abruptly with age: it was the brand of choice of demonstrates the need to document fur- 24.5%-*5.8% of males aged 12 through 17 years but was chosen by only ther the evidence that tobacco adver- 12.7%±3.6%of males aged 18 through 24 years;for females, 21.7%±13.7% tising does target young people to aged 12 through 17 years chose Camels, while only 5.5%±3.2% aged 18 smoke. through 24 years preferred this brand. Both Marlboro and Camel brands had a higher market share in California in 1990 compared with that for the United See also pp 3145, 3149, and 3185. States in 1986.Of interest is that the market share for Camel increased among the younger smokers but was more evenly distributed for Marlboro. We hypothesized that credible evi- Conclusions.—Perception of advertising is higher among young smokers; dence against tobacco advertising would market-share patterns across age and sex groups follow the perceived adver- include (1) recent survey data that tising patterns; and changes in market share resulting from advertising occur showed that heavily advertised tobacco mainly in younger smokers. Cigarette advertising encourages youth to smoke brands were noticed more by minors and should be banned. than by older adults;(2)market share of (JAMA. 1991;266:3154-3158) cigarette brands in minors that reflected the pattern of recognition of advertis- ing; and (3) the market share of a cig- CIGARETTES are one of the most are attractive to children and inspire arette brand that only recently has tar- heavily marketed consumer products in them to experiment with smoking, geted young people has increased dis- the United States.''For nearly two de- which is highly addictive.'," The to- proportionately in the young.To address cades,public health advocates and rep- bacco industry has argued that adver- this third issue,it would be important to resentatives of the tobacco industry have tising is necessary for tobacco compa- choose a brand for which there has been hotly debated whether cigarette adver- nies to maintain market shares of this a major new advertising campaign in tising encourages children to start smok- legal product and has claimed that ad- recent years and compare its perfor- ing 1.1 The central concern of the public vertising by individual tobacco compa- mance with a brand that has been heavily health community is that the themes nies targets adults only and encour- advertised for years. and images used in tobacco advertising ages regular smokers to switch One brand that has undergone a ma- brands or to maintain brand jor new advertising campaign is Camel.13 From the Population Studies for Cancer Prevention, loyalty.&1' Despite evidence that 90% Camel cigarettes were first introduced University of California, San Diego (Drs Pierce and of recent adult smokers started smok- in 1913 by RJ Reynolds(RJR),and their Burns,Mss Gilpin and Whalen,and Mr Rosbrook);the in regularly before the age of 18 National Cancer Institute, Bethesda, Md (Mr Shop- land) hop- g z y g advertising approach remained up- land);and the Tobacco Control Section,California De- years, both tobacco product manu- changed until 1988,with such themes as partment of Health Services,Sacramento(DrJohnson). facturers and representatives of the "Camel. Where a man belongs." How- Reprint requests to Population Studies for Cancer Prevention, University of California, 2251 San Diego major association of advertisers have ever, r in February1988 RJR decided to Ave.suite B-111,San Diego,CA 92110(Dr Pierce). consistently denied that cigarette ad- update the Camel image and introduced 3154 JAMA,December 11, 1991—Vol 266,No.22 Tobacco Advertising and Young People—Pierce et al the "'smooth character" multimedia ad- other adults were interviewed.The sur- boards or in magazines. What brand of vertising campaign,which features a car- vey team interviewed all adolescents be- cigarette was advertised the most?" toon camel.It has been claimed that,since tween the ages of 12 and 17 years in the the introduction of this campaign in 1988, screened households.The response rates Calculation of "the character has reversed the fortunes were 75.3%and 78.4%for the adult and Confidence Intervals(Cls) of Camel cigarettes,now the No.6 brand teenage surveys, respectively. The For all percentages in the figures and in the $40 billion market" and that the aforementioned sampling probabilities tables, we have provided 95% CIs. In "smooth character is proving just how were used as part of the design effect to the CTS,these were derived by a vari- much a well-executed image blitz can do develop an initial weighting formula to ant of the jackknife procedure,in which for a decrepit brand.""The Tobacco Re- provide population estimates.Poststrat- 33 subsamples were taken from the full porter"also indicated how this campaign ification weighting ensured that the sam- survey file, and sample weights were halted slumping sales,"contributing to a ple was representative of the state pop- computed according to the same proce- hefty 14 percent change in the brand's ulation by age, sex, county-region, ed- dure as for the full sample.19 Variances market share during 1990." ucation, and race-ethnicity. were estimated on the basis of the de- Several public health professionals We compared data from the CTS with viations of the subsample percentages have claimed that this advertising cam- data collected in a similarly designed and the percentage for the full sample. paign targets children in particular. national survey, the Adult Use of To- The variances were then used to com- Even within the tobacco industry, one bacco Survey(AUTS)."The AUTS was pute the 95% CIs on the basis of the hears some suggestions that the new a national sample of 13 031 adults older critical value from Student's t distribu- Camel campaign does just that. Bird13 than 17 years. The response rates for tion. For the 1986 data, a design effect quotes a"source close to Marlboro"as the AUTS and for the in-depth surveys (1.5),which represents the cost in lack having said: "When you see teenage were 85.5% and 86.9%, respectively. of precision for choosing a sampling de- boys—people the cigarette companies Poststratification weighting was under- sign other than a simple random sam- aren't supposed to be targeting in the taken in this survey to ensure that the ple,published with the survey results," first place—going crazy for this guy[the population was representative of the was used to inflate the initial variance, cartoon camel],you know they're[RJR] adult population of the United States by which was computed as if the study sam- hitting their target."The RJR company age,sex,education,and race-ethnicity. ple were a simple random sample. has countered by claiming that"Camel smokers are principally white males over Definitions of Smoking Status RESULTS the age of 21 years."15 The definitions of current smoking Naming of the Most Camel is a brand that has undergone were those standards that are used in Advertised Cigarette Brand a change in its marketing strategy from all national surveys• in the United One third (33.6%) of all adults (,18 one clearly targeted at adult men to one States.2,18 Adults were asked whether years old) identified Marlboro as the that is now accused of aiming its heavy they had smoked 100 cigarettes in their most advertised brand;Camel was iden- advertising at minors. In this article, lifetime and whether they smoked now. tified by 13.7% of adults as the most we compare the performance of Camel Former smokers were those who had advertised brand. Among teenagers with that of Marlboro, which has had smoked 100 cigarettes but did not cur- (aged 12 through 17 years),41.8%iden- the largest advertising budget for many rently smoke, and those who smoked tified Marlboro and 28.5% identified years and has been the consistent mar- fewer than 100 cigarettes were consid- Camel as the most advertised brand. ket leader in recent times. We present ered to have never smoked. Teenagers No more than 3%of either the adult or evidence that Camel and Marlboro cig- were classified as current smokers if teenage respondents named any other arettes are perceived by minors as be- they reported smoking a cigarette within single brand. ing heavily advertised;that minors when the previous 30 days. We further clas- In each age group(except the young- compared with adults differentially sified teenagers who were not current est, in which the 95% CIs overlapped), smoke heavily advertised brands; and smokers as follows:respondents who had Marlboro was the brand most omen named that the market share of minors and smoked(but not within the 30-day time as the most advertised cigarette,and both young adults has increased dispropor- frame)were considered to have exper- Camel and Marlboro were named more tionately for Camel cigarettes compared imented with smoking or to be former often by adolescents than by adults(Fig with Marlboro over the past 5 years. smokers;those who had never smoked 1). However, the pattern of recognition METHODS but indicated that they might smoke in for Marlboro advertising across age the future were categorized as adoles- groups was somewhat different from the Data Sources cents contemplating smoking; and, fi- pattern for Camel advertising. The per- In 1990,the California Department of nally, teenagers who said they would centage who named Marlboro as being Health commissioned a series of Cali- never smoke were placed in a"never" the most advertised increased with age fornia Tobacco Surveys (CTS). Herein category. among adolescents, peaking at 48.1% we report data from in-depth telephone Survey Questions among the 16- through 17-year-old age interviews of 24 296 adults and 5040 teen- y group before declining among adults. agers. These interviews were obtained In both the adult and the teenage CTS, However,the youngest respondents(aged after a stratified Waksberg random-digit information on brand smoked was ob- 12 through 13 years)most often named dialing method16 was used to screen tained by asking current smokers,"What Camel as the most advertised brand,with 32<>135 households to recruit partic- brand do you usually buy?"In the 1986 34.2%of this age group identifying Camel ipants(J.Waksberg,unpublished data, AUTS, current smokers were asked, as the most advertised brand of cigarettes. December 17, 1984). "What brand of cigarette do you usually Moreover,the percentage of respondents smoke now?" In the 1990 CTS, all re- who named Camel steadily declined by Response Rates and Survey Methods spondents were asked the following age to 22.7%for 16-through 17-year-olds, All adult smokers and former smok- question about cigarette advertising: 19.8%for 18-through 24-year-olds, and ers who had quit within the last 5 years `Think back to the cigarette advertise- less than 100/c for all respondents aged 45 were interviewed;a random 28%b of all ments you have recently seen on bill- years and older. JAMA, December 11, 1991—Vol 266,No.22 Tobacco Advertising and Young People—Pierce et at 3165 Table 1.-Brands Purchased by Smokers in California 70 Brand Purchased,%* Semple c 60 Age,yt Marlboro Camel Other Size,n vi 50 Males . 12-17 55.4±12.6 24.518.0 18.0113.4 131 40 18-24 71.3±5.1 12.713.6 16.014.5 567 30 ,g 25-29 61.7±7.6 13.8±6.9 24.515.8 470 20 30-44 47.7t2.9 10.5±2.1 41.8±3.3 1579 W,& >_45 21.1±2.7 8.112.8. 70.8t3.3 1316 0 x Females 12-17 63.3t15.7 21.7±13.7 13.4±9.6 102 1$ZA25�N X65 18-24 69.4±7.1 5.5±3.2 25.1±7.3 461 Age,y 25-29 49.5±7.2 3.6±2.5 46.9±8.3 467 30-44 33.0±4.0 2.3±0.9 64.7±4.2 1500 >_45 12.7±3.1 2.2±1.2 85.1±3.6 1594 Fig 1.-Percentage of respondents who identified Marlboro(shaded bars)or Camel(solid bars)as the `Numbers given are percentages±95/confidence intervals. most advertised brand of cigarettes, by age of tin the youngest age group of both sexes,a small percentage did not purchase cigarettes but obtained them by respondent. Straight lines on top of bars indicate other means. 95%confidence intervals. Table 2.-Differences in Marlboro and Camel Brands Purchased by Smokers in California in 1990 Compared With Smokers in the United States in 1986 70 Market Share* 60 Adolescent Girls 1986 1990 Rate Ratio 50 Age,y n Marlboro Camel n Marlboro Camel Marlboro Camel 40 Males 30 18-29 571 49.6±5.2 5.8±2.4 1037 66.815.0 13.2±3.7 1.4 2.3 e 20 rte; 30-44 950 26.4±3.6 6.8±2.0 1579 47.7±2.9 10.5±2.1 1.8 1.5 vi 10 ?45 799 11.7±2.8 7.012.2 1316 21.1±2.7 8.1±2.8 1.8 1.2 m 0 Females r 18-29 619 39.414.9 1.011.0 928 59.0±4.8 4.5±2.1 1.5 4.5 70 C 60 Adolescent Boys 30-44 830 17.0±3.2 1.0±0.9 1500 33.0t4.0 2.3±0.9 1.9 2.3 _45 879 4.5±1.7 1.5±1.0 1594 12.7±3.1 2.211.2 2.8 1.5 cc 50 40 *Numbers given are percentages±95%confidence intervals. 30 20 10 current smokers named it more often adolescent male market: 79.9% of the than did respondents from the other 12- through 17-year-old male smokers �0 mel-t �� o*e6 groups. Camel cigarettes were named reported purchasing these brands(Ta- often ).The market share lled GJ`<g�� �Qe���Go�eR`Q�a��eJe�SR` templat ng startingbto smoke s who e(about byth twobrandsdeclined steadily owith 36%), although 95% CIs for the differ- the increasing age of the smoker:84%of ent smoking status groups overlapped. the 18- through 24-year-old market; Among girls (Fig 2, top), Marlboro 75.5% of the 25- through 29-year-old Fig 2.-Percentage of adolescent respondents who was identified as the most advertised market; and 58.2% of the 30- through identified Marlboro(shaded bars)or Camel(solid brand by more than one third of each 44-year-old market (Table 1, top). In bars)n the most advertised brand ig cigarettes,by smoking status group. As among boys, contrast to a relatively dual decline smoking status of respondent.Straight lines on top g l;z' p• g Y s Y i� of bars indicate 95 confidence intervals. girls who reported some experience with in market share across the different age smoking were the ones who most often groups for Marlboro,the decline in Cam- Identification of Brand in named Marlboro. In all but one group el's market share occurred abruptly be- Adolescents by Smoking Status (95%CIs did not overlap), Camel ciga- tween adolescent and adult groups.Mar- and Sex rettes were named less frequently than ket share decreased from 24.5%among Among teenagers,the brand that was Marlboro;among the group of teenage male adolescents to about 12% among named the most advertised varied by girls who were contemplating starting men between the ages of 18 and 44 years; smoking status among boys(Fig 2,bot- to smoke,the percentage who identified 95%CIs between the teenage and first tom)and girls(Fig 2,top).Among boys Camel as the most advertised brand adult group did not overlap. who had smoked in the month before (35%)was slightly higher(95%CIs over- The market-share pattern for females the survey,56.7%named Marlboro and lapped)than the percentage who iden- for Marlboro and Camel was somewhat 28.4% named Camel (95% CIs did not tified Marlboro. similar to that for males (Table 1, bot- overlap). Among boys who had never Reported Preference for Cigarette tom).The two brands shared 85%of the smoked but who were contemplating Brands,California a for g adolescent female market and 74.9%of smoking,44.5%identified Marlboro and the 18-through 24-year-old market;the 36.7%identified Camel as the most ad- The pattern of cigarette brands pur- market-share again declined rapidly with vertised.Thus,Marlboro was most fre- chased by California smokers paralleled age. Marlboro was the most often pur- quently named the most advertised the reported perception of advertising chased brand:63.3%of teenage female brand in all male adolescent groups;more (Table 1). smokers and 69.4% of 18- through 24- than 40%of all boys named it, and the Marlboro and Camel dominated the year-old female smokers reported pur- 3156 JAMA,December 11, 1991-Vol 266,No.22 Tobacco Advertising and Young People-Pierce et al • + chasing that brand.For Camel,the step- changes in market share caused by in- boro as the most advertised brand. function decline in the postadolescent creased targeting of advertising would Thus, the pattern of market share for market share was even more dramatic occur differentially in the youngest age each brand is different and appears to among female smokers than among male groups. mirror the pattern of recall of advertis- smokers. Camel cigarettes reached a In California in 1990, people gener- ing for that particular brand. market share of 21.7% in 12- through ally perceived Marlboro to be the most A final purpose of our analysis was to. 17-year-old girls who smoked; market advertised brand of cigarettes,followed relate changes in the market share of share then decreased to under 6%in all by Camel. These two brands were Camel compared with Marlboro since older age groups(95%CIs barely over- named most advertised by 70%of ado- the start of the new smooth character lapped). lescents and 47%of adults.Some health Camel campaign.We were particularly professionals have suggested that the interested in determining whether the Increases in Camel Market Share cartoon character advertising for the estimated change in market share for Between 1986 and 1990 Camel brand preferentially targets and the Camel brand occurred in the We examined whether the preference influences adolescents and even chil- younger age groups to which the ad- for Camel cigarettes by California smok- dren.13 This impression is bolstered by vertising appears to be targeted. Un- ers in 1990 was recent by comparing the the fact that RJR spent$100 million in fortunately, the data that we had for brand preference of California smokers 1990 both on its advertising campaign this assessment were less than optimal; in 1990 to the preferences of all US smok- and on promotional giveaways of items the only brand data that we had avail- ers in 1986. Rate ratios compare the that are appealing to adolescents. Our able were for the whole nation in 1986 percentage of California smokers who data support these concerns.The great- compared with California only in 1990. bought Camel and Marlboro in 1990 with est recognition of the Camel advertising In addition, the 1986 data set included the percentage of US smokers who campaign occurred in the youngest age adults only(18 years and older)and so bought these two brands in 1986(Table group examined in this study (ages 12 did not include the majority of teenag- 2). The rate ratios greater than 1 indi- through 13 years). Indeed, the level of ers. Accordingly, we examined the ra- cate that the percentage of smokers who recognition was inversely related to the do of change across adult age groups reported purchasing these brands of cig- age of the subject. This nearly linear between the Camel and Marlboro arettes was considerably higher in Cal- inverse trend was not demonstrated brands.We hypothesized that if Camel ifornia in 1990 than it was in the United with all cigarette advertising, as evi- advertising more effectively targets States in 1986 in each age and.sex group. denced by our results for Marlboro. younger people than does Marlboro ad- Rate ratios are given separately for While Marlboro has enormous recogni- vertising, then the increase in market three age groups and for males(top)and tion as the most advertised brand in all share for Camel will disproportionately females (bottom). age groups, peak recognition occurred come from the younger age groups The market share for Marlboro was at ages 16 through 17 years. Not only compared with the pattern of increase 40%to 80%higher among males in Cal- was the perception of Camel advertis- in market share for Marlboro. ifornia in 1990 compared with 1986 US ing highest among 12-to 13-year-olds, The market share for both Marlboro data. The market-share increase for but it was also particularly high among and Camel appears to have increased Camel was 230% in the youngest age those adolescents who were at consid- between 1986 and 1990. The market group and was 50%and 20%in the two erable risk of starting to smoke. Al- share for Marlboro appears to have in- older age groups.The proportionate in- though there is an electronic-media ban creased more in those older than 30 years creases in market share of both brands on tobacco advertising,tobacco compa- than in those younger than 30 years in were higher among females than among nies have devised other means to pro- both men and women. In contrast, the males for all age groups.Among females, mote their products to minors.20 market share for Camel has shown a the Camel brand had a low market base Marlboro and Camel had the largest marked increase among men and women in 1986.The percentage of female smok- market shares of all cigarette brands in younger than 30 years than among those ers between the ages of 18 and 29 years California in 1990.This was particularly older than 30 years. who purchased Camel cigarettes in Cal- the case in young smokers; these two We conclude that tobacco advertis- ifornia in 1990 was 450% higher than brands had 80% of the market share ing, particularly of Camel cigarettes, that for the country as a whole in 1986. among 12-through 17-year-old smokers has been effective in targeting adoles- In both males and females,Marlboro's and 84%among 18-through 24-year-old cents in the United States.(Indeed,the rate ratio increased somewhat consis- smokers. The pattern of market share 1990 CTS were undertaken during a ma- tently across age groups.However,this for both brands was the same as the jor antismoking advertising campaign was not the case for Camel: the major pattern for recognition of the advertis- in California, so our results may actu- source of Camel's market-share increase ing. In both males and females, the ally underestimate the problem.)It was appears to have come from the younger proportion of smokers who reported those in the 12- through 13-year age smokers. smoking Camel cigarettes decreased group who had the highest recall of the COMMENT with age. Furthermore, the market advertisements, and recall was partic- share for Camel among 12-through 17- ularly high among those who did not yet The data in this article add to the year-old boys was almost twice that smoke but who were considering smok- evidence'that tobacco advertising dif- among 18- to 24-year-old men. Simi- ing.These advertisements appear to in- ferentially targets young people, par- larly,the market share for Camel in 12- fluence these prospective new smokers ticularly minors. We have considered through 17-year-old girls was many to start smoking the brand that is ad- some expected consequences should to- times higher than in the next highest vertised. Our data suggest an associa- bacco advertising target minors: first, age group (18 through 24 years). The tion between the brands identified as the perception of the advertising would peak market share for Marlboro was in most highly advertised and market share be inversely related to age;second,the the 18-to 24-year-old age group for both for the same brands. Not only is the pattern of market share across age and males and females,and the age pattern market share of brands similar to recall sex groups would follow the pattern of of the market share was similar to the of the most advertised brands, but the the perceived advertising; and third, age pattern for identification of Marl- brand that appears to be aimed the most JAMA, December 11, 1991—Vol 266, No.22 Tobacco Advertising and Young People—Pierce et al 3157 at adolescents has demonstrated a dif- is strong evidence that most smokers start to smoke24•25 may be in jeopardy. ferential increase in market share in the become addicted when they are minors Given the potential harm to the health. youngest adolescents overtime.Resolv- and do not understand the long-term of future generations, public policy ing this issue will require a longitudinal consequences of smoking.2,11 Overcom- should, as a matter of urgency, extend study, which we hope to undertake in ing this addiction is a lifelong struggle the ban on tobacco advertising to cover the near future. for many smokers'21 and the failure rate not only electronic media but also all Our results suggest that tobacco ad- for cessation attempts is extremely other forms of cigarette advertising and vertising is causally related to young high.21,22 Unless those responsible for promotion.2-1-28 people becoming addicted to cigarettes; protecting our young take action quickly, This study was supported by contract 89-97872 the sum of this evidence is considerable the declines that we have seen recently from the California Department of Health Ser- although not yet complete. Also, there in the proportion of our youngsters who vices,Tobacco Control Section,Sacramento. References 1. Davis R. Current trends in cigarette advertis- ucts Before the Subcommittee on Health and the ing.N Engl J Med. 1991;324:913-917. ing and marketing. N Engl J Med. 1987;316:725- EnzriromnentoftlwCommittee onEnergyandCom- 21. Pierce JP.The quitting process.In:Institute 732. merce,99th Congress,2nd Sess(1986)(testimony for the Study of Smoking Behavior and Policy. 2. Surgeon General. The Health Consequences of of G.Weil). Smoking Cessation: The Organization, Delivery Smoking:25 Years of Progress:A Report of the 11. Pierce JP, Naquin M, Gilpin E, Giovino G, and Financing of Services.Cambridge,Mass:Har- Surgeon General.Rockville,Md:US Dept of Health Mills S,Marcus S.Smoking initiation in the United vard University Press;1990. and Human Services;1989.US Dept of Health and States: a role for worksite and college smoking 22. Surgeon General.The Health.Consequences of Human Services publication(CDC)89-8411. bans.J Natl Cancer Inst. 1991;83:1009-1013. Smoking:Nicotine Addiction:A Report of the Sur- 3. Raftery J.Advertising and smoking:a smoul- 12. Goad GP. Canada's tobacco ad ban is over- geon General. Rockville,Md:US Dept of Health dering debate.Br J Addict. 1989;84:1241-1246. turned by judge. Wall St J.July 29, 1991. and Human Services;1988.US Dept of Health and 4. Centers for Disease Control. Cigarette adver- 13. Bird L.Joe Smooth for president. ADweeks Human Services publication(CDC)88-8406. tising-United States,1988.MMWR.1990;39:261- Marketing Wk.May 20, 1991:20-22. 23. Hatziandreu EJ,Pierce JP,Lefkopoulou M,et 265. 14. Doolittle DE.Joe Camel takes Reynolds over al.Quitting smoking in the United States in 1986. 5. Chapman S,Fitzgerald B.Brand preference and the hump. Tobacco Rep. February 1991:14-18. J Natl Cancer Inst. 1990;82:1402-1406. advertising recall in adolescent smokers:some im- 15. Griscom TE. The 'smooth character'appeal. 24. Fiore MF, Novotny TE,Pierce JP, Hatzian- plications for health promotion.Am JPublic Health. Washington Post.June 1,1991:21. dreuE,Patel K,Davis R.Trends incigarette smok- 1982;72:491-494. 16. Waksberg J. Sampling methods for random- ing in the United States:the changing influence of 6. Klintzner M, Gruenewald PJ, Bamberger E. digit dialing.J Ant Stat Assoc. 1978;73:4046. gender and race.JAMA. 1989;261:49-55. Cigarette advertising and adolescent experimen- 17. US Dept of Health and Human Services. To- 25. Pierce JP,Fiore MC, Novotny TE, Hatzian- tation with smoking.Br J Addict.1991;86:287-298. bacco Use in 1986:Methods and Basic Tabulations dreu E,Davis R.Trends in cigarette consumption 7. Chapman S. On the limitations of econometric From Adult Use of Tobacco Survey.Rockville,Md: in the United States:educational differences are analysis in cigarette advertising studies.Br J Ad- Public Health Service;1986. increasing.JAMA. 1989;261:56-60. dict. 1989;84:1267-1274. 18. US Dept of Health and Human Services.Smok- 26. Warner KE.Effects of antismoking campaign: 8. Waterson W. Advertising and cigarette con- ing Tobacco and Cancer Program: 1985 Report. an update.Am J Public Health. 1989;79:144-151. sumption.Lond Advertising Assoc.December 1982. Bethesda,Md:National Institutes of Health;1986. 27. Tye JB,Warner KE,Glantz SA.Tobacco ad- 9. Hearings on Advertising and Tobacco Products Publication NIH 86-2687. vertising and consumption: evidence of a causal Before the Subcommittee on Health and the Envi- 19. Efron B. The Jackknife, the Bootstrap and relationship. J Public Health Policy. 1987;8:492- ronmtent of the Committee on Energy and Com- Other Resampling Plans. Philadelphia, Pa:Soci- 508. merce,99th Congress,2nd Sess(1986)(testimony ety for Industrial and Applied Mathematics;1982. 28. Pierce JP. Time to ban cigarette advertising of J.O'Toole). 20. Blum A. The Marlboro Grand Prix: circum- and continue the`Quit for Life'campaigns.Med J 10. Hearings on Advertising and Tobacco Prod- vention of the television ban on tobacco advertis- Aust. 1990;152:113-114. 3158 JAMA, December 11, 1991-Vol 266,No.22 Tobacco Advertising and Young People-Pierce et al Active Enforcement of Cigarette Control Laws in the Prevention of Cigarette Sales to Minors Leonard A. Jason, PhD; Peter Y. Ji; Michael D. Anes; Scott H. Birkhead Objective.—To assess the effect that cigarette legislation would have on re- METHOD ducing merchant sales rates of cigarettes to minors and the affect on adoles- This study was initiated as a result of Cent smoking behavior. a private citizen's complaint regarding a Design.—Observational survey of merchant selling behaviors and adoles- minor's possession of cigarettes to Of- cent smoking habits before and after passage of legislation. ficer Bruce Talbot of the Woodridge,Ill Setting.—The setting for the merchant survey was Woodridge, III(population (population,25 200),police department. 25200), a suburban community of Chicago. The surveys were distributed to In August 1988, Officer Talbot sent a adolescents in the local junior high school. letter to all cigarette vendors in Woo- Participants.—Convenience sample of both merchants and adolescent Stu- diidge detailing the state law prohibit- dents. ing cigarette sales to minors. In addi- Intervention.—Passage of community antismoking legislation. tion, following a media focus on a De- Main Outcome Measures.—Percentage of stores selling cigarettes to Paul University study of cigarette sales minors in Woodridge and percentage of students who had experimented with to minors in Chicago .and P.Y.J., unpublished data, 1991)91),, OOfficer Talbot cigarettes or were regular smokers. invited one of the authors (L.A.J.) to Results.—Merchant sales rates in Woodridge decreased from a baseline of assess cigarette sales rates in Wood- 70%before legislation to less than 5%in 1.5 years of compliance checking af- ridge. These measurements were done ter legislation. Student surveys showed that the rates of cigarette experimen- in August and November 1988 and Feb- tation and regular use of cigarettes by adolescents were reduced by over 50%. ruary 1989. The DePaul University re- Conclusion.—Cigarette control lbws can be effective in significantly reduc- search team also distributed a question- ing the rate of cigarettes sold by merchants and rates of cigarette use by ad- naire to students at the local junior high olescents.Key elements of successful legislation implementation are consistent school in March 1989 to determine the compliance checking and heightened community awareness of the problems number of adolescent smokers and their and prevalence of adolescent smoking. smoking habits. In May 1989, new leg- (JAMA. 1991;266:3159-3161) islation was passed restricting cigarette sales in Woodridge,and cigarette sales were tracked after the legislation was DESPITE the health hazards of smok- tions were sent to merchants. Despite passed. ing,approximately 3000 adolescents be- Buffalo's educational efforts, a citywide During each checking period,minors come new smokers daily.'A recent study investigation showed that 77%of stores 12 to 13 years of age(all of whom were by DiFranza and Tye2 has estimated receiving educational packages sold cig- rated by independent judges as looking over$1 billion in illegal sales of tobacco arettes to minors.'In Santa Clara,Calif, less than 18 years of age)were sent into products to minors.Although state laws an aggressive campaign was used to alert stores to purchase a pack of cigarettes. prohibit sales to minors, minors fre- merchants to cigarette laws. There was Unobtrusively, a Woodridge police of- quently have little difficulty in purchas- an initial reduction of cigarette sales to ficial or a DePaul University research ing cigarettes.Across the nation,active minors from 74% to 39%,6 yet vending assistant observed the transaction. tobacco-control investigators have sent machine sales were unchanged.A 1-year Vending machine sales were also tracked minors, under supervision, into stores follow-up showed that the merchant sales by sending minors to purchase cigarettes to purchase cigarettes. In Santa Clara had rebounded to 60%.' from these machines. There was 100% County, California, 74%of stores sam- agreement between two independent pled in this manner sold tobacco prod- judges as to whether an illegal sale oc- ucts to minors.' In Massachusetts, an See also pp 3168 and 3188. curred. 11-year-old child was able to purchase Sales assessments conducted before cigarettes from stores 75%of the time.' The following study sought to assess, legislation to determine if Officer Tal- Campaigns to alert merchants to cig- by periodic checking, the effects of li- bot's letter to merchants was effective, arette laws are insufficient for restricting censing and enforcement of legislative were made in August and November cigarette sales. In Buffalo, NY, letters provisions on cigarette selling behav- 1988 and February 1989. In the spring describing the city's cigarette sales law, iors of merchants and minors' accessi- of 1989,Officer Talbot and other Wood- warning signs,and enforcement instruc- bility to cigarettes. Through distribu- ridge officials drafted cigarette control tion of questionnaires before and after legislation modeled after the city's li- From the Department or Psychology, DePaul Unl- passage of legislation,we also sought to quor control laws.This was done to treat versity,Chicago,Ill. determine the effect of the law on the the sales of tobacco and alcohol, both DPaulReprint Un requests sity,2323 NDSema rtment ary Ave,Cof hicago,IL prevalence of adolescent cigarette use age-restricted products,in precisely the 60614(Dr Jason). in a local junior high school. same manner. Using the liquor laws as JAMA, December 11, 1991—Vol 266, No.22 Active Enforcement of Cigarette Control Laws—Jason et al 3159 I Sales Rates and Offense Data Before and After Passage of Community Smoking Legislation actment of the ordinance are shown in the Table. In addition, in each of the. No.of Ates No.of StoresFirstSecond Data Stores Rats,x Complying Ofenae Offense time periods predating the legislation, 9•fore Passage sales from three vending machines in August 1988 20 70 s Woodridge were shown to be 100%. November 1988 20 80 8 The results of compliance checking of February 1989 19 79 a Woodridge merchants conducted after After Passage the passage of the legislation are shown June 1989 23 35 1s s in the Table. August 1989 22 36 is a a From March through May 1989, the November 1989 22 0 22 0 0 interval between the compliance checks January 22 0 22 0 0 a 28 1 0 before and after passage of the legisla- April 19900 327 a 28 1 00 tion,the number of vending machines in July 19so Woodridge increased from three to six. December 1990 30 3 1 ° In the first assessment after passage of the legislation(June 1989),three of six vendingmachinessold cigarettesinWood- a guide had the additional effect of tak- during compliance checkingperiods,a con- ridge.The number of vending machines ing advantage of existing civil enforce- gratulatory note was sent from the may- decreased from six in January 1990 to ment structures, such as the mayor's or's office. two in December 1990. In January, office,for issuing sales licenses and fol- The possession clause of the Wood- April,and July 1990,there were no vend- lowing up violations.The new cigarette ridge ordinance allows police officers to ing machine sales. In December 1990 legislation,containing licensing,enforce- issue a ticket to any minor caught with there was one vending machine sale. ment, and possession provisions, was illegal tobacco products.This ticket car- Fifty minors were cited for posses- passed May 1, 1989, with a 30-day en- ries a fine of$25.The minor can pay the sion of cigarettes in the 1.5 years after forcement grace period. Vendors were fine immediately at-the police depart- passage of the ordinance. These minors issued cigarette sales licenses and billed ment,in a manner similar to paying for were assessed a$25 fine and their cig- a licensing fee of$50. a parking ticket, or can wait for a civil arettes were confiscated.Four of the 50 Reports of the cigarette sales assess- hearing on the possession charge. minors were repeat offenders during this ment and the passage of the law were The police department in Woodridge time period. aired on local television stations and pub- has played an important role in the en- A total of 680 local seventh- and lished in local newspapers to inform com- forcement of penalties.First,after pas- eighth-grade students were surveyed munity members. During the grace pe- sage of the ordinance, all officers were in March 1989,-before the law was riod,Officer Talbot personally delivered informed of the provisions by a depart- passed. Results from this survey indi- a copy of the law and a tip sheet describ- mental memorandum from the chief of cated that 46%of the students had ex- ing all forms of valid age identification police.Second,if a store incurred a 1-day perimented with cigarettes(eg,had tried issued by the state of Illinois to every suspension, the mayor's office notified cigarettes on at least one occasion)and cigarette vendor in Woodridge.Vendor's the police,and all officers patrolling the 16% were regular smokers. In April questions about the law were answered area were reminded at morning roll call 1991,almost 2 years after passage of the during these face-to-face store visits,each to keep a watchful eye on the store. Woodridge ordinance,639 local seventh- of which took approximately 15 minutes. The police also maintained heightened and eighth-grade students were sur- All store visits were concluded in 1 day. community awareness in Woodridge by veyed,and 23%reported experimenta- The Woodridge police department sending letters to two different popu- tion with cigarettes, with only 5% de- planned quarterly"stings"to check mer- lations on a regular basis.At the end of scribing themselves as regular smok- chant compliance after the grace period, every school year,police sent a letter to ers.Survey return rates were at 90%or and all stores were checked regularly(Ta- merchants warning them that during above both before and after passage of ble). When a violation occurred,the po- the summer months more adolescents the legislation. lice officer wrote a report on the violation would be in their stores,reminding them Other important data from the April and sent it to the mayor's office(the mayor of the ordinance, and asking for their 1991 survey revealed that 77% of the is the tobacco commissioner under the continued support. Also, at the begin- smokers cited friends,parents,siblings, ordinance,as well as liquor commissioner). ning of every school year, Woodridge or others as sources of their cigarettes, The mayor's office then sent a letter in- schoolchildren were given an informa- 17% cited stores or vending machines forming the merchant of his or her right tive letter about the ordinance to take outside of Woodridge as their source, to appeal within 10 days.If the merchant home to their parents.This letter asked and 6%cited stores or vending machines chose to appeal,a civil hearing was held; for continued compliance and stressed within Woodridge.Additionally,86%of otherwise,the mayor issued a warning to the importance of preventing adolescent student respondents knew of the Wood- the merchant, and either suspended the tobacco,access. ridge law and 69% felt the law would cigarette-sellinglicense for 1 ormore days, A follow-up questionnaire was distrib- either prevent their procurement of cig- and/or imposed a monetary fine of up to uted tojunior high school students in April arettes or make it harder for them to $500.First offenses incurred a warning in 1991,almost 2 years after passage of the obtain cigarettes. the first year of enforcement,with a fine ordinance. This questionnaire sought to COMMENT and/or suspension imposed if a second vi- determine the effects of the Woodridge olation occurred during the subsequent ordinance on the number of adolescent Two significant findings have emerged compliance check. In the second year of smokers and their smoking habits. from this study. Principally, to our enforcement, the tobacco commissioner RESULTS knowledge,Woodridge is the first com- has tended to enforce the ordinance more munity to successfully reduce cigarette stringently than in the first year. If the Results of the assessment of Wood- sales rates to minors to a minimal level merchant did not sell cigarettes to minors ridge merchants'sales rates before en- as a result of legislation. The key ele- 3160 JAMA, December 11,1991—Vol 266,No.22 Active Enforcement of Cigarette Control Laws—Jason et al i mems of the legislation are vendor li- taining positive police-merchant inter- caught.Legislation was effective in Woo- censing,active compliance checking,and action.Thank-you letters from the may- dridge because(1)police conducted reg- penalties for merchant sales violations or's office helped to instill a sense of ularly scheduled compliance checks and and minors'possession of cigarettes.Sec- merchant solidarity and promoted mer- used uniform checking methods,(2)leg- ond, the combined effects of this legis- chant-community ties. Finally, the po- islation was coupled with a strong ed- lation and a community awareness of lice department's letter describing the ucational message from the police to com- the problem of adolescent smoking have ordinance that was given to schoolchil- 'munity members at the beginning and substantially reduced the frequency of dren to take home to their parents and end of every school year, and (3) the adolescent experimentation with ciga- police efforts toward enforcement helped community's success received consistent rettes and regular smoking. to keep families focused on the issue of media exposure. The licensing process in Woodridge is adolescent smoking. efficient because it takes advantage of School survey results indicate that the We would like to thank students Nick Cicarelli, age 13 years,existing enforcement procedures de- Woodridge law has substantially reduced Kelter,age 13 years and Jenny Haut age 12 yen Vega,age 13 years, arts signed to control sales of alcohol to mi- the number of adolescents who smoke. for participating in the Woodridge compliance nors.The purpose of licensing is to mo- The reduction of merchant cigarette checks. tivate a store to monitor itself in played p or role in a major a- sales has la keeping We also thank Richard Russell,principal of Jef- ferson ferson Junior High School in Woodridge,Ill,for al- rette sales,rather than face license sus- cigarettes from minors. Another possi- lowing access to students for our surveys and pension. If a temporary suspension is ble contribution to this reduction is the Woodridge Police Department Juvenile Officer made, it is likely to have a greater im- possession clause of the ordinance.The Will Sperling for his helpful comments. pact than a fine, because so much of a possession clause may provide an addi- Last, we would like to thank Sergeant Bruce store's profit is made through the sale of tional deterrent to experimental ci a- cBuzz"Talbot for his continued effort,support,and p g p g contribution to the Woodridge City Tobacco Ordi- cigarettes, especially in large stores. rette use by minors.Although opponents nance, without which this article would not have The legislation and enforcement suc- of penalties for possession in cigarette been possible. cess in Woodridge must also be attrib- legislation have insisted that these References uted to the commitment of the police clauses are a case of"blaming the vic- department and mayor's office to con- tim"for cigarette use,we feel that pos- 1. Pierce R Fiore Trends Novotny TE, ekinHatzg in the EJ,Davis RM.Trends in cigarette smoking in the tinually reinforce to the community the session clauses are one part of an overall United States:projects to the year 2000.JAMA. problem of adolescent smoking.The in- plan—one part that may have a deter- 1989;261:61-65. volvement of the local and Chicago met- rent effect. 2. DiFranza JR,Tye JB.Who profits from tobacco ropolitan area media in reporting on cig- Traditional educational approaches to sales to children?JAMA. 1990;263:2784-2787. arette sales before passage of the ordi- reducing merchant cigarette sales have 3. Altman DG,Foster V,Rasenick-Douss L,Tye I1 g g g JB. Reducing the illegal sale of cigarettes to mi- nance,the ordinance passage itself,and resulted in only short-term reductions nors.JAMA. 1989;261:80-83. subsequent successes fostered commu- presumably because an educational mes- 4. DiFranza JR, Norwood BD,Garner DW,Tye nity pride in the initiative. The police sage instills little motivation for change. JB. Legislative efforts to protect children from visits to merchants to clarify the law, Without the possibility of direct action tobacco.JAMA.ors'a cess 7-3389. � I> Y � 5. Reducing minors' access to tobacco. Tobacco internal police briefings regarding mer- merchants will fail to respond to edu- Youth Rep.Autumn 1990;4:17. chant violations,the police letter to mer cational programs to reduce the number 6. STAT campaign cuts illegal sales of tobacco to chants at the beginning of the summer, of adolescent smokers because of the minors in half. Tobacco Youth Rep. Autumn and the earl age-identification ti tremendous profits involved in cigarette 7. Mo:1. yearly g P A $� 7. More being done to prevent cigarette sales to sheets were all integral parts of main- sales and the certainty of not being minors. Tobacco Youth Rep.Autumn 1989;4:6. JAMA,December 11, 1991—Vol 266, No.22 Active Enforcement of Cigarette Control Laws—Jason et al 3161 No-Smoking Laws in the United States ' , An Analysis of State and City Actions to Limit Smoking in Public Places and Workplaces Nancy A. Rigotti, MD, Chris L. Pashos, PhD Objective.—To assess the prevalence, content, and growth of state and city may have the greatest impact on smok- laws restricting smoking in public places and workplaces in the United States ing behavior because adults spend more and to identify factors associated with their passage. time at work than in any other single Design.—A mailed survey of city clerks in US cities with a population of place outside the home." 25 000 or greater(N=980) and review of existing data sources confirmed the Despite their potential importance, status of smoking restrictions in 902(92%)of the cities in the sample.State laws little is known about the extent of laws were identified by contacting each state's Legislative Reference Bureau(100% restricting smoking or the effects of response). Content of laws was coded using previously developed categories. these laws,particularly at the local go v- ernment level.It would be useful to know Main Outcome Measures.—Prevalence,comprehensiveness, and cumula- the scope of existing legislation and,con- tive incidence of no-smoking laws in states and in cities with a population of versely,areas of need,to foster the wide- 25000 or greater. spread enactment of effective laws.This Results.—By July 1989,44 states and 500(51%)of the cities in our sample is particularly important when the to- had adopted some smoking restriction,but content varied widely.While 42%of bacco industry is moving aggressively cities limited smoking in government buildings, 27% in public places, 24% in to block new legislation and neutralize restaurants,and 18%in private workplaces,only 17%of cities and 20%of states current laws (Tobacco Observer. Octo- had comprehensive laws restricting smoking in all four of these sites.The num- ber 1988;13:1;Governing. May 1989:34; ber of city no-smoking laws increased tenfold from 1980 to 1989. City and New York Times. December 24, no-smoking laws were independently associated with population size, geogra- sary t). This information is also lanecws on sary to evaluate the effects of laws on phy,state tobacco production,and adult smoking prevalence.Laws were more critical measures, such as smoking be- common in larger cities, Western cities, and states with fewer adult smokers. havior. Laws were less common in tobacco-producing states and in the South. Previously, we summarized federal Conclusions.—No-smoking laws are more widespread than previously ap- regulations and state laws through 1987 preciated,especially at the local level,reflecting a rapid pace of city government and developed a scale to rate their com- action in the 1980s. Nonetheless, comprehensive laws, which are most likely prehensiveness.2 City,town,and county to provide meaningful protection from environmental tobacco smoke exposure, government actions have not been an- remain uncommon and represent a major gap in smoking control policy. Laws alyzed. Local government action has are most needed in smaller and non-Western cities and in states that produce been more difficult to follow because the number of communities to monitor is tobacco and have a higher proportion of smokers. large and there is no central registry of (JAMA. 1991;266:3162-3167) local legislation. Two groups—Ameri- cans for Nonsmokers Rights(ANR), a CONCERN about the health effects of smoking on virtually all domestic airline national advocacy group, and Tobacco- environmental tobacco smoke(ETS)ex- flights—most public-sector actions have Free America(TFA),a coalition of vol- posure and growing public antismoking occurred at the state and local levels. untary health organizations—indepen- sentiment have led to a wave of activity The content of city and state no-smok- dently compile laws and produce peri- limiting smoking in public places and ing laws varies widely,and the result is odic reports. Their results are similar workplaces in the United States."Ac- apatchwork of smoking restrictions that but not identical.In August 1988,ANR tions range from federal, state, and lo- affect a growing number of Americans identified 321 local ordinances with pro- cal legislation to limits adopted volun- at work and in such public places as visions for"significant nonsmoker pro- tarily by private businesses, hospitals, government buildings,retail stores,res- tection";2 months later,TFA reported hotels, and schools. While the federal taurants, theaters, and sports arenas. that 380 local communities had laws re- government has taken some action— The intent of these laws and policies stricting smoking in public places.2,8,9 most prominently, legislation banning is to reduce individuals'ETS exposure, One year later,ANR and TFA estimates but the impact may be considerably had risen to 397 and 440,respectively.10,11 From the Institute for the Study of Smoking Behavior broader.These laws represent a critical These figures include city and county and Policy,John F.Kennedy School of Government, new direction in tobacco-control policy, laws and municipalities of all sizes.How- Harvard University(Drs Rigotti and Pasho ),and the with the potential to influence social ever, because the do not s stemati- General Internal Medicine Unit, Massachusachu setts Gen- p y y eral Hospital, Harvard Medical School (Dr Rigotti), norms governing tobacco smoking and, cally survey communities to obtain in- Boston,Mass.Dr Pashos is now with the Department ultimately,to affect smoking behavior. formation, ANR and TFA may miss of Health Care Policy,Harvard Medical School. Smokingrestrictions are most likely to some local actions underestimatingthe An earlier version of this report was presented at the y � Seventh World Conference on Tobacco and Health, have these effects if they are compre- extent of local no-smoking legislation. Perth,Australia,April 4,1990. hensive in coverage. Limits on work- The purpose of this study was to de- Reprint requests to the General Internal Medicine lace smoking should provide the eat- scribe the prevalence, incidence, and Unit, Massachusetts General Hospital, Fruit Street, p g p � Boston,MA 02114(Dr Rigotti). est protection from ETS exposure and content of state and local no-smoking 3162 JAMA,December 11, 1991—Vol 266,No.22 No-Smoking Laws—Rigotti&Pashos laws'and to identify factors associated or workplace"and to mail us a copy of ered major tobacco producers;jointly, with their passage. To do so, we up- any law. A second mailing was sent to they account for over 90%of US tobacco dated and extended our previous work nonresponders.We received responses acreage and output.11 Data on adult(age on state no-smoking laws, and we de- from 753 of 831 cities (91% response ,18 years)smoking prevalence by state veloped a method to monitor the prev- rate). For the analysis, a city was were obtained from the 1989 Current alence and describe the comprehensive- counted as having a smoking restriction Population Survey.is ness of local legislation in a defined sam- if it met one of two criteria: (1)we had To identify factors characterizing cit- ple,US cities with a population of 25 000 a copy of a city or town ordinance or ies with no-smoking laws in 1989, we or greater. regulation or (2) both compendia re- conducted similar analyses,using as co- We sought to answer these questions: ported that the city had a no-smoking variates city population,geographic lo- (1) What proportion of US cities and ordinance."," County ordinances and cation, state tobacco production, and states have adopted restrictions on laws in cities with a population of less state smoking prevalence.Because city- smoking in public and at work?(2)How than 25 000 were excluded. specific smoking prevalence is not avail- comprehensive are state and local smok- able for a national sample, we used the ing restrictions? (3) How rapidly have Categorization of Laws corresponding state data as a proxy.To these laws spread throughout the United We reviewed each law or regulation adjust for the potential confounding ef- States?(4)Do states and cities with laws to determine whether it met our defi- fect of coexisting state laws on local gov- differ from those without laws in size, nition of a no-smoking ordinance:legis- ernment action, we added state legis- geographic location,tobacco production, lation (or regulation) that specifically lation as an independent variable in anal- or smoking prevalence? restricted or banned tobacco smoking in yses of city laws. Analyses were done Our previous work suggested that one or more public places or workplaces (1) including all smoking restrictions state laws were less common in South- for the purpose of protecting individu- (laws and regulations)and(2)limited to ern,tobacco-producing states. For this als from ETS exposure. A law was ex- laws. Because results were similar,we analysis,we hypothesized that both city cluded if its primary intent was protec- report the results of the former, more and state laws would be less common in tion from fire or explosion hazard.Local comprehensive analysis. Southern states, tobacco states, and laws that had been preempted by sub- The significance of relationships found states with more smokers. sequent state laws were also excluded. on univariate analyses was tested with METHODS Four states had such laws in 1989(Flor- X2 analysis,X2 test of trend,and Fisher's ida, Oklahoma, Pennsylvania, and Vir- Exact Test.16," The independence of Data Collection ginia),but only the Florida law nullified these relationships was determined by States.—To identify current state existing local ordinances, and we ex- conducting multiple logistic regression laws, we updated our previous compi- cluded Florida local ordinances from the analyses to estimate the probability that lation by requesting a copy of any state analysis. All laws were coded by a sin- a state or city had(1)any local smoking no-smoking law from the director of each gle rater. The content of each law or law and (2) a comprehensive no-smok- state's Legislative Reference Bureau in regulation was categorized according to ing ordinance. For the univariate and September 1989. Responses were ob- the types of places in which smoking multivariate analyses, we considered tained from 49 of 50 states and the Dis- was restricted or banned, using cate- that cities whose status was not known trict of Columbia;a telephone call to the gories developed for our previous anal- (n=78[8%1)had no law. Relationships nonresponding state (Missouri) con- ysis': (1)one to three public places, (2) did not change when we repeated the firmed that it had no law. more than three public places, (3) res- analysis with these cities excluded. Cities.—Using 1980 census data,12 we taurants, and (4) private-sector work RESULTS identified all US cities and towns with a sites. We added a fifth category, gov- population of 25 000 or greater(N=980); ernment buildings, for this analysis. City No-Smoking Laws this was our sample. We updated pop- Laws that limited or banned smoking in Sources of Information.—We con- ulation figures for cities with a popula- all these categories were considered firmed the presence or absence of a law tion of 50 000 or greater using`1986 comprehensive laws.For city laws cited restricting.smoking in 902(92%)of the data.13(Updated figures for cities with by both ANR and TFA, we used the 980 US cities in our sample.In response a population of 25 000 to 50 000 were not categorizations in their manuals to de- to our mailing, we obtained copies of available.)We used two data sources to termine coverage. Accuracy of this smoking restrictions from 404 cities;275 determine whether a smoking restric- method was confirmed by comparing the of these had not been previously re- tion was present: (1) information pro- categorization of these data sources with ported,while 129 were cited by another vided by city clerks in response to a our rater's coding of the actual law,us- source (ANR or TFA). An additional mailed request and(2)existing compen- ing the 40 laws obtained from TFA. . 109 cities were classified as having con- dia of laws.10,11 We requested informa- fumed laws because both ANR and TFA tion from all cities in the sample not Analyses cited them as having laws.10," We ex- cited by both sources as having laws We calculated the prevalence of state eluded laws from Florida cities(n=13) (n=831). We reasoned that there was laws as of December 31, 1989, and the because state law had nullified them by little chance that a city cited by two prevalence of city laws as of July 1,1989. 1989.City clerks confirmed the absence independent sources did not have a law. To identify characteristics of states with of a smoking restriction in 389 cities We confirmed this by obtaining from no-smoking laws in 1989,we conducted (40%), and the 13 Florida cities were TFA copies of 100%of ordinances from univariate and multivariate analyses,us- also classified as lacking laws.We clas- a random sample of 40 of the 149 cities ing as covariates geographic location, sified 78 cities(8%)as unknown;we had cited by both ANR and TFA as having state tobacco production, and state no information about 73 of them and five laws. From April through June 1989, a smoking prevalence. Geographic loca- were cited by only one source. mailing to city clerks asked them to in- tion was categorized by US.Census re- Prevalence and Content.—By July dicate whether the city had passed or gion (n=4). Six states (Georgia, Ken- 1, 1989, over half of local governments was considering"any law or regulation tucky,North Carolina,South Carolina, in our sample had taken some action to that restricts smoking in a public place Tennessee, and Virginia) were consid- limit smoking. Smoking restrictions in JAMA, December 11, 1991—Vol 266, No.22 No-Smoking Laws—Rigotti&Pashos 3163 public places or workplaces had been Table 1.—Smoking Restrictions Adopted by US Cities With a Population of 25 000 or Greater(1989) adopted by 513 (52%) of the US cities Urban with a population of 25 000 or greater. Population in After excluding Florida towns, 500 cit- cities,No. Millions,No. ies (51%) had laws in force (Table 1). Any smoking restriction Soo (51) 67.8 (68) Eighty percent of these restrictions No restriction 402 (41) 26.7 (27) were city ordinances;399 cities had en- status unknown 78 (8) 4.7 (5) acted laws and 112 had adopted admin- Total 960 (100) 99.2 (100) istrative regulations (11 cities had Smoking restricted in both).Almost all administrative actions Government buildings 416 (42) 58.4 (59) limited smoking in government build- 1 to 3 Public places 314 (32) 54.5 (55) ings,while laws were generally broader >3 Public places 260 (27) 48.0 (48) in coverage. Table 1 displays what city Restaurants 239 (24) 44.4 (44) smoking restrictions covered. Cities Private-sector workplaces 173 (18) 36.5 (37) most often limited smoking in govern- Comprehensive law* 165 (17) 35.7 (36) ment buildings (42%), followed in de- *Restricts smoking in government buildings, more than three public places, restaurants, and private-sector scending order by restrictions on smok- workplaces. ing in one to three public places(32%), more than three public places (27%), restaurants (24%), and private work- places(18%). Comprehensive laws had 600 been adopted by 165 cities (17%). The extent of local government action on 500 smoking appears greater when ex- pressed in terms of population covered (Table 1). Over two thirds of the nearly 3 400 100 million Americans residing in cities with a population of 25 000 or greater o 300 were covered by some type of local z6 smoking restriction,and over one third 200 lived in cities with comprehensive no- smoking ordinances. Incidence.—Figure l displays the cu- EM 100 mulative incidence of local laws between 1980 and July 1, 1989. It demonstrates 0 a tenfold increase in the number of local 1980 1981 1982 1983 1984 1885 1986 1987 1988 1989 laws over the decade.Two types of laws Year grew most rapidly:comprehensive laws 111111111 and those limited to government Comprehensive Law ® 1 to 3 Public Places build- ings.Each year between 1982 and 1988, ® Restaurants and Public Places SM Government Buildings Only more laws passed than in the preceding 033 >3 Public Places C7 Other year.On an annualized basis,fewer laws passed in 1989, but data are complete Fig 1.—Cumulative number of local no-smoking laws in US cities with a population of 25000 or greater from only through July 1. 1980 through 1989. Variability.—Table 2 displays the variability of local smoking restrictions among cities grouped by population,ge- To examine the relationship be- State No-Smoking laws ography,state tobacco production,and tween city laws and smoking preva- Prevalence and Incidence.—By the state no-smoking law. The prevalence lence, we added 1989 state smoking end of 1989,all but six states had adopted of all types of smoking restrictions in- prevalence (used as a proxy for city some restriction on smoking in public creases with city population size.Cities prevalence) to the multivariate mod- places or workplaces, but these laws in the six major tobacco-producing states els (Table 3). It was independently varied widely in comprehensiveness.Ta- have significantly fewer laws,especially associated with a city's likelihood of ble 4 shows what state no-smoking laws more comprehensive laws,while West- having a no-smoking law, even after covered.Most states(n=35)limited smok- ern cities have more laws than other adjusting for the effects of city popu- ing in public places, but fewer than half regions. There were fewer city laws in lation, location, and state tobacco pro- required restaurants to provide no-smok- states with comprehensive state laws, duction by multiple logistic regres- ing sections(n=22),and only 12 addressed presumably because state law obviated sion. Thus, cities in states with fewer smoking in private work sites. Only 10 the need for local government action. smokers were more likely to have states had comprehensive laws.Figure 2 There were also fewer city laws in states adopted no-smoking laws. The addi- displays the cumulative incidence of state with no state smoking laws than in states tion of smoking prevalence to the laws since 1970,demonstrating that,while with noncomprehensive laws. Multiple analysis did not markedly alter previ- prevalence has increased since the mid- logistic regression analysis confirmed the ously described relationships between 1970s,comprehensive laws are largely a independence of these univariate rela- city laws and population, region, or product of the 1980s. tionships for all smoking laws and for tobacco production. Results of multi- Variability.—Table 4 also demon- comprehensive laws (Table 3). There variate analysis were similar whether stratesthe variability inno-smoking laws were no significant two-way interactions administrative regulations were in- by geographic region and tobacco gro- in these models. cluded or excluded. duction.Comprehensive laws were more 3164 JAMA,December 11,1991—Vol 266,No.22 No-Smoking Laws—Rigotti&Pashos Table 2.-Prevalence of Local No-Smoking Laws by Population,Region,Tobacco Production,and State Law Local No-Smoking Laws,No.(%) No.of Any Govemment >3 Public Private Comprehensive Cities Law` Bulidings' Places Restaurants Workplaces Lawt All cities 980 500 (51) 416 (42) 260 (27) 239 (24) 173 (18) 165 (17) Population(in thousands) 25-50 524 229 (44) 191 (36) 96 (18) 89 (17) 64 (12) 61 (12) 50-100 273 142 (52) 115 (42) 78 (29) 75 (27) 52 (19) 49 (18) 100-250 122 76 (62) 65 (53) 48 (39) 41 (34) 30 (25) 29 (24) >250 61 53 (87) 45 (74) 38 (62) 34 (56) 27 (44) 26 (43) P# <.0001 <.0001 <:0001 <.0001 <.0001 <.0001 Geographic region West 257 180 (70) 152 (59) 130 (51) 123 (48) 110 (43) 106 (41) Midwest 284 134 (47) 112 (39) 53 (19) 43 (15) 31 (11) 30 (11) South 252 109 (43) 92 (37) 60 (24) 50 (20) 21 (8) 21 (8) Northeast 187 77 (41) 60 (32) 17 (9) 23 (12) 11 (6) 8 (4) P§ <.0001 <.0001 <A001 <,0001 <.0001 <.0001 Major tobacco-producing statell Yes 74 29 (39) 23 (31) 11 (15) 11 (15) 3 (4) 3 (4) No 906 471 (52) 393 (43) 249 (27) 228 (25) 170 (19) 162 (18) P§ .034 .040 .018 .047 .0007¶ .00111 State no-smoking law# None/minimal 209 97 (46) 79 (38) 47 (22) 42 (20) 26 (12) 25 (12) Noncomprehensive 617 365 (59) 301 (49) 207 (34) 191 (31) 143 (23) 136 (22) Comprehensivet 154 38 (25) 36 (23) 6 (4) 6 (4) 4 (3) 4 (3) P§ ... 00001 <.0001 <.0001 <.0001 <.0001 <.0001 *Includes both laws and administrative regulations.Relationships do not change when regulations are excluded. tRestricts smoking in more than three public places,restaurants,and private-sector workplaces. tx'test of trend. W tests,unless otherwise marked. IlMajor tobacco-producing states are Georgia,Kentucky,North Carolina,South Carolina,Tennessee,and Virginia. IlRsher's Exact Test. #A minimal law does not restrict smoking in restaurants,workplaces,or more than three public places.A noncomprehensive law restricts smoking in more than three public places but not in workplaces. prevalent in Northeastern states,while Table 3.-Prevalence of Local No-Smoking Laws by Population,Region,Tobacco Production,State Law, Southern states had the fewest laws of and Smoking Prevalence* all types.Major tobacco-producing states Any Local No-Smoking Law, Comprehensive Law,t also had few laws. Since all six major Odds Ratio(95% Odds Ratio(95% tobacco states are in the South,we strat- Confidence Interval) Confidence Interval) ified the data to separate the effects of Without Smoking Prevalence region and tobacco production.The anal- Population$ 1.90 (1:62-2.24) 1.86 (1.53-2.26) ysis demonstrates that both region and GtSouth c region§ 0.48 (0.30-0.74) 0.14 (0.08-0.26) tobacco production affect the prevalence Midwest 0.32 (0.22-0.48) 0.16 (0.09-0.26) of state laws: Southern states not pro- Northeast 0.32 (0.21-0.50) 0.09 (0.04-0.19) ducing tobacco have fewer laws than Major tobacco-producing statell 0.39 (0.20-0.78) 0.20 (0.05-0.77) states in other regions, but they have State no-smoking lavA more laws than their tobacco-producing None/minimal 6.62 (3.48-12.60) 8.27 (2.61-26.18) Southern neighbors.We also examined Noncomprehensive 6.01 (3.51-10.30) 5.57 (1.95-15.89) the relationship between state no-smok- Wlth smoking Prevalence ing laws and adult smoking prevalence. Population# 1.91 (1.62-2.25) 1.88 (1.54-2.29) Although the 1989 adult smoking prev- Geographic region§ South 0.83 (0.47-1.47) 0.23 (0.11-0.48) alence was lower in states with laws Midwest 0.53 (0.32-0.86) 0.24 (0.13-0.45) than in states without laws, smoking Northeast 0.49 (0.29-0.81) 0.12 (0.05-0.28) prevalence was not independently Major tobacco-producing statell 0.40 (0.29-0.81) 0.20 (0.05-0.77) sociated with any type of state law wheenn State no-smoking lavA geographic region or tobacco produc- No/minimal 8.98 (4.56-17.68) 2.30 (3.62-41.74) tion was included in a multiple logistic Noncomprehensive 7.19 (4.12-12.57) 7.15 (2.41-21.18) regression analysis (data not shown). 1989 State smoking prevalence 0.90 (0.84-0.96) 0.91 (0.83-0.99) COMMENT *Multiple logistic regression analysis. This study demonstrates that state and tL&v restricting smoking in more than three public places,restaurants,and private-sector workplaces. #Coded as four categories:25 000 to 50 000,50 000 to 100 000,100 000 to 250 000,and more than 250 000. city laws limiting smoking in public places §Dummy variables,with the West as the reference category. and workplaces in the United States are IlMajor tobacco-producing states are Georgia,Kentucky,North Carolina,South Carolina,Tennessee,and Virginia. 11Dummy variables,with comprehensive state law as the reference category.A minimal law does not restrict more widespread than previously appre- smoking in restaurants,work sites,or more than three public places.A noncomprehensive law restricts smoking in ciated,particularly at the local level.The more than three public places but not in work sites. JAMA,December 11, 1991-Vol 266,No.22 No-Smoking Laws-Rigotti&Pashos 3165 1 c Table 4.—State No-Smoking Laws:Prevalence and Variability State No-Smoking laws,No.(9+) No.of Any Government >3 Public Private Comprehensive States law* Buildings Places Restaurants Workplaces Lawt All states$ 51 44 (86) 31 (61) 35 (69) 22 (43) 12 (24) 10 (20) Geographic region Northeast 9 9 (100) 8 (89) 9 (100) 7 (76) 7 (78) 5 (56) West 13 12 (92) 11 (85) 11 (85) 7 (54) 2 (15) 2 (15) Midwest 12 11 (92) 9 (75) 10 (83) 6 (50) 2 (17) 2 (17) South 17 12 (71) 3 (18) 5 (29) 2 (12) 1 (6) 1 (6) P§ .1916 .0001 .0002 .0052 .0006 .0312 Major tobacco-producing statell Yes 6 3 (50) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) No South 11 9 (82) 3 (27) 5 (45) 2 (18) 1 (9) 1 (9) Other 34 32 (94) 28 (82) 30 (88) 20 (59) 11 (32) 9 (26) P§ ... .0130 <0001 <.0001 .0038 .1299 .3053 *Law restricting smoking in at least one public place. tRestricts smoking in more than three public places,restaurants,and private-sector workplaces. #Includes the District of Columbia. §Ail significance tests are Fisher's exact tests or extensions for a 2 x 4 table. ((Major tobacco-producing states are Georgia,Kentucky,North Carolina,South Carolina,Tennessee,and Virginia.Because these are all Southern states,the nontobacco states are stratified to permit comparison of Southern states with and without major tobacco production. analysis indicates that legislation devel- oped at the state level in the 1970s and 50 moved to cities a decade later, where it spread rapidly. By mid-1989, all but six states and over half of US cities had taken 40 some action,and nearly all urban Amer- icans were covered by a state or local smoking restriction.Despite this impres- 30 - sive level of action on smoking,broad re- strictions on smoking remain relatively a uncommon. By 1989, only 20%of states z 20 -and 17% of cities had adopted compre- hensive no-smoking ordinances--those most likely to provide meaningful protec- 10 - - tion from ETS exposure. From a public health perspective,further state and lo- 0 cal action is clearly indicated.States and 19701971197219731974197519761977197819791980198119821983198419851986198719881989 communities need to adopt comprehen- sive no-smoking laws or amend current legislation to broaden coverage.Action is Q 1 to 3 Public Places ® >3 Public Places especially needed in smaller cities and in ® Restaurants and Public Places = Private Work Sites tobacco-producing states and non-West- ern states. We identified more local laws than Fig 2.—Cumulative number of state no-smoking laws in the United States from 1970 through 1989. previous reports1"I for several reasons. First,our method permitted a near-com- Our method appears to be a feasible While both state and local laws pro- plete ascertainment of laws from a de- approach to monitoringthe trend of state vide protection from ETS exposure, fined sample of cities, providing a bet- and local smoking restrictions nation- their content differs. The comprehen- ter estimate of the prevalence of smok- wide. Because the sample is limited to siveness of most state laws is limited.A ing restrictions in US cities than pre- cities with a population of 25 000 or greater proportion of city laws are com- vious work. Second, we included both greater,we do not know the prevalence prehensive. Our data also indicate that administrative regulations and legislated of smoking restrictions in smaller com- city and state no-smoking laws are re- ordinances,both of which have the force munities or in counties; therefore, our lated phenomena. City no-smoking or- of law. Our goal was to note the full count underestimates the total number dinances are less common in states with extent of local government action in- of city, town, and county no-smoking comprehensive no-smoking laws, pre- tended to reduce ETS exposure, even laws in the United States.According to sumably because these cities have less though some actions were limited in our analysis of laws cited by other need for action. However, there are scope. Others have excluded adminis- sources,10'" approximately half were more city ordinances in states with non- trative regulations or laws covering only passed by counties(10%to 15%)or small comprehensive state laws than in states government buildings.10 We suggest that towns not in our sample(35%). Adding with no state law. This is understand- these limited actions provide some ETS these laws to our results, we estimate able if state and city laws are not passed protection and, more important, can that by 1989 over 700 cities,towns,and independently.States that have passed serve as a community's start toward counties in the United States had a noncomprehensive state law probably more comprehensive action. adopted smoking restrictions. have antismoking coalitions that are also 3166 JAMA,December 11, 1991—Vol 266,No.22 No-Smoking Laws—Rigotti 8 Pashos active at the local level. States with no guage of no-smoking laws is heteroge- and New York Times. December 24, state law may have weaker antismoking neous,and our categories cannot incor- 1988:1).'Although the industry has less forces or face greater tobacco industry porate the full variability of these stat- successfully opposed local than state or opposition at both the state and local utes. Even within categories, the de- federal laws,it has begun to use its eco- levels.This is supported by our finding gree of smoking restriction required by nomic power to mount more vigorous that,as expected,there are fewer state law varies-particularly for workplace opposition to proposed legislation and and city laws in states where tobacco is laws. In many cases, more strongly to sponsor referenda to overturn laws a major crop. Local laws and compre- worded legislation could improve the that pass(so far unsuccessfully).�21,21The hensive state laws are also more com- level of ETS protection.Finally,we did industry has organized and funded mon in the West, probably-reflecting not obtain a copy of 109 laws claimed by "smokers'rights"groups to oppose new the earlier start and greater strength of both ANR and TFA. Relying on their legislation (Wall Street Journal. April nonsmokers'rights activism in that re- categorizations might have introduced 11, 1988:29).122 It has also attempted to gion.2 misclassification errors, although our prevent or override local no-smoking or- Our analysis found support for the analysis of 40 laws for which we had dinances by supporting weak state laws hypothesis that no-smoking laws are both hard copies and ANR and TFA that specifically preempt stronger local more easily adopted where there are categorizations did not detect consistent action.By 1989,four states had adopted fewer smokers. Local laws were more errors that would introduce a system- these laws(Florida,Pennsylvania,Okla- common in states with a lower smoking atic bias. homa, and Virginia)and bills had been prevalence in 1989. An alternative hy- Important questions remain concern- filed in others (Governing. May 1989: pothesis, that no-smoking ordinances ing the progress and impact of no-smok- 34).2 A related tactic is to promote state contribute to reducing smoking preva- ing laws. Whether these laws actually "smokers'rights"legislation that seeks lence, cannot be tested with these data achieve their intent of reducing ETS to protect smokers from discrimination because they contain no information on exposure and whether they also reduce in the workplace and elsewhere.By May smoking prevalence after 1989.Two fac- smoking behavior is not yet known be- 1991, 13 states had passed these laws tors tend to blunt any association of laws cause individual laws have rarely been and they had been introduced in 12 others and smoking behavior.First,the range evaluated.2 Anecdotal reports and one (New York Times.February 19,1991:A1 of state smoking prevalences is narrow, systematic study have demonstrated and May 20,1991:B1).Whether these ac- limiting the power of the analysis. Sec- that comprehensive laws are well ac- tions will slow the momentum of efforts ond,city-specific smoking prevalence is cepted,feasible for health departments to protect the public from ETS exposure not available for a national sample. We to implement,and achieve a reasonable bears close watching and demonstrates used state smoking prevalence as a proxy level of compliance.1,11,11 Whether the the need for regular monitoring of state for city prevalence, which essentially rapid pace of state and local no-smoking and local no-smoking legislation. eliminated within-state variability.Any legislation continues is another impor- resulting misclassification, however, tant question.We report that fewer lo- This study was supported by grant 1-1103- would tend to reduce any relationship cal laws passed in 1989 than in previous CA48386-01 from the National Cancer Institute, between smoking behavior and no-smok- years, but our data are complete only Bethesda, Md. Dr Rigotti is a Teaching and Re- ing laws. through July 1, 1989. This issue merits search Scholar of the American College of Physi- There are several limitations to this close monitoring, because gaps in non- dans. g p We thank Tobacco-Free America (especially study. First, it cannot determine the smoker protection are considerable,and Angela T.Mickel)and Americans for Nonsmokers' degree of compliance with written stat- the proliferation of smoking-control laws Rights for sharing information;Kathy McNiff for utes.If compliance is not good,our data is being more actively opposed by the programming; MDn ;aid Singe r,D Pand,Thomas a a chap- will overestimate the extent of actual tobacco industry(Tobacco Observer.Oc- man Walsh, PhD,for advice in study design and smoking restriction. Second, the lan- tober 1988;13:1;Governing.May 1989:34; review of the manuscript. References 1. The Health Consequences of Involuntary Smok- 8. Americans for Nonsmokers'Rights.Matrix of lar Disease Risk Factors.Bethesda,Md:National ing:A Report of the Surgeon General.Atlanta,Ga: Local Smoking Ordinances.Berkeley,Calif:Amer- Heart,Lung,and Blood Institute;1991.(A sepa- Centers for Disease Control; 1986. US Dept of icans for Nonsmokers'Rights;1988. rate fact sheet for each state contains state smok- Health and Human Services publication CDC 87- 9. Tobacco-Free America.State Legislated Actions: ing prevalence data from the 1989 current popu- 8398. Limitations on Smoking in Public Places.Wash- lation survey.) 2. Reducing the Health Consequences of Smoking: ington, DC: Tobacco-Free America Legislative 16. Rosner B.Fundamentals of Biostatistics.2nd 25 Years of Progress: A Report of the Surgeon Clearinghouse;1988. ed.Boston,Mass:Duxbury Press;1986. Geneml.Atlanta,Ga:Centers for Disease Control; 10. Pertschuk M,Shopland DR.Major Local Smok- 17. Mehta CR,Patel NR.A network algorithm for 1989. US Department of Health and Human Ser- ing Ordinances in the United States. Bethesda, performing Fisher's Exact Test in r x c contingency vices publication CDC 894411. Md:Americans for Nonsmokers'Rights and Na- tables.J Am Stat Assoc. 1983;78:427434. 3. Rigotti NA.Trends in the adoption of smoking tional Cancer Institute; 1989. US Department of 18. Martin MJ.The San Francisco experience with restrictions in public places and worksites. N Y Health and Human Services publication NIH 90- regulation of smoking in the workplace:the first 12 State J Med. 1989;89:19-26. 479. months.Am J Public Health. 1986;76:585-586. 4. Repace JL, Lowrey AH. A quantitative esti- 11. Tobacco-Free America. State Legislated Ac- 19. Rigotti NA, Bourne D, Rosen A, Locke J, mate of nonsmokers'lung cancer risk from passive tions on Tobacco Issues.Washington,DC:Tobacco- Schelling TC.Worksite compliance with a no-smok- smoking.Environ Int. 1985;11:3-22. Free America Legislative Clearinghouse;1989.At- ing law:a randomized community intervention trial. 5. Sorensen G, Rigotti NA, Rosen A, Pinney J, tachment E. Am J Public Health.In press. Prible R.Effects of a worksite nonsmoking policy: 12. County and City Data Book,1988.Washing- 20. Intimidation in Albuquerque.Washington,DC: evidence for increased cessation. Am J Public ton,DC:US Bureau of the Census;1983. Smoking Control Advocacy Resource Center,Ad- Health. 1991;81:202-204. 13. City Employment in 1986.Washington, DC: vocacy Institute;1989. 6. Stillman FA, Becker DM, Swank RT, et al. US Bureau of the Census;1988.Series GE-86,No.2. 21. Hanauer P.Proposition P:anatomy of a non- Ending smoking at The Johns Hopkins Medical 14. Grise VN,Griffin KF. The U.S. Tobacco In- smokers' rights ordinance. N Y State J Med. Institutions:an evaluation of smoking prevalence dustry.Washington,DC:Economic Research Ser- 1985;85:369-374. and indoor air pollution.JAMA.1990;264:1565-1569. vice,US Department of Agriculture;1988:16.Eco- 22. Industry Creates`Smokers'Rights'Campaign. 7. Samuels B,Glantz SA.The politics of local to- nomic Report 589. Washington,DC:Smoking Control Advocacy Re- bacco control.JAMA. 1991;266:2110-2117. 15. Data Fact Sheet:United States:Cardiovascu- source Center,Advocacy Institute;1991. JAMA,December 11, 1991-Vol 266, No.22 No-Smoking Laws-Rigotti&Pashos 3167 The Effects of Combining Education and Enforcement to Reduce Tobacco Sales to Minors A Study of Four Northern California Communities Ellen Feighery, MS; David G. Altman, PhD; Gregory Shaffer, MA Objective.—To examine the effects of a community education and law en- olds, 9.5%in 14-and 15-year-olds, and forcement intervention on illegal tobacco sales to minors. 19.1%in 16-and 17-year-olds. The rate Design.—A 2-year, before and after trial with retail stores as the unit of anal- for 16- and 17-year-olds is just 2% less ysis. than that of the overall California adult Setting.—Implementation occurred in four suburban California communities population.This study also identified ex- with populations of 25 000 to 100 000. perimenters: 12.9% of 12- and 13-year- Participants.—All the retail stores in one intervention community and half the olds, 29.4%of 14-and 15-year-olds, and 40.8%of 16-and 17-year-olds.Nationally, retail stores, randomly selected, in the other three intervention communities in 1988, more than 3 million Americans (n=169)were visited by minors aged 14 to 16 years with the intent to purchase under the age of 18 years consumed al- tobacco. most 1 billion packs of cigarettes and 26 Intervention.—Ongoing community and merchant education and four law million containers of smokeless tobacco, enforcement operations were conducted. accounting for approximately 3% of an- Main Outcome Measures.—Over-the-counter and vending machine sales of nual tobacco industry profits.I These data tobacco to minors were the primary outcomes. illustrate that minors are obtaining to- Results.—Among a cohort of stores visited by minors at the pretest(n=104) bacco readily. in June 1988,71%sold tobacco over the counter and 92%sold tobacco through vending machines.At posttest 2 in May 1990,24%sold tobacco over the counter See also pp 3159 and 3186. and 93%sold tobacco through vending machines. Of the 31 stores issued ci- tations, 16 were followed into the courts where the fines were dismissed or re- Indeed,in field trials around the coun- duced. try,minors have purchased tobacco suc- Conclusions.—Education alone had a limited effect on reducing illegal cessfully from stores and vending ma- tobacco sales to minors. It did promote community support for more aggressive chines 70%to 100%of the time (E.F., enforcement strategies. Education plus enforcement decreased significantly unpublished data,1991)."'In May 1990, over-the-counter sales;vending machine sales were unaffected by these inter- Louts Sullivan, MD, Secretary of the ventions.The lack of support at the judicial level may temper the effectiveness Department of Health and Human Ser- of enforcement. Legislative remedies addressing judicial obstacles and vend- vices, wrote: "Access of minors to to- bacco is a major problem in every state ing machine sales are needed. of the nation. About three-fourths of (JAMA. 1991;266:3168 3171) the million outlets which sell cigarettes to adults also sell cigarettes to minors. EASY access to tobacco products by smokers become addicted to tobacco by These stores ignore the laws of their adolescents is a major public health prob- the age of 18 years,generally before it states because enforcement is almost lem. Approximately 75% of current is legal for them to purchase tobacco nonexistent.""The Inspector General's products.' The importance of prevent- Office of the Department of Health and ing early tobacco use is reinforced by Human Services found blatant disregard From the center for Research in Disease Prevention, data illustrating that tobacco is the ini- of the laws that prohibit the sale of to- Stanford University School of Medicine,Palo Alto,Calif Health tial drug of reference for young people bacco to minors b merchants law en- (Ms Feighery and Dr Altman)and the North Bay Health g P Y g p p Y Resources Center,Petaluma,Calif(Mr Shaffer). and that its use is associated with other forcement agencies,and communities at Reprint requests to the Center for Research in Dis- drug use.1-1 A 1990 report of tobacco use large. Specifically, they found only 11 ease Prevention,Stand,Palo university 303( of Medi- in Californias found that current smok- active state and local jurisdictions where Medi- cine,1000 Welch Rd,Palo Alto,CA 94303(Ms Feigh- � ery). ing status was 3.3%in 12-and 13-year- these laws were enforced.12 In one of 3168 JAMA,December 11,1991—Vol 266,No.22 Education and Enforcement to Reduce Tobacco Sales—Feighery et al these jurisdictions, Woodridge, Ill, il- generated from the telephone company names of violators and stores. legal sales to minors in the 26 stores yellow pages and the county health de- In May 1990,following police enforce- licensed to sell tobacco were eliminated partment's listing of eating establish- ment activities and continued education, due to a tobacco retailer licensing ordi- ments.The outlets included grocery,li- 15 male and female minors aged 14 to 16 nance and active law enforcement.This quor and convenience stores, restau- years who had not participated previ- impressive finding should be interpreted rants, pharmacies, and gas stations. ously in the project visited 145 stores in in light of the fact that in the four vil- In September 1988,a comprehensive the four cities with the intent to pur- lages contiguous to Woodridge,tobacco educational intervention directed at mer- chase tobacco(posttest 2). Of the orig- sales to minors occurred in 94% of the chants, law enforcement agencies, and inal 169 outlets visited In June through stores. Thus,tobacco is still readily ac- the community at large was begun.The August 1988,104 were revisited in May cessible to Woodridge minors.The cur- intervention included widely publiciz- 1990. The 65 stores not visited in May rent study evaluates an intervention, ing the results of the pretest through -1990 either went out of business,did not also cited in the Inspector General's re- the local media; making presentations sell tobacco products, or could not be port,that combined merchant and com- to city councils,the county board of su- located by project staff members. At munity education with active enforce- pervisors,and community organizations; posttest 2, 41 of the 145 stores visited ment of the California law by local police and mailing educational packets to all had not been visited previously.To ex- departments. The goal was-to reduce tobacco retailers in the four cities. The amine the representatives of the 104 tobacco sales to minors by 50%. packets included a cover letter that de- stores that were visited at the pretest Laws regulating sales of tobacco prod- scribed the results of the first survey and posttest 2, two analyses were con- ucts to minors differ from state to state. and why it was important to comply ducted.The first,an analysis of pretest California law prohibits the sale and pur- with the law,a copy of the Iaw,warning data,compared stores with pretest data chase of tobacco products to and by any- stickers for cash registers,employee ed- only(n=65)to stores with pretest and one under the age of 18 years.Retailers ucation materials, and a list of individ- posttest 2 data(n=104).The second,an who break this law are subject to a mis- uals and community organizations sup- analysis of posttest 2 data, compared demeanor with a first-offense fine of porting the project. stores with posttest 2 data only(n=41) $200,a second-offense fine of$500, and In December 1988, half of the stores to stores with pretest and posttest 2 a third-offense fine of$1000.As of Jan- visited at the pretest were selected ran- data(n=104).Neither of these analyses uary 1, 1989,minors caught purchasing domly and visited by eight of the orig- was significant,suggesting that the 104 t6bacco are subject to a fine of$50 or 25 inal 20 minors(posttest 1).A total of 83 stores visited at the pretest and post- hours of community service work. stores were visited.Because the results test 2 were representative of the entire METHODS of this visit fell short of project expec- sample of 210 different stores visited tations to reduce sales by at least 50%, over the course of the intervention(169 In 1988, the Solano County Cancer face-to-face interviews were conducted at the pretest, 41-at posttest 2). Prevention Program embarked on a com- with 17 merchants to discover why the Because the California law was not munity-wide effort to reduce the illegal educational effort did not achieve the enforced prior to the intervention, we sale of tobacco to minors. The four tar- results of a similar study in Santa Clara wanted to document how judges inter- geted cities in Solano County,California County, California.5,13 Merchants re- preted it. In the summer of 1990, staff (Benicia,Fairfield,Vacaville,and Vallejo), ported the following:(1)frustration that members followed half of the store clerks with populations ranging from 25 000 to minors could purchase tobacco products cited for selling tobacco to minors 100000, account for approximately 85% easily from other local sources,making through the county court system and of the county's 340000 residents. These it a disincentive to change their prac- interviewed each judge after court ses- are suburban communities separated tices; (2) knowledge that the law was sions to obtain explanations of their from each other by 8 to 24 km. not enforced and a belief that without decisions. Tobacco sales to minors (yes or no) the sanctions of active enforcement,. RESULTS was the primary outcome variable of business as usual was acceptable; and the study. Data were collected at the (3)belief that the most effective method Overall, 73% (n=169) of stores sold pretest (June through August 1988); to stop sales to minors was by active en- tobacco to minors at the pretest, 68% posttest 1, after an education-only in- forcement. The results of the store sur- (n=83) sold tobacco at posttest 1, and tervention (December 1988); and post- veys and the merchant interviews were 31%(n=145)sold tobacco at posttest 2. test 2, after an education plus law en- communicated to local police departments Over-the-counter sales dropped from forcement intervention(May 1990).Data with requests to enforce the law. 72% at the pretest (n=144), to 62% were analyzed using the McNemar non- In November 1989,a law enforcement (n=69)at posttest 1,and to 21%(n=122) parametric test. intervention was added to the ongoing at posttest 2. Vending machine sales In June through August 1988 (pre- educational intervention.Four police de- were 84%at the pretest(n=25),93%at test),20 youths ranging in age from 14 partment enforcement operations (ie, posttest 1 (n=14), and 83% (n=23) at to 16 years were recruited through local "stings") were conducted by three po- posttest 2. Table 1 presents data from community agency contacts and escorted lice departments. Upcoming police vis- the cohort sample of stores visited at to 169 stores in the four intervention its to stores were announced in local both the pretest and posttest 1,and Ta- cities to purchase tobacco. The 14- to newspapers. A total of 90 stores were ble 2 presents data from the cohort of 16-year-old age group was selected be- visited by underage police cadets;34% stores visited at both the pretest and cause this is generally when experimen- of the stores sold tobacco products and posttest 2. tation and adoption of smoking behavior received citations.Each enforcement ef- Following the issuance of citations, occurs. The stores comprised approxi- fort required about 8 hours of each po- 16 of the 31 merchants who received mately half of all the retail outlets in lice department's time,half of which was citations were followed through the ju- each city with the exception of Benicia spent visiting the stores and the dicial system by Solano County Cancer where all tobacco retailers were sur- other half on paperwork. The results Prevention Program staff members to veyed. The outlets in the other three of this police activity were reported track the disposition of the cases. In cities were selected randomly from lists in the local media, including the seven cases, the judges suspended the JAMA, December 11, 1991—Vol 266,No.22 Education and Enforcement to Reduce Tobacco Sales—Feighery et ai 3169 Table 1.—Cohort of Stores Visited at Pretest and Posttest 1 to Determine Percentage of Stores Selling County study, one of which was San Tobacco to Minors Jose,the 11th largest city in the 11nited %Change Pretest States, were primarily urban. Second, June 1988 December 1988 to Posttest 1t while we used local print and radio me- Pretest, Posttest 1, (95%Confidence dia extensively, we were generally un- Variable No.of stores(%) No.of stores(%) P. Interval,%)# able to capture the attention of the larger All stores 77 (77) 77 (65) NS —12 (1,-24) San Francisco Bay area television me- Over-the-counter sales sa (73) 64 (59) NS —14 (0,-28) dia market as was accomplished in the Vending machines sales 13 (92) 13 (92) NS —0 (21,-21) Santa Clara study.We did,however,get *NS indicates a significance level>.05. some television coverage when police de- tPerceni in pretest—percent in posttest 2. partments started issuing citations. $95%confidence interval calculated for change using sample with data at pretest and posttest 1. There are several disappointing out- comes of the intervention to date.First, Table 2.—Cohort of Stores Visited at Pretest and Posttest 2 to Determine Percentage of Stores Selling vending machine sales were unaffected. Tobacco to Minors This provides support for the elimina- %Change Pretest tion of all tobacco vending machines,an June 1988 May 1990 to Posttest 2t action called for by Department of Pretest Posttest 2 (95%Confidence Health and Human Services Secretary Variable No.of stores(%)* No.of stores(%) Pt Interval,%)§ Louis Sullivan and several prohealth All stores 104 (75) 104 (35) <.0001 —40 (-26,-55) groups. Community experiences with w Over-the-counter sales 87 (71) 87 (24) ool -47 (-31,-64) Vending machines sales 14 (93) 14 (93) NS 0 partial vending machine bans (eg, vol- (20,-20) untary surveillance of machines by re- *The numbers for over-the-counter sales(87)and vending machine sales(14)do not add up to the number for tapers, the use of locking devices or to- all stores(104)because three stores were dropped from the analysis since tobacco was purchased differently kens, or limiting machines to adult loca- between tests leg,one time by over-the-counter and one time by vending machine). tNS indicates a significance level>.05. tions)in limiting vending machine sales of #Percent in pretest—percent in posttest 2. tobacco to minors are not encouraging. §95%confidence interval calculated for change using sample with data at pretest and posttest 2. The second outcome was the judges' dismissals of charges and reductions of sentences, placed those cited on infor- nity and merchant education alone. fines when store clerks who received mal probation,and imposed$50 fines as Given the small amount of time spent by citations appeared in court. This was a condition of probation.One judge sus- each police department to implement particularly disappointing in light of the pended the sentences of all nine persons the intervention, our findings suggest broad community support and police issued citations and although they were that enforcement is a feasible way to commitment to enforcement.Continued found guilty, the individuals were not reduce tobacco sales to minors. lack of punishment would eventually act required to pay a fine. While education alone yielded a rel- as a deterrent to police involvement and Interviews were conducted with two atively small reduction in illegal sales of would weaken the effectiveness of po- judges to learn why these decisions were tobacco products to minors, it did gar- lice enforcement efforts in the commu- made. Three primary reasons were ner broad community support and set nity. Additionally, merchants would given. First,the judges were reluctant the stage for the more punitive action of again flaunt the law knowing that cita- to establish criminal records for citizens police department enforcement. Edu- tions would not be upheld in court. Fol- with no prior criminal history. Since it cating merchants and the community at lowing this study, however, when new is a misdemeanor in California to sell large about the law demonstrated to the citations were processed through the tobacco to a minor, convicted violators police departments that a serious effort same courts, the judges were far less will have lifetime criminal records.Sec- was made to obtain voluntary compli- lenient than they were when first con- ond,the judges believed that a$200 fine ance with the law. When educational fronted with this type of case.This sug- placed an unfair burden on store em- efforts fell short of project goals,police gests that the court system can be stim ployees,many of whom worked for min- departments were willing to take ac- ulated to respond when repetitive vio- imum wage. California law states that tion.Additionally,because police depart- lations occur. the clerk who sold tobacco illegally, ments are public servants and therefore Third, although it is clear that mer- rather than the manager or store owner, sensitive to community pressure, pre- chant behavior can be affected signifi- is cited. Third,judges believed that vi- senting evidence of the problem and sup- cantly by active enforcement of the law olations of the law on tobacco sales to port from the local media and commu- by police departments, multiple prob- minors were far less serious than most nity leaders influenced their decisions lems with access laws exist. Given that of the other cases they face daily. As a to enforce a law that is largely ignored the dockets of criminal courts are gen- result,judges treated violators leniently, throughout the state and nation. erally overloaded, these courts are in- particularly first-time offenders.Unlike Since this project modeled its educa- appropriate vehicles for the disposition the California law regulating sales of tional intervention on the Santa Clara of access violations.Also,judges'reluc- alcohol to minors, judges interpreted County project,"we expected to achieve tance to establish criminal records for the tobacco sales law as providing them similar reductions in over-the-counter citizens with no prior convictions must with discretion in determining the pen- sales due to education alone. In retro- be recognized. Both of these problems alties, if any, imposed. spect,there were two major differences lend support for access laws to be con- in program implementation that may ac- sidered as civil rather than criminal of- COMMENT count for our educational intervention fenses and processed administratively This study illustrated that enforce- not reducing tobacco sales significantly. rather than through the criminal justice ment of laws regulating sales of tobacco The first difference is the type of com- system. products to minors had a significant ef- munity. The four intervention cities in Fourth,the responsibility for enforce- fect on over-the-counter sales above and Solano County are small and suburban, ment has implications for the success of beyond that obtained through commu- whereas the cities in the Santa Clara access laws. To increase enforcement, 3170 JAMA,December 11, 1991—Vol 266,No.22 Education and Enforcement to Reduce Tobacco Sales—Feighery et al authority should rest with both law en- this study should be considered in light ing use of tobacco or are already in the forcement agencies and with local or of several potential design limitations. early stages of use.Unfortunately,there state government agencies such as pub- First,the design did not allow a test of are virtually no data on the relationship he health departments. Secretary Sul- the independent effects of education and between reductions in tobacco access livan and other tobacco control activists enforcement.Thus,it is unclear whether and youth smoking prevalence. In the propose the establishment of a licensing enforcement alone would have achieved case of alcohol,however,there is a mod- system that would fund enforcement and the same outcomes.Second,a truer cross erate amount of literature on the effects tie compliance with laws to selling to- section of merchants would have been of raising the minimum age of purchase, bacco products. obtained if sampling occurred on sev- alcohol availability, and prohibition of Fifth,there is the issue of who should eral different days and times.Third,the consumption."" Although this study be culpable and for what. Some have absence of data from some stores at the demonstrated that active enforcement argued that the owner or manager of a three data collection points may limit 'of sales-to-minors laws is an effective store selling tobacco products should be the conclusions derived, although sta- and viable way to reduce sales of to- fined for not setting and enforcing pol- tistical analysis of potential differences bacco products to minors,the most fun- icies. The establishment of a licensing was not significant. damental question has yet to be an- system would address this issue because The results of this study illustrate the swered-what effect does decreased ac- stores would be punished for violations, complexity of the problem regarding ac- cess by underage youth have on their including the loss of their license to sell cess to tobacco products by underage use of tobacco? any tobacco.Others have suggested that youth. An over-the-counter sales rate minors should be held accountable by of 21%still provides minors with access making possession and use of tobacco to tobacco, although it is possible that This project was sponsored by the Solano County illegal.The effects of these different op- this level of sales may serve as an ef- Cancer Prevention Program of the North Bay impediment to minors who are Health Resources center, Petaluma, Calif, and tions need further evaluation. fective im Iewas supported by a grant from the Henry J.Kai- The interpretation of the findings of not yet addicted or who are contemplat- ser Family Foundation,Menlo Park,Calif. References 1. Surgeon General. The Health Consequences of 7. DiFranza JR,Tye JB.Who profits from tobacco Health. 1991;81:891-893. Smoking:Nicotine Addiction:A Report of the Sur- sales to children?JAMA. 1990;263:2784-2787. 14. Ashley MJ, Ranldn JG. A public health ap- geon General.Washington,DC:US Dept of Health 8. Kirn TF.Laws ban minors'tobacco purchases, proach to the prevention of alcohol-related health and Human Services;1988.US Dept of Health and but enforcement is another matter. JAMA. problems. Annu Rev Public Health. 1988;9:233- Human Services publication 88-8406. 1987;257:3323-3324. 271. 2. Fleming R, Levanthal H,Glynn K,Ershler J. 9. Altman D, Foster V, Rasenick-Douss L, Tye 15. US Department of Health and Human Ser- The role of cigarettes in the initiation and progres- JB.Reducing the illegal sale of tobacco to minors. vices.Surgeon General's Workshop on Drunk Driv- sion of early substance abuse. Addict Behay. JAMA. 1989;261:80-83. ing: Proceedings. Washington, DC: US Dept of 1989;14:261-272. 10. DiFranza JR,Norwood BD,Garner DW,Tye Health and Human Services;1989. 3. Greydanus DE.Routing the modern Pied Piper JB. Legislative efforts to protect children from 16. Holder HD. Environmental restrictions and of Hamelin.JAMA. 1989;261:99-100. tobacco.JAMA. 1987;257:3387-3389. effective prevention policy. Adv Subst Abuse. 4. Yamaguchi K,Kandel DB.Patterns of drug use 11. Model Sale of Tobacco Products to Minors Con- 1987;1(suppl 1):405-432. from adolescence to young adulthood,II:sequences trol Act:AModel law Recommended for Adoption 17. Holder HD,Saltz RF.Research opportunities of progression.Am J Public Health. 1984;74:668- by States or Localities to Prevent the Sale of To- in environmental and community prevention strat- 672. bacco Products to Minors. Washington, DC: US egies.Presented at the Institute of Medicine Panel 5. Yamaguchi K,Kandel DB.Patterns of drug use Department of Health and Human Services;May on Opportunities for Research on Prevention of from adolescence to young adulthood,III:predic- 24, 1990. Alcohol-Related Problems;January 20,1988;Wash- tors of progression. Am J Public Health. 12. Youth Access to Cigarettes:A Report of the ington,DC. 1984;74:673-681. Offwe of the Inspector General, New York. New 18. Milgram GG,Nathan PE. Efforts to prevent 6. University of California(San Diego),California York,NY:Office of the Inspector General;1990. alcohol abuse.In:Edelstein BA,Michelson L,eds. Department of Health Services. Tobacco Use in 13. Altman DG,Rasenick-Douss L,Foster V,Tye Handbook of Prevention.New York,NY:Plenum California, 1990. Sacramento,Calif:Department JB.Sustained effects of an educational program to Press;1986:243-262. of Health Services;1990. reduce sales of cigarettes to minors.Am J Public JAMA,December 11, 1991-Vol 266,No.22 Education and Enforcement to Reduce Tobacco Sales-Feighery et al 3171 Special Communications Clinical Interventions in Tobacco Control A National Cancer Institute Training Program for Physicians Marc Manley, MD; Roselyn P. Epps, MD; Corinne Husten, MD; Thomas Glynn, PhD; Donald Shopland SMOKING is the leading preventable tients who smoke.1,6-9 The trials,involv- • Patients of trained physicians who cause of death in our country. Smoking ing more than 1000 physicians and 30 000 received routine reminders were up to kills 434 000 people a year, more than patients,were conducted in various med- six times more likely to stop smoking 1000 every day'; it accounts for about ical care provision settings across North than patients of control physicians. 85%of all lung cancer deaths,about 80% America(eg,private offices,public clin- SUMMARY OF NCI of all chronic obstructive pulmonary dis- ics, health maintenance organizations, RECOMMENDATIONS ease deaths,and 30%of all heart disease and residency programs).Although con- deaths. In addition, smoking costs this ducted in nonresearch environments, It is important for both the physician country $52 billion annually in health these were randomized,controlled trials. and the patient who smokes to realize care and other costs.' In all but one trial,smoking cessation that smoking cessation is not a single rates were calculated for a physician's event;rather,it is a process that takes See also pp 3139 and 3183. entire practice, not for individual pa- place over time. A person who smokes tients. Thus, all smokers in each phy- moves through stages from being unin- sician's practice were followed,not just terested in stopping, to thinking about In spite of the magnitude of this health those who received interventions from change, to making a concerted effort hazard,most physicians have never re- a physician. Long-term cessation was to stop, to finally maintaining ceived training in techniques to help pa- confirmed biochemically. Even with abstinence."" Most people make sev- tients stop smoking. Many physicians these very strict evaluation criteria,ces- eral attempts before succeeding com- believe they are unprepared and unsuc- sation rates of up to 15% were docu- pletely, and relapse is a normal part of cessful in treating patients addicted to mented in the trials'; ie, up to 15% of the smoking cessation process. nicotine.1,4 However, recent scientific patients who smoked were verified as Since stopping smoking is a process, evidence suggests that physicians can smoke-free at the end of approximately the physician and office staff must pro- help smokers stop and thus reduce the 1 year. Certain groups of patients had vide a repeated,consistent message and incidence of smoking-related diseases. even higher cessation rates. In one offer appropriate treatment at every In 1989, the National Cancer Insti- study,patients who agreed to stop smok- visit.The treatment of smokers is anal- tute(NCI)used clinical trial results and ing and returned for several follow-up ogous to the management of patients consensus development to produce ree- visits had long-term cessation rates of with mild or moderate hypertension: ommendations for physicians who treat almost 30%.8 screening is an integral activity for the patients who smoke. The trial results Although physician training did in- office staff at every office visit, with and subsequent recommendations are crease the use of smoking interventions, physician treatment and follow-up for described briefly below; more details the training alone did not significantly patients who exhibit the risk factor. have been published elsewhere.'There alter smoking cessation rates among pa- As with hypertension, successful in- is now a national program to train phy- tients.' Only when trained physicians tervention involves the office staff and sicians in these brief but effective smok- were routinely reminded to intervene requires effective use of the patient's ing cessation techniques. with all patients who smoked(by using medical record. In tobacco control, ef- SCIENTIFIC BASIS FOR THE chart stickers or other devices)did smok- fective clinical intervention requires rou- ing rates among patients significantly tine discussion of the subject by as many TRAINING decrease.4,6,11 health professionals as possible. Such In 1984, the NCI began support of Results across all five trials may be discussions may be brief,but they must five randomized,controlled intervention summarized as follows: be frequent and routine. trials of brief training and intervention • Training physicians in the treat- Health professionals should be re- protocols for physicians to use with pa- ment of nicotine dependence will result minded to intervene whenever they see in more consistent and more effective a patient who smokes. The NCI recom- From the Cancer Control Science Program,Division patient care. mendations include a reminder system of Cancer Prevention and Control,National Cancer In- 0 When trained physicians are rou- within each office practice.Medical prac- stitute,Bethesda.Md. tinely prompted to intervene with patients tices differ, so office procedures will Reprint requests to the National Cancer Institute, who smoke the can achieve significant v but the goal of organizing an office 9000 Rockville Pike, EPN-241, Bethesda,MD 20892 r y g=� vary, g g g (Dr Manley). reductions in smoking prevalence. for smoking intervention is to ensure 3172 JAMA,December 11,1991—Vol 266,No.22 Clinical Interventions in Tobacco Control—Manley et al Intervention Plan for Patients Who Smoke intervention (assist and arrange) is a A training program now in progress patient self-help program that is encour- will train 100 000 US physicians in smok- Ask all patients about smoking aged, guided, and supported b h sl in intervention techniques. If each Advise all smokers to stop g � � � Pp Y P Y �- g q State your advice clearly, for example, "As your cians and nurses. trained physician can help just 10% of physician,I must advise you to stop smoking now" his or her patients who smoke to stop Personalize the message to quit DESCRIPTION OF THE TRAINING p Assist patients who want to stop now each Y p year, the United States will have Help the patient select a quit date PROGRAM 3 million more ex-smokers annually.The Provide self-help materials Consider prescribing nicotine gum(or transdermal From the results ofthe trials discussed NCI training program is a major effort nicotine patches, when available), especially for above, the NCI produced a manual of to reduce the prevalence of smoking by highly addicted patients Arrange follow-up visits practical smoking cessation techniques rapidly implementing nationwide the re- Set a follow-up visit within 1 to 2 weeks after the quit for use by clinicians.10 Based on infor- sults of definitive clinical trials. date mation in the manual a 3-hour course Have a member of the office staff call or write the Individual physicians and professional organiza- patient within 7 days after the initial visit,reinforcing for physicians and nurses was designed. tions throughout the nation are welcome to partic- the derision to stop and reminding the patient of To teach physicians throughout the ipate in this project and can get information by the quit date Set a second follow-up visit in 1 to 2 months nation,the NCI has established and con- ming 1.800-4-CANCER.(Individual copies of the tinues to seek collaborative relationships NCI manual"can be obtained through calls to this with organizations that Share aCommit- number. Continuing medical education credit is g routinely offered for participation in the NCI ment to cancer prevention and can reach training.) that all patients who smoke are identi- practicing physicians.Collaborating or- I9eferences fied, monitored, and appropriately ganizations to date include the Ameri- treated at every office visit. Organiza- can Cancer Society, American Medical 1. Schultz years of potential Smoking-attributable mortality and life lost—United States, 1988. tional procedures are not complex and Association, Association of American MMWR.1991;40:62-71. usually meld smoothly into a physician's Medical Colleges,American Medical Wo- 2. us Department of Health and Human Services. Smoking and Health:A National Status Report.2nd practice. men's Association, Society of Teachers ed.Rockville,Md:US Department of Health and Hu- Such procedures usually include se- of Family Medicine, Association of man Services;1990:38.US Dept of Health and Human Services publication(CDC)87-8396. lecting a smoking cessation coordinator Teachers of Preventive Medicine,Amer- 3. Wells KB,Ware JE,Lewis CE.Physicians'attitude for the office,making the office tobacco- ican College of Preventive Medicine, in counseling patients about smoking. Med Care. free, implementinga way to,identify man state medical societies and sev- 4. Oc eneJ365. Y � Y r 4. Ockene JK,Kristellaz J,Goldberg R,et al.Increas- and monitor smokers,reminding health eral large health maintenance orgam- ing the efficacy of physician-delivered Smoking inter- ventions:a randomized clinical trial.J Gen Intern Med. professionals to intervene, and using a ZationS. 1991;6:1-8. protocol that involves staff members in The goal of the program is to train S. Glynn TJ, Manley MW, Pechacek TF. Physician- the intervention and follow-up.10 At min- 100 000 physicians.An evaluation of the initiated smoking cessation programs: the National Cancer Institute trials.In:Engstrom P,ed.Advances imum,patients charts should be plainly training programs impact is now in in Cancer Control.New York,NY:Alan R Liss Inc; marked with a sticker or other visible progress. 1990 identifier Of the patient's smoking Sta- 6. Cohen SJ,Stookey GK,Katz BP,Drook CA,Smith Il g DM. Encouraging primary care physicians to help tus. Many practices use additional re- COMMENT smokers quit:a randomized,controlled trial.Ann In- 48-652. minder systems, includingflow sheets There is a tremendous disparity be- 7. Med. gss;llo:Coates Y � � p Y 7. Cummings SR, Coates TJ, Richard RJ, et al. to promote systematic treatment. tween the magnitude of the health prob- Training physicians in counseling about smoking ces- When treating a patient who smokes, lems caused by smoking and the exper- sation: a randomized trial of the "Quit for Life" 7. physicians and other health profession- tise of physicians in treating this addic- program. Wilsong DMC,Taylor DW,Gilbert JR,et al.A ran- als need a methodical plan. The NCI- tion.Clinical trials have clearly established domized trial of a family physician intervention for recommended approach (see Table) is that physicians can significantly reduce amoking cessation.JAMA. 1988;260:1570-1574. 9. Kottke TE,Brekkle ML,Solberg LI,Hughes JR. often referred to as "the four As": (1) smoking among their patients by using Arandomized trial to increase smoking intervention by ask about smoking; (2)advise smokers brief intervention techniques. Adoption physicians:doctors helping smokers,round I.JAMA. 1989;261:2101-2106. to stop;(3)assist patients willing to stop; of a routine intervention by physicians 10. Glynn T,Manley M.How To Help Your Patients and(4)arrange follow-up.5 The four As nationwide will result in marked reduc- Stop Smoking:A National Cancer Institute Manual for Physicians. Bethesda, Md: US Department of intervention plan is a general approach tion in smoking prevalence and, there- Health and Human services;1989.National Institutes that can be adapted for use in very little fore,reduced incidence ofcancerand other of Health publication 89-3064. time, often 3 minutes or less. The a - diseases caused b smoking. There are 11. Prochaska J, okingenoe toC.ward Stages and processes Il �' g' of self-change in smoking:toward an integrative model proach is designed for use in almost any documents now available that encapsu- of change.J Consult Clin Psychol.1983;51:390-395. outpatient setting. When a patient de- late findings from the clinical trials for 12. Glynn TJ,Boyd GM,Gruman JC. Essential ele- ments of self-help/minimal intervention strategies for cides to stop smoking,the recommended easy use by practicing physicians, smoking cessation.Health Educ Q.1990;17:329-345. JAMA,December 11, 1991—Vol 266,No.22 Clinical Interventions in Tobacco Control—Manley et al 3173 I Nicotine Replacement Therapy During Pregnancy Neal L. Benowitz, MD CIGARETTE smoking is a major pre- smoked before pregnancy.' Although 4600 infant deaths in the United States ventable cause of fetal death and injury, 39% of these smokers quit during each year.'These deaths are related to yet many women continue to smoke cig- their pregnancy, about 20%of the pop- a higher than usual incidence of prema- arettes throughout pregnancy.The most ulation of pregnant women continued to ture delivery related to abruptio pla- effective pharmacologic adjunct to smok- smoke. centae,placenta previa,and premature ing cessation therapy is nicotine replace- Other recent surveys report smoking rupture of the membranes in smoking ment, including nicotine chewing gum rates during pregnancy of 21%in 1985 mothers,as well as intrauterine growth or transdermal nicotine. However, the and 1986 in 25 states and the District of retardation. Furthermore,babies born use of nicotine replacement therapy is Columbia,'27.5%in 1986 in Missouri,' to smoking mothers have a four times contraindicated during pregnancy. The 19.9% in 1985 in Idaho, 26.1% in 1988 greater risk of a low Apgar score if the analysis presented herein suggests that and 1989 in New York,"and 23%in 1989 mother smokes two packs per day com- the benefits of nicotine replacement ther- in Ohio.'Among pregnant teenagers in pared with babies of nonsmoking moth- apy to aid smoking cessation in preg- Washington State,32%and 37%smoked ers (with correction for gestational nant women who cannot stop smoking during pregnancy in 1984 and 1988, re- age)." without such therapy substantially spectively.I Of note is that heavier smok- Smoking does not appear to increase outweigh the risks of continued smok- ers are much less likely to quit sponta- the risk of teratogenicity. Some studies ing or the risks of nicotine replacement neously during pregnancy than are show adverse effects of maternal ciga- per se. lighter smokers,presumably reflecting rette smoking on child development."," PREVALENCE OF CIGARETTE a high level of dependence. Thus, al- However, these studies are somewhat SMOKING DURING PREGNANCY though the average number of cigarettes difficult to interpret because of confound- smoked per day declines during preg- ing with socioeconomic class,education Although the prevalence of smoking nancy,'the prevalence of heavy smok- of the parent, and passive smoking. has declined in recent years and many ing, defined as one pack(20 cigarettes) The major effect of cigarette smoking women quit smoking when they become or more per day, during pregnancy re- on birth weight has been termed the pregnant,20%to 25%of pregnant Amer- mains substantial: 13.1%, 12,1%, and fetal tobacco syndrome,defined in terms ican women continue to smoke through- 11.6%in the Missouri, Idaho,and New of the following four points'': out pregnancy; many of these women York State studies, respectively.',, 1. The mother smoked five or more are heavy smokers.The National Health cigarettes per day throughout the preg- Interview survey of women pregnant in RISKS OF CIGARETTE SMOKING nancy. 1985 or with a birth in the previous 5 DURING PREGNANCY 2. The mother had no evidence of hy- years found that 32% of white women Cigarette smoking during pregnancy pertension during pregnancy. substantially increases the risk of spon- 3. The newborn had symmetrical taneous abortion,prematurity,low birth growth retardation at term(greater than From the Departments of Medicine,Psychiatry,and weight and perinatal mortality. Smok- 37 weeks' ion defined as a birth Pharmacy,University of California,San Francisco,and ' eegestation),)r the Division of Clinical Pharmacology and Experimen- ers have an increased risk of spontane- weight of less than 2500 g and a pond- tal Therapeutics,San Francisco General Hospital Med- ously aborting a chromosomally normal eral index (weight in grams/length in ical Center, fetus, with an odds ratio of 1.2 to 1.8 centimeters cubed)exceeding 2.32. This report was submitted to the Food and Drug Ad- g ministration Drugs of Abuse Advisory Committee,May compared with nonsmokers.8-10 The in- 4. No other cause of intrauterine 8,1991.The opinions expressed are those of the au- cidence of low birth weight (<2500 g) growth retardation is obvious. thor and do not necessarily reflect those of the Food increases with increasing cigarette con- It is clear that cigarette smoking rep- and Drug Administration. Reprint requests to the Division of Clinical Pharma- sumption,and 21%to 39%of low-birth- resents a major(if not the major)known cology and Experimental Therapeutics,San Francisco weight births have been attributed to risk to the fetus, and pregnant women General Hospital Medical Center, 1011 Potrero Ave, maternal cigarette smoking.'It is esti- should be strongly encouraged and as- Bldg 30, 5th Floor, San Francisco, CA 94110 (Dr Benowitz). mated that maternal smoking results in sisted in stopping smoking.Prospective 3174 JAMA,December 11, 1991—Vol 266,No.22 Nicotine Replacement During Pregnancy—Benowitz studies of smoking cessation therapy smoked, even though the label says its vated levels of catecholamines in amni- during pregnancy indicate that smoking use is contraindicated in pregnancy. otic fluid in pregnant women who smoke cessation does result in improved birth POTENTIAL RISKS OF NICOTINE during their third trimester indicate weight. IN PREGNANCY sympathetic activation in the fetus,con- sistent with fetal hypoxia and/or direct BENEFIT OF NICOTINE Nicotine could contribute to tobacco- effects of nicotine." Nicotine can also REPLACEMENT THERAPIES related reproductive disorders; how- inhibit production of prostacyclin,a po- Nicotine replacement therapies can ever, the nature and magnitude of its tent local vasodilator and inhibitor of enhance cessation rates in smokers who adverse effects are as yet unknown. platelet aggregation, in arteries. Be- have been unable to stop smoking on While evidence for a causal link between duced prostacyclin-like activity has been their own. Tobacco smoking is main- cigarette smoking and reproductive dis- noted in umbilical arteries of cigarette- tained in most smokers by addiction to orders is well established, the patho- -smoking mothers.32 The above findings nicotine." Tobacco dependence can be physiology is not.Smokers are exposed suggest that nicotine contributes to the treated successfully. Behavioral thera- to several thousand chemicals in tobacco adverse effects of cigarette smoking on pies are successful for some smokers. smoke, including nicotine and carbon reproduction,probably by acting on the Pregnant women who smoke are often monoxide. Nicotine and carbon monox- uteroplacental circulation. highly motivated to quit and may be ide are suspected to contribute to re- Fetal hypoxemia has also been con- more responsive than are other patients productive disturbances.An assessment sidered to be a contributory cause of to advice that they stop smoking. Pre- of the potential contribution of nicotine behavioral abnormalities, such as hy- natal smoking cessation programs have is important for making decisions about peractivity, short attention span, and resulted in quitting rates of 9% to risks vs benefits of nicotine substitution lower scores on spelling and reading 43%.7.15,16 But many smokers, particu- therapy during pregnancy. tests,which occur at a higher frequency larly heavy smokers who are more de- A likely cause of growth retardation in children whose mothers have smoked pendent on nicotine, are still unable to in fetuses of smoking mothers is induc- throughout pregnancy than in those born quit. tion of fetal hypoxia and/or ischemia, to nonsmoking mothers. In addition, Because of the pharmacologic nature which could be produced by both carbon chronic prenatal exposure to nicotine of the addiction process,pharmacother- monoxide and nicotine. The developing via maternal infusions in rats results in apy of tobacco dependence makes sense. fetus is normally in a state of relatively neurobehavioral disturbances in the off- At this time, the only pharmacologic low oxygen tension.Oxygen delivery to spring.m Other neonatal rat studies sug- therapy shown to be effective in smok- the fetus is enhanced by adaptations that gest that fetal exposure to nicotine may ing cessation is nicotine replacement include the presence of fetal hemoglo- result in abnormal neurochemical de- therapy.Nicotine replacement therapy, bin,which has a greater affinity for ox- velopment of the brain.Rats exposed to using nicotine polacrilex chewing gum ygen than does adult hemoglobin;higher nicotine during gestation show delayed or transdermal nicotine delivery sys- hematocrit;higher fetal cardiac output; neural maturation and persistent defects tems, can reduce tobacco withdrawal and high red blood cell 2,3-diphospho- in central catecholaminergic and cholin- symptoms and enhance the efficacy of glycerate level,which facilitates the re- ergic neuronal activity.34-" It is specu- behavioral therapy. A meta-analysis of lease of oxygen from hemoglobin to tis- lated that nicotinic cholinergic recep- nicotine chewing gum therapy indicated sues.Carbon monoxide impairs oxygen tors normally function to modulate neu- an overall smoking cessation rate (in availability by binding avidly to fetal ronal maturation in the fetus and that men and nonpregnant women) at 6 hemoglobin,thereby reducing the avail- excessive neonatal exposure to nicotine months of 27% for nicotine chewing ability of oxygen to the fetus.'A ma- impairs that development. gum comparedwith 18% for placebo ternal 10% blood carboxyhemoglobin RELATIVE RISKS OF chewing gum. 8 The trials were per- level, which can be observed in a two- CIGARETTE SMOKING AND formed in smoking cessation clinics that pack-per-day cigarette smoker, can be NICOTINE REPLACEMENT were able to provide expert behavioral associated with a 10% to 15% higher NICOTI PREGNANCY counseling. Another trial has shown a carboxyhemoglobin level in the fetus dose-response relationship for nicotine than in the mother. This has been Cigarette smoking delivers not only therapy and smoking cessation out- equated to a 60%reduction in fetal blood nicotine and carbon monoxide but also come.11 In this trial,treatment with 4-mg flow.24,25 many other toxic chemicals to smokers. nicotine chewing gum was found to be Nicotine may contribute to fetal is- Smoking in general delivers more nic- more effective than 2-mg nicotine chew- chemia by its effects on the placental otine at a more rapid rate, resulting in ing gum in highly dependent smokers. circulation. Nicotine infusion in preg- more intense cardiovascular and central This rate of effectiveness was compa- nant sheep increases uterine vascular stimulation,than does nicotine from nic- rable with that obtained with 2-mg nic- resistance and reduces uterine blood otine chewing gum or transdermal nic- otine chewing gum in smokers with a flow,effects that appear to be mediated otine delivery systems. Therefore,nic- medium or low level of dependence. by catecholamine release.' Smoking otine replacement therapies are likely Thus, nicotine replacement appears to acutely and chronically reduces placen- to present substantially lower risk than be of particular value in more-depen- tal blood flow in pregnant women,pre- cigarette smoking, particularly heavy dent smokers. Similar data for smoking sumably due to nicotine.27,26 Both ciga- smoking, during pregnancy. cessation rates have been observed with rette smoking and nicotine chewing gum the use of transdermal nicotine delivery increase fetal heart rate during the sec- Delivery of Nicotine systems.20,21 and trimester in humans,consistent with Cigarette smokers systemically ab- Of note is a recent survey of Michigan sympathetic neuroactivation.29 During sorb about 1 mg of nicotine and 10 to 15 family physicians indicating that 51% the third trimester in humans,cigarette mg of carbon monoxide per cigarette." were dissatisfied with the effectiveness smoking or nicotine gum chewing de- Thus,a pack-per-day smoker consumes of available smoking cessation methods.22 creases fetal heart rate and reduces fe- on average 20 mg of nicotine and 200 to Twelve percent prescribed nicotine tal breathing movements,both of which 300 mg of carbon monoxide.Concentra- chewing gum to pregnant women who may be signs of fetal hypoxia.29,30 Ele- tions of nicotine and carboxyhemoglo- JAMA,December 11, 1991—Vol 266,No.22 Nicotine Replacement During Pregnancy—Benowitz 3175 1 bin in the blood tend to build up through- tial tolerance to effects of nicotine, tol- venous nicotine and cigarette smoking out the day with regular cigarette smok- erance, at least to some effects, is not (which results in average plasma nico-' ing,plateauing at levels of 20 to 35 ng of complete. For example, heart rate ac- tine concentrations 175%of those with nicotine per milliliter and 5% to 10% celeration and catecholamine release either alone)were similar to the effects carboxyhemoglobin.Significant concen- have been shown to persist'overnight in of either smoking or intravenous nico- trations of nicotine and carboxyhemo- habitual cigarette smokers." Persis- tine alone. This indicates a flat dose- globin persist overnight,even while the tent sympathetic neural activation is of response curve and suggests that if a smoker sleeps, and smokers typically importance because this could be a meth- person were to use nicotine replacement awaken with plasma nicotine concentra- anism of adverse effects on the fetus. products and smoke as well,the cardio- tions of about 5 ng/mL and carboxyhe- Presumably, the same pattern of sym- vascular effects will not be enhanced. moglobin concentrations of 3%to 5%. pathetic neural activation occurs in us- SUMMARY AND CONCLUSIONS The average systemic absorption of ers of nicotine polacrilex chewing gum nicotine from one piece of 2-mg nicotine or transdermal nicotine delivery sys- Epidemiologic evidence indicates a polacrilex chewing gum is 1 mg." No tems. causal and dose-related relationship be- carbon monoxide or other toxins are ab- tween cigarette smoking and adverse sorbed. With the use of 12 pieces of Cardiovascular Effects in Humans reproductive outcomes.Smoking cessa- nicotine chewing gum per day(which is The major cardiovascular effects of tion during pregnancy reduces the risk more than most smokers use),about 12 nicotine result from sympathetic neural of reproductive problems and can be the mg of nicotine are absorbed, resulting activation."In healthy people,cigarette opportunity for a permanent change in in blood nicotine levels with chewing smoking increases systolic blood pres- life-style to reduce the risk of later smok- gum averaging one third to one half those sure about 10 mm Hg and diastolic blood ing-related chronic diseases. Because observed in cigarette smokers(although pressure 5 mm Hg,increases heart rate cigarette smoking results in exposure in some people concentrations of nico- 10 to 20 beats per minute,and increases to many chemicals, it is impossible to tine may be similar when smoking and cardiac output,owing to both increased ascertain the contribution of nicotine per chewing the gum). heart rate and cardiac contractility.Pe- se from epidemiologic data. Studies of Transdermal nicotine delivery sys- ripheral vascular changes include cuta- the pharmacologic and toxicologic effects tems typically deliver about 15 to 20 mg neous vasoconstriction, systemic veno- of nicotine in animals and experimental of nicotine per day.37 Peak plasma con- constriction, and increased muscular studies of the effects of cigarette smok- centrations of nicotine typically range blood flow.As discussed previously,pla- ing and nicotine in humans suggest that from 10 to 15 ng/mL and tend to fall cental blood flow may decrease, pre- nicotine may contribute to adverse re- somewhat overnight (owing to declin- sumably due to contraction of the al- productive outcomes. Mechanisms of ing rates of nicotine release from the ready maximally dilated uterine blood particular concern include reduction of patch). Thus, the daily dose of nicotine vessels. uteroplacental blood flow and direct ef- and peak blood levels of nicotine from As comparative data on effects on the fects on the developing fetal brain. nicotine chewing gum or transdermal uteroplacental circulation in humans are Nicotine replacement therapy is delivery systems are lower than those not available, it is reasonable to com- clearly beneficial as an adjunct to smok- of one-pack-per-day cigarette smokers. pare the systemic cardiovascular effects ing cessation therapy, particularly in Transdermal delivery systems do result of smoking and nicotine replacement more highly dependent smokers. It is in persistent exposure to nicotine that therapies to consider relative risks.The the more dependent(ie,heavier)smoker at some times of day,particularly over- cardiovascular effects of single admin- who is at more risk for adverse repro- night, may exceed those found in ciga- istrations of nicotine chewing gum(two ductive outcome and who is less likely to rette smokers. 2-mg pieces)are similar in nature but a stop smoking when becoming pregnant. Pharmacodynamic Considerations smaller magnitude those of smoking Nicotine replacement therapy is likely a single cigarette. Studies of the car- to present less risk to the fetus than Nicotine from cigarette smoke is ab- diovascular effect of nicotine gum cigarette smoking. Cigarette smoking sorbed rapidly through the lungs and chewed throughout the day have shown exposes the mother and the fetus to high into the circulation and results in tran- a similar increase in blood pressure but levels of carbon monoxide and similar or sient high arterial blood concentrations a smaller increase in heart rate com- higher concentrations of nicotine com- that are delivered to the brain and other pared with ad libitum cigarette smok- pared with those from nicotine replace- organs. The effects of nicotine are ing.41 ment therapies. In addition, the more greater when doses are administered Circadian cardiovascular effects of rapid delivery of nicotine from cigarette rapidly compared with when the same transdermal nicotine delivery systems smoke produces higher concentrations dose is given more slowly."The rapid have not yet been reported in the med- of nicotine in the brain and other organs high-dose delivery of nicotine to the brain ical literature. In unpublished studies and more intense physiological effects. is thought to be responsible for much of from my laboratory,the effects of trans- There is some concern that 24-hour-per- the psychological stimulation and reward dermal nicotine on blood pressure and day sustained concentrations of nicotine, associated with cigarette smoking. heart rate were less than those of cig- such as result from transdermal deliv- Nicotine from polacrilex chewing gum arette smoking. ery systems, might have different ef- or transdermal delivery systems is ab- Of potential relevance to predicting fects than the rising and falling levels of sorbed slowly and does not produce the effects of transdermal nicotine are stud- nicotine seen in cigarette smokers.How- mental stimulation or euphoria reported ies of the effects of 14-hour intravenous ever, significant levels of nicotine are after cigarette smoking.The blunted re- infusions of nicotine, with or without present overnight even in smokers when sponse is due both to lower concentra- concomitant cigarette smoking.42 Circa- they do not smoke overnight,and avail- tions reaching the brain and other tar- than cardiovascular effects,including 24- able cardiovascular studies comparing get organs and to the development of hour urinary catecholamine excretion, cigarette smoking and nicotine replace- acute tolerance,which is well known to were similar during intravenous nico- ment therapies show, if anything, occur to many effects of nicotine. tine and cigarette smoking conditions. smaller circadian cardiovascular effects Despite the development of substan- Of note, the combined effects of intra- with nicotine replacement com- 3176 JAMA,December 11, 1991—Vol 266,No.22 Nicotine Replacement During Pregnancy—Benowitz 1 pared with cigarette smoking. testing of nicotine replacement therapy This research was supported in part by grants Thus, I conclude that the benefits of in pregnant women who smoke,partic- DA02277 and DA01696 from the National Insti- tutesnicotine replacement therapy to aid ularl m those who smoke 20 or more of Health. p py y I thank Kaye Welch for preparing the manu- smoking cessation in pregnant women cigarettes per day and who have failed script. who cannot stop smoking without such behavioral therapies. Such testing is _ therapy substantially outweigh the risk medically indicated and ethically accept- that Financial has Disen closure.-Dr e.consultant Benowitz t �d re h a of continued smoking or nicotine replace- able based on current concepts of ben- research funded by the several pharmaceutical ment therapy per se, at least in heavy efits vs risks of nicotine replacement companies that manufacture or market nicotine smokers. I recommend formal clinical therapy. replacement medications. References 1. Fingerhut LA,KleinmanJC,KendrickJS.Smok- domized clinical trial. Br J Obstet Gynecol. ing and fetal breathing movements. Br J Obstet ing before, during, and after pregnancy. Am J 1987;94:293-300. Gynaecol.1976;83:262-270. Public Health. 1990;80:541-544. 16. Ershoff DH, Quinn VP, Mullen PD, Lairson 31. Divers WA Jr,Wilkes MM,Babaknia A,Yen 2. Williamson DF, Serdula MK, Kendrick JS, DR. Pregnancy and medical cost outcomes of a SSC.Maternal smoking and elevation of catechola- Binkin NJ.Comparing the prevalence of smoking self-help prenatal smoking cessation program in a mines and metabolites in the amniotic fluid.Am J in pregnant and nonpregnant women,1985 to 1986. HMO.Public Health Rep. 1990;105:340-347. Obstet Gynecol. 1981;141:625-628. JAMA. 1989;261:70-74. 17. Benowitz NL.Pharmacologic aspects of ciga- 32. Ahlsten G,Ewald U,Tuvemo T.Prostacyclin- 3. Stockbauer JW,Land GH.Changes in charac- rette smoking and nicotine addiction. N Engl J like activity in umbilical arteries is dose-depen- teristics of women who smoke during pregnancy: Med. 1988;319:1318-1330. dently reduced by maternal smoking and related to Missouri,1978-88.Public Health Rep.1991;106:52- 18. Lam W, Sze PC, Sacks HS, Chalmers TC. nicotine levels.Biol Neonate. 1990;58:271-278. 58. Meta-analysis of randomised controlled trials of nic- 33. Peters DAV, Taub H,Tang S. Postnatal ef- 4. Centers for Disease Control.Cigarette smoking otine chewing-gum.Lancet. 1987;2:27-30. fects of maternal nicotine exposure. Neurobehav among reproductive-aged women-Idaho and New 19. Tonnesen P,Fryd V,Hansen M,et al.Effect Toxicol. 1979;1:221-225. York.MMWR. 1990;39:659-662. of nicotine chewing gum in combination with group 34. Navarro HA,Seidler FJ,Whitmore WL,Slot- 5. Centers for Disease Control.Effects of mater- counseling on the cessation of smoking.N Engl J kin TA.Prenatal exposure to nicotine via maternal nal cigarette smoking on birth weight and preterm Med. 1988;318:15-18. infusions:effects on development of catecholamine birth-Ohio, 1989.MMWR. 1990;39:662-665. 20. Hurt RD,Lauger GG,Offord KP,Kottke TE, systems.JPharmacol Exp Ther.1988;244:940-944. 6. Davis RL,Tollestrup K,Milham SJr.Trends in Dale LC.Nicotine-replacement therapy with use of 35. Navarro HA,Seidler FJ,Eylers JP,et al.Ef- teenage smoking during pregnancy: Washington a transdermal nicotine patch:a randomized double- fects of prenatal nicotine exposure on development State, 1984 through 1988. AJDC. 1990;144:1297- blind placebo-controlled trial. Mayo Clin Proc. of cental and peripheral cholinergic neurotransmit- 1301. 1990;65:1529-1537. ter systems:evidence of cholinergic trophic influ- 7. Sexton M,Hebel JR.A clinical trial of change in 21. Tennesen P,Norregaard J,Simonsen K,Sawe ences in developing brain.J Pharmacol Exp Ther. maternal smoking and its effect on birth weight. U.A double-blind trial of 16-hour transdermal nic- 1989;251:894-900. JAMA. 1984;251:911-915. otine patch in smoking cessation.N Engl J Med. 36. Benowitz NL,Jacob P III,Savanapridi C.De- b. Surgeon General. The Health Consequences of 1991;325:311-315. terminants of nicotine intake while chewing rico- SmokingforWomen:AReportoftheSurgeonGen- 22. HicknerJ,Cousineau A,Messimer S.Smoking tine polacrilex gum. Clin Pharmacol Ther. e al,1983.Washington,DC:US Government Print- cessation during pregnancy: strategies used by 1987;41:467-473. ing Office; 1983:191-249. US Dept of Health and Michigan family physicians.JAm Board Fam Pract. 37. Benowitz NL, Chin K, Denaro CP,Jacob P Human Services publication 410-889/1284. 1990;3:39-42. III.Stable isotope methodology for studying trans- 9. Stillman RJ,Rosenberg MJ,Sachs BP.Smok- 23. Longo LD.Carbon monoxide:effects on oxy- dermal drug absorption:the nicotine patch. Clin ing and reproduction. Fertil Steril. 1986;46:545- genation of the fetus in utero.Science.1976;194:523- Pharmacol Ther. 1991;50:286-293. 566. 525. 38. Benowitz NL,Kuyt F,Jacob P III.Influence 10. Centers for Disease Control.The Health Ben- 24. Longo LD.The biological effects of carbon mon- of nicotine on cardiovascular and hormonal effects efW of Smoking Cessation:A Report of the Sur- oxide on the pregnant woman,fetus,and newborn of cigarette smoking. Clin Pharmacol Ther. geon General,1990.Rockville,Md:Public Health infant.Am J Obstet Gynecol. 1977;129:69-103. 1984;36:74-81. Service,Office on Smoking and Health;1990. US 25. Bureau MA,MonetteJ,Shapcott D,et al.Car- 39. Benowitz NL. Increased 24-hour energy ex- Dept of Health and Human Services publication boxyhemoglobin concentration in fetal cord blood penditure in cigarette smokers. N Engl J Med. No.CDC 90-8416. and in blood of mothers who smoked during labor. 1986;315:1639-1640. 11. Garn SM,Johnston M,Ridella SA,Petzold AS. Pediatrics. 1982;69:371-373. 40. Benowitz NL,Porchet H,Sheiner L,Jacob P Effect of maternal cigarette smoking on Apgar 26. Resnik R,Brink GW,Wilkes M.Catecholamine- III.Nicotine absorption and cardiovascular effects scores.AJDC. 1981;135:503-506. mediated reduction in uterine blood flow after nic- with smokeless tobacco use:comparison with cig- 12. Naeye RL,Peters EC.Mental development of otine infusion in the pregnant ewe.J Clin.Invest. arettes and nicotine gum. Clin Pharmacol Ther. children whose mothers smoked during pregnancy. 1979;63:1133-1136. 1988;44:23-28. Obstet Gynecol. 1984;64:601-607. 27. Lehtovirta P, Forss M. The acute effect of 41. BenowitzNL.Toxicityof nicotine:implications 13. Fried PA,Watkinson B.36-and 48-month neu- smoking on intervillous blood flow of the placenta. with regard to nicotine replacement therapy. In: robehavioral follow-up of children prenatally ex- Br J Obstet Gynecol. 1978;85:729-731. Pomerleau OF,Pomerleau CS,eds.Nicotine Re- posed to marijuana,cigarettes,and alcohol.J Dev 28. Philipp K, Pateisky N, Endler M. Effects of placement:ACritical Evaluation.New York,NY: Behav Pediatr. 1990;11:49-58. smoking on uteroplacental blood flow.Gynecol Ob- Alan R Liss Inc;1988:187-217. 14. Nieburg P,Marks JS,McLaren NM,Reming- stet Invest. 1984;17:179-182. 42. Benowitz NL,Jacob P III. Intravenous nico- ton PL. The fetal tobacco syndrome. JAMA. 29. Lehtovirta P,Forss M,Rauramo 1,Kariniemi tine replacement suppresses nicotine intake from 1985;253:2998-2999. V. Acute effects of nicotine on fetal heart rate cigarette smoking. J Pharmacol Exp Ther. 15. MacArthur C,Newton JR,Knox EG.Effect of variability.Br J Obstet Gynecol.1983;90:710-715. 1990;254:1000-1005. antismoking health education on fetal size:a ran- 30. Manning FA,Feyerabend C.Cigarette smok- JAMA,December 11, 1991-Vol 266, No.22 Nicotine Replacement During Pregnancy-Benowitz 3177 Law and Medicine Helene M. Cole, MD, Section Editor Tobacco Liability and Public Health Policy Larry 0. Gostin, JD; Allan M. Brandt, PhD; Paul D. Cleary, PhD IN MARCH 1991,the US Supreme Court agreed to hear the ical evidence concerning the risks of their product through ad- case of Cipollone v Liggett Group,'a tobacco liability suit.The vertising that reassured smokers about the safety of ciga- Cipollone case gained notoriety in 1988 when it became the rettes and constituted an"implied warranty"of their product. first instance in which a jury awarded damages to a plaintiff Chesterfields were advertised as"mild,"which Rose Cipol- to compensate harms caused by the use of cigarettes. Since lone reportedly understood to mean safe. Liggett advertised the 1950s, more than 300 such suits had been filed, few had in 1952:`Play Safe,Smoke Chesterfield.Nose,throat,and ac- come to trial,and none had been victorious prior to the Cipollone cessory organs not adversely affected by smoking Chester- ewe.2,3 The Supreme Court will decide whether the federally fields."Liggett claimed to base this promotion on"the results mandated warning labels on cigarette packages preempt tort ac- of a continuing study by a competent medical specialist."In tions claiming that manufacturers failed to adequately warn or other advertisements the company claimed: "Chesterfield misled consumers about the health hazards of smoking. First to Give you Scientific Facts in Support of Smoking." The Court's decision is likely to have profound implications After being urged by her husband to quit smoking, Mrs not only for tobacco use and policy in the United States,but Cipollone switched in 1955 to L & M Filters (Liggett and for the impact of product liability law on fundamental issues Myers),which were advertised as the new"miracle product," of public health and social policy.This article reviews the Ci- the "alpha cellulose"filter that is "just what the doctor or- pollone case and its procedural history;evaluates the impact dered."' (RJ Reynolds Tobacco Company, Winston-Salem, on law and social policy of preempting tobacco litigation;and NC,similarly advertised:"More doctors smoke Camels than examines the multiple important roles of the judicial system any other cigarette."')In a deposition taken before her death, beyond that of providing simple damages.We argue that to- Cipollone explained that"through advertising, I was led to bacco litigation provides a forum for the evaluation of smoking assume that they were safe and they wouldn't hurt me."" behavior and corporate responsibilities.Such litigation is,we Her lawyers presented as evidence hundreds of documents suggest,one component of a comprehensive program for re- recovered through the discovery process,outlining research, ducing cigarette consumption, which currently causes ap- advertising,and public relations strategies of manufacturers proximately 400 000 deaths each year in the United States." to combat the negative impact of the medical evidence. Dur- ing the 1950s,for example,the companies duplicated animal studies that demonstrated that constituents in tobacco could Rose Cipollone, like many women in the 1940s, began to cause cancers."At the trial,a former researcher for Liggett smoke as a teenager.For the next 40 years she smoked a pack and Myers testified that data linking cigarettes to cancer in and a half a day. Although she had tried to stop on several animals in experiments which he conducted had been with- occasions, these attempts failed. According to her own as- held by the company.12 A 1972 memorandum outlined the sessment in legal depositions, she had become addicted. She industry's strategy of"creating doubt about the health charge smoked Chesterfield cigarettes, manufactured by Liggett without actually denying it; advocating the public's right to and Myers, Durham, NC,until 1955,because she felt it was smoke without actually urging them to take up the practice "glamorous"and she wanted to"imitate the pretty girls and and encouraging objective scientific research as the only way movie stars"depicted in the cigarette advertisements. Mrs to resolve the question of health hazard."13 Cipollone kept on smoking even after developing a malignant The federal judge in Newark, NJ, H. Lee Sarokin, said tumor that resulted in the removal of part of her right lung that the evidence warranted the conclusion that the tobacco in 1981 and the entire lung in 1982. industry had participated in a"conspiracy vast in its scope, During the early 1950s, the first major epidemiological devious in its purpose, and devastating in its results" by studies of smoking began to appear in the medical literature. obscuring the dangers of smoking and misleading the public These studies,widely'reported in the public press,found that about the risks of their product.14 Lawyers for the industry cigarette consumption greatly increased the likelihood of lung argued that Mrs Cipollone chose to smoke fully aware of the cancer,heart disease,and stroke."According to Cipollone's debate about health risks. According to defense lawyers, Mrs lawyers,the tobacco industry responded to the growing med- Cipollone had continued to smoke although she was fully aware of the"health controversy."'This was a woman who did what From the American Society of Law and Medicine,Boston,Mass(Mr Gostin),the she wanted to do,"explained one tobacco company lawyer." Department of Social Medicine, University of North Carolina School of Medicine, LEGAL HISTORY Chapel Hill(Dr Brandt),and the Department of Health Care Policy,Harvard Medical School,Boston,Mass(Dr Cleary). Rose Cipollone brought suit against Liggett Group Philip Reprint requests to the Office of the Executive Director,American Society of Law , and Medicine.765 Commonwealth Ave,16th Floor,Boston,MA 02215(Mr Gostin). Morris, New York, NY, and Lorillard, New York, NY, on 3178 JAMA,December 11,1991—Vol 266,No.22 The Cipollone Case—Gostin et al August 1, 1983, a year before her death from cancer at age electronic communications"after January 1,1971.In the 1984 58. She alleged that the manufacturers failed to inform con- act, Congress for the first time required warnings to be sumers adequately of the health risks of smoking,intention- placed in advertisements as well as on packages.I The critical ally neutralized the effect of the federally mandated health issue for the Supreme Court is whether the labeling act warnings through their advertising and public relations cam- preempts post-1965 claims against cigarette manufacturers paigns,knowingly misrepresented the health hazards of smok- from persons harmed by smoking from bringing tort actions, ing, and ignored and withheld from the public medical and premised on the adequacy of health warnings or the manu- scientific evidence.16 facturers' suppression of health information and intentional The cigarette manufacturers defended Cipollone's claims deception of consumers.' Such preemption would insulate by asserting that they were preempted by the Cigarette manufacturers from most lawsuits brought by injured Labeling and Advertising Act of 1965. The federal District smokers. Court held that none of Cipollone's claims were preempted by The preemption doctrine is founded upon the supremacy the labeling act."The Court of Appeals reversed the ruling, clause of the Constitution that renders "This Constitution, concluding that the act preempts state law damage actions and the Laws of the United States. . . . the supreme Law of arising from smoking after January 1, 1966,based on the cig- the Land."" Congress can preempt state law by expressly arette companies'advertising or promotion of cigarettes or saying so,36 by regulating so comprehensively as to`leave no the adequacy of their warnings as to the hazards of smoking.13 room" for state regulation'17 or where federal law conflicts Following that opinion,the case proceeded to trial and the with state law33 because the state law stands"as an obstacle jury returned a verdict of$400000 for Cipollone on a breach to the accomplishment and execution of the full purposes and of warranty claim arising from the misleading advertise- objectives of Congress."39 The question for the Supreme Court ments of the manufacturers before 1966. This jury verdict in the Cipollone case is whether the availability of state was the first in the history of the extensive tobacco litigation common law tort remedies to persons harmed by smoking in which a plaintiff was awarded damages. The Court of cigarettes undermines the accomplishment and execution of Appeals, however, overturned the verdict on technical the purposes and objectives of the labeling act. grounds.19 The case is currently before the Supreme Court. We present three legal arguments why the Supreme Court The Cipollone case was the first federal Court of Appeals should rule against the tobacco industry in the Cipollone case to decide that the labeling act preempted state tort case. First,the Supreme Court recognizes a basic presump- liability. Four additional US Courts of Appeals later con- tion against preemption,40 particularly in a field in which the curred with the Cipollone rnling.3423 The tobacco industry, states have historical authority,41 such as the police power to however, has been less successful in having litigation dis- protect the public health and safety.42.43 The federal courts missed in state courts.24 The New Jersey Supreme Court" have ruled against preemption in analogous cases in nuclear and the Texas Court of Appeals'held that the labeling act energy",'and pesticides,4,41 despite the existence of labeling did not preempt any of the plaintiffs'claims. The Minnesota laws and other federal regulation. Supreme Court,"as well as appellate courts in other states,21,21 Second, the legislative history suggests that the primary ruled that failure-to-warn claims are preempted, but claims purpose of the labeling act was to protect consumers of cig- based on misrepresentation or deception are permitted be- arettes,not the economic interests of the manufacturers.43 In cause they"do not conflict with the objectives of the labeling the months following Surgeon General Luther Terry's report act.117 The Cipollone appeal to the Supreme Court argues that smoking causes lung cancer,49 the FTC proposed ad- that the preemption doctrine applied to smoking cases de- ministrative rules that would have required warning labels to values the state's interest in their citizens'health and safety be placed on all packages and advertising.50 The FTC had and leaves injured citizens without a remedy. already concluded that cigarette advertisements were false THE CIGARETTE LABELING ACT: and deceptive practices because they did not disclose known THE PREEMPTION DOCTRINE health hazards. Spurred by the FTC proposal (which was viewed as usurping Congressional authority), Congress en- Congress enacted the Cigarette Labeling and Advertising acted the Cigarette Labeling and Advertising Act of 1965. Act in 1965.30 It required warning labels on all cigarette The public and legislative perspective of the day was that in packages reading"Caution:Cigarette Smoking May be Haz- light of the compelling evidence of the harm of cigarettes, ardous to Your Health." The act contained a preemption smokers should be apprised of the risks by the federal gov- provision:"No statement relating to smoking and health(other ernment in the name of public health, despite the potential than the congressionally mandated warnings) shall be re- economic repercussions for the tobacco industry. 2 In a case quired on any cigarette package."The labeling act was sub- concerning Federal Communications Commission regulation stantially amended in 197031 to read, "The Surgeon General of tobacco advertising,the District of Columbia Circuit Court Has Determined that Cigarette Smoking is Dangerous to stated that"if we are to adopt[the tobacco industry's]anal- Your Health." It was amended again in 198432 by the Com- ysis of the labeling act,we must conclude that Congress acted prehensive Smoking Education Act to include four warnings to curtail the potential flow of information lest the public on a rotational basis.The warnings include specific reference learn too much about the hazards of smoking for the good of to lung cancer, heart disease, emphysema, and pregnancy the tobacco industry and the economy. We are loathe to complications;the benefits of quitting smoking;the risks of impute such a purpose to Congress. . . ."53 fetal injury and low birth weight;and the presence of carbon Third,in the legislative debates,persons in Congress spe- monoxide in cigarette smoke.Each amendment required stron- cifically addressed state`laws"or"regulations,"suggesting ger warnings.Congress added another preemption provision only that state legislatures or administrative agencies could in the 1970 act: "No requirement or prohibition based on not add a layer of regulation onto federal requirements.',' smoking and health shall be imposed" by a state on manu- The preemption provisions apparently answered the tobacco facturers that adhere to the labeling requirements.This freed industry's principal objection to warning labels: the fear of federal agencies to enter the field. The Federal Trade Com- states' mandating different labels.53 Thus, while Congress mission(FTC)required that all advertisements contain the preempted states from requiring different labels, it left cig- same warning label as used on cigarette packages.33 Congress arette companies free to make additional disclosures.87,58 also prohibited advertising of cigarettes "on any medium of Neither the language of the labeling act nor its legislative JAMA,December 11, 1991—Vol 266, No.22 The Cipollone Case—Gostin et al 3179 history would appear to justify the sweeping and unprece- adolescents,and there is evidence that age of onset is related dented immunity from litigation sought by the tobacco in- to the persistence of smoking.Young persons are more likely dustry. It would be ironic if cigarette manufacturers, which than adults to be influenced by the types of social models so steadfastly resisted the labeling act,50 effectively used it as presented in cigarette advertisements.61 Juries are more a shield against liability. likely to view litigation on behalf of children and adoles- cents in a sympathetic light because minors are perceived AND FAIR REPRESENTATION as being less able to understand, and to resist, misleading advertisements. Underlying the technical issue of preemption are two cen- A Supreme Court decision against preemption would sim- tral claims made by plaintiffs—the tobacco industry failed in ply allow plaintiffs to present claims that cigarette manufac- its duty to adequately warn consumers of salient health risks turers did not meet their legal responsibilities to adequately and to fairly represent its product. Product liability law im- warn and not to mislead consumers. This would put the poses an obligation on manufacturers to disclose to consum- tobacco industry in the same position as manufacturers of ers risks associated with the use of its product that are not other products ranging from drugs,vaccines,and pesticides, already well known to the public.To be sure,injured smokers to ladders and automobiles. may have difficulty in convincing juries that they were un- SMOKING POLICY aware of the health risks of tobacco. Nevertheless,if given the opportunity,plaintiffs could seek to demonstrate through Although the American Medical Association,American Can- objective evidence that the industry was in possession of cer Society, American Heart Association, American Lung relevant public health information that it withheld from con- Association,and American Public Health Association all filed sumers;that the information withheld went beyond the Sur- amicus curiae briefs in the Cipollone case urging the court to geon General's health warnings and could have affected con- allow tobacco litigation,68 many observers, especially critics sumers'choices to smoke;and that the warning labels alone of the liability system,oppose such litigation. They perceive were ineffective in conveying relevant information and dis- plaintiffs like Mrs Cipollone as attempting to shift to pro- closure through public health advertising,packet inserts,and ducers the consequences of their choice to smoke.11 The social other means would have been more effective. functions of tobacco liability litigation,however,need careful Behavioral research, for example, suggests that few peo- evaluation before summarily barring access to the courts. ple are aware of and understand warning labels.110 The factors The value of such litigation is measured by the usefulness of affecting interpretation of the information about risk are the courts in providing (1) a forum for assessing complex complex and include both the context and wording of the questions about corporate responsibilities in an industry al- message. Cigarette manufacturers "tell" consumers many ready required to print warning labels; (2) a remedy for things. The messages about health risks receive little em- smokers injured as a result of corporate failure to warn or phasis, are not supported by other visual and verbal mes- presentation of misleading information;(3)a regulatory mech- sages associated with smoking, and are counteracted by anism for industries that produce inherently dangerous prod- manufacturers' statements discrediting the scientific basis ucts by holding them to defined standards of disclosure and for the warning labels they are compelled to put on their fair advertising, and encouraging design of safer products; packaging. and (4) a method of allocating responsibility for disease in Even if one believes that cigarette manufacturers should society and individual injuries induced by unsafe products. be exempt from the duty to warn consumers about pertinent health information,should the industry also be exempt from A Forum the obligation not to mislead consumers?Misrepresentation The transcending question is whether public policy should 'Ss based on a duty to tell the truth, not a duty to warn provide a forum for a series of credible questions posed in . . . (asserting) the falsity of what the cigarette manufac- current litigation. Did the tobacco industry knowingly con- turer has chosen to say,"not what it has failed to say. The ceal information it had in its possession concerning the dan- Minnesota Supreme Court said that in order to preempt gers of smoking? Did the industry knowingly mislead the misrepresentation claims "we would have to assume that public through deceptive advertising that provided incorrect Congress intended that act to be a license to he . . . (to allow information in promoting cigarettes?Do the mandated warn- manufacturers)deceitful advertising practices."61 ing labels fulfill all responsibilities on the part of the industry Precedents for misrepresentation claims go back as far as to consumers? Finally, what is the nature of individual and the 1930s when the FTC found that RJ Reynolds was en- corporate responsibility if a product is known or suspected of gaging in"false,deceptive and misleading"practices for ad- being dangerous? vertising that"the wind and physical condition of athletes will A Supreme Court decision in the Cipollone case against not be impaired by the smoking of Camel cigarettes,as many preemption opens up the courts as a forum for the resolution as one likes . . . and that the smoking of Camels is not dis- of these important questions.The courts can carefully weigh advantageous to breathing capacity during an athletic con evidence and arguments on both sides of these claims.Given test."62 the complexities of the issues at stake,the courts are likely Currently, the antismoking lobby's most promising mis- to provide the best available forum for their elucidation and representation cases involve advertisements that appear to adjudication. A Supreme Court decision that allows suits to encourage adolescents to purchase cigarettes. While the in- be brought, of course, makes no assumption about the res- dustry denies that it advertises to.minors (who are legally olution of individual lawsuits which would be adjudicated on barred from purchasing cigarettes),6'advertisements in teen- their merits. Juries in the past have been hostile to oppor- age and sports magazines reach a predominately youthful tunistic claims for damages in the absence of solid evidence audience with images conveying adventure and risk taking.6' in relation to one or more of the questions raised above. Sponsorship of sporting events also enables young viewers to see brand name cigarettes prominently displayed on televi- sion(such as on racing cars),without any health warning and Not only are the courts an appropriate forum for the as- despite the statutory ban on broadcast advertising.65,ss A sessment of risk and responsibility,they also have tradition- high proportion of smokers start smoking while they are ally provided a remedy for persons injured through unlawful 3180 JAMA,December 11, 1991—Vol 266,No.22 The Cipollone Case—Gostin et al 1 corporate behavior. The common law has for centuries pro- These are telling policy arguments against product liability vided compensation to persons injured by products as a result litigation. Yet, if a case could ever be made for the manu- of inadequate warnings or deceptive advertising by manu- facturer's bearing the cost of injuries, cigarette companies facturers.Yet,no federal or state statute provides a remedy would be a prime candidate. Under product liability theory, even if smokers can prove that cigarette manufacturers were the burden of compensation should be shifted to those who in possession of relevant health information and deliberately are best able to spread the cost,to those in the best position withheld it from the public or that they intended to deceive. to control the danger, and to those who secure economic A decision by the Supreme Court that the labeling act pre- benefit from the sale of the hazardous product.89 The tobacco empted all tort actions, therefore, would leave the public industry, on each count, is best able to bear the costs of without legal recourse.The Supreme Court has,in the past, smoking-induced injuries. been unwilling to permit such a result,70 stating that Con- gress would not, without comment, remove all means of Assumption of Risk judicial recourse for consumers injured by illegal conduct." Who bears the responsibility for the health consequences of smoking?The tort system, of course,has another feature A Regulatory Mechanism frequently unspoken. Tort litigation allocates blame and as- While many people in public health perceive product lia- signs responsibility. Who is responsible for the burden of bility law as being harmful to the development of drugs, disease imposed by cigarette smoking? Americans tend to vaccines,and medical devices,"it could have a salutary effect believe that individuals can and should take control of their in regulating inherently dangerous products like tobacco and health,that individuals must make decisions to prevent dis- alcoholic beverages. The tort system may force the tobacco ease and promote their own health.90 Cigarette smoking has industry to disclose more health information to the public in become, in many respects, the preeminent example of this the form of stronger warnings,packet inserts, and/or mod- model of health-related behavior. According to this logic, ification of its advertising campaign.Successful duty-to-warn individuals should heed public health advice and government suits could result in the industry's disclosing clear, simple, warnings, act responsibly, and stop smoking. If individuals and more powerful messages to the consumer. suffer deleterious health consequences as a result of smoking, A Supreme Court decision favoring the Cipollone case they have no one to blame but themselves. It is, of course, might even encourage the FTC to take a proactive position this argument that has severely limited the success of tobacco in remedying misrepresentation by ordering cigarette com- liability litigation in the past. panies to advertise the dangerousness of their product. The Assigning responsibility and affixing blame,however,are FTC orders requiring affirmative disclosures and corrective more complicated than they might appear.The Surgeon Gen- advertising have been upheld by the courts.73 eral's warning on cigarette packages is countermanded through Forcing tobacco companies to plough back part of their prof- advertising, as well as social and cultural norms and peer its to provide a more balanced informational environment may pressures.More important,a person's"choice"to stop smok- reap public health benefits,including fewer new smokers re- ing is affected by the powerful addictive qualities of tobacco cruited, decreased consumption by current smokers, and use.91 Recent sociodemographic data suggest that cigarette more smokers quitting.1,75 A genuine threat of liability would smoking is increasingly stratified by educational status and focus the attention of the tobacco industry on the effects of socioeconomic class.92 Some individuals, then, are clearly at misleading advertising on children,women,and ethnic minor- greater risk to become and remain smokers than others." ities, where particular brands and advertising campaigns Many factors beyond rational assessment of health risks in- have been focused.71-11 The industry would,at minimum,have fluences smoking behavior. to reassess the economic value of creating adventuresome, Given these variables,it becomes difficult to conclude that healthy, safe,and erotic images in stark contrast to the per- smokers bear all the responsibility if they become ill; con- functory and bland warnings of the Surgeon General.79 versely, individuals are able to exert some control over be- A principal criticism of the regulatory effect of the tort haviors like cigarette smoking(almost 40 million Americans system is that it will deter manufacturers from researching have quit since 1964).59 More than one party may be respon- or marketing products that pose a substantial risk of liability, sible for an injury;some courts hearing tobacco litigation may such as contraceptive devices'or vaccines.8 'Additionally, conclude, as did the jury in Cipollone,that both the smoker damages imposed by the tort system may be passed on to the and the companies bear partial responsibility for the injuries. consumer as price increases, making valuable health care The courts have traditionally and(for the most part)effec- products unaffordable and driving up insurance premiums.' tively performed this role of adjudicating responsibilities for While public policy militates against increased cost for drugs, injury. To deny access to the courts to individuals who seek vaccines, and medical devices, it pushes in exactly the op- to prove that they have been injured by cigarettes subverts posite direction for tobacco.Here,public health strategy may the traditional judicial function of weighing evidence and ar- seek an increase in cost,which is likely to result in a decrease guments to affix responsibility for harms that have occurred. in use of the product.While some price inelasticity is evident This is a complicated social calculus that only the courts may in the sale of cigarettes, particularly to highly dependent determine with clarity, on a case-by-case basis. smokers, a rise in price is likely to discourage sales, partic- ularly to younger and poorer people. Critics of the tort system also argue that compensation to We express no preference about the disposition of any injured consumers is neither fair nor efficient.86 The tort particular case involving the liability of the tobacco industry. system is not fair because it does not compensate all those Each case must be adjudicated on the basis of the strengths who are injured through wrongdoing and may even reward or weaknesses of the evidence presented.At the heart of our some who are not entitled to it. Juries also provide awards argument is that significant social costs will be incurred if the which are disproportionate to the need,either because they Supreme Court preempts most tobacco litigation.These costs have an antipathy to the industry being sued or because they include losing the court as a forum for the adjudication of view it as a "deep pocket."87 The tort system is inefficient complex issues of smoking policy, a remedy for injured per- because the true costs of compensation are magnified many sons, a regulatory mechanism for the manufacture of an in- times over by lawyers'fees,administration,and court costs.88 herently dangerous product, and a process for assigning re- JAMA, December 11, 1991—Vol 266, No.22 The Cipollone Case—Gostin et al 3181 1 sponsibility for disease in society and individual ill health velopment Commission,103 SCt 1713(1983). 46. Ferebee v Chevron Chemical Co,736 F2d 1529(DC Cir 1984). induced by smoking. The tobacco industry should face the 47. Wisconsin Public Intervenor v Mortier,111 SCt 2476(1991). same stark choices that manufacturers in other sectors face. 48. US Code Cong&Admin News.1965:2350-2361. S Advisory Committee to the Surgeon General the Public Health Service. They must make more adequate disclosures, curtail their Smoking and Health:Report of the Advisory Committee to the Surgeon General of advertising, or compensate injured parties through the tort the Public Health Service.Washington,DC;Office of the Surgeon General,US Dept SYStem.� of Health,Education and Welfare;1%4.Public Health Service publication 1103. 50. Fritsehler A.Smoking and Politics:Policymaking and the Federal Bureaa- Tobacco liability suits are only one aspect of a comprehen- cracy.New York,NY:Appleton-Century-Crofts;1969. sive program to control tobacco consumption in the United 51. Federal Register.1964;29:8324. 52, Brandt AM.The cigarette,risk and American culture.Daedalus.1990;119:155- States. Other strategies to reduce tobacco-induced disease 176. would include more education, research on techniques for 53. Banzhaf v FCC,405 F2d 1082,1089(DC Cir 1968),cert denied,396 US 842 cessation, higher taxes, selective bans, and further le sla- (lsss). !� 54. Cipollone I,593 F Supp at 1159.61. tive and regulatory control of advertising. Even a more vig- 55. Common law claims challenging adequacy of cigarette warnings preempted un- der the federal Cigarette Labeling and Advertising Act of 1%5:Cipollone v Liggett (croup.St Johns Law Rev.1986;32:754,762. governments, however, would not necessarily obviate the 56. HR Rep No.449,89th Cong,1st Sess 4,9(1965). role of litigation.Perhaps the worst outcome in the Cipollone 57. Morrison AB.Brief of Amici Curiae Am Cancer Society,Am Heart Assn,Am case would be a decision b the Supreme Court that anted Pub Health Assn,and Public Citizen.Kyte and Cann v Philip Morris.Mass Sup Jud y p Ct,SJC-5165. preferred status to the tobacco industry when all other man- 58, Tribe L.Federalism with smoke and mirrors.Nation.June 7,1986:788-789. ufacturers were still required to make the tough choices 59. Surgeon General.Reducing the Health Consequences of Smoking:35 Years of tl g Progress:AReport of the Surgeon General.Washington,DC:US Dept of Health and imposed by the tort system. Human Services;1989.US Dept of Health and Human Services publication CDC 8s-8411. References 60. Leventhal H,Cleary PD.The smoking problem:a review of the research and theory in behavioral risk modification.Psychol Bull.1980;88:370405. 1. Cipollone v Liggett Group,cert granted,111 S Ct 1386(March 25,1991). 61. Forster v RJ Reynolds Tobacco Co,437 NW2d 655,662(Minn 1989). 2. Greenhouse L.Courts to say if cigarette makers can be sued for smokers'can- 62. In re RJ Reynolds Tobacco Co,46 FTC 706,720-727-28(1950). cer.New York Times.March 26,1991. 63. RJ Reynolds Tobacco Co.We don't advertise to children.Time.April 9,1984:91. 3. Riley P.The product liability of the tobacco industry:has Cipollone v Liggett Advertisement. Group finally pierced the cigarette manufacturers aura of invincibility?Boston Coll 64. Altman D.How an unhealthy product is sold:cigarette advertising in magazines, Law Rev.1989;30:1103-1178. 1960-1985.J Commun.1987;37:95-106. 4. Death toll from smoking is worsening.New York Times.February 1,1991. 65. Ledwith F.Does tobacco sports sponsorship on television act as advertising to 5. Hammond EC, Horn D. Smoking and death rates: report on 44 months of children?Health Educ J. 1984;43:85-88. follow-up on 187,783 men,I:total mortality.JAMA.1958;166:1159-1172. 66. Blum A.The Marlboro grand prix:circumvention of the television ban on to- 6. Hammond EC, Horn D. Smoking and death rates: report on 44 months of bacco advertising.N Engl J Med.1991;324:913-917. follow-up on 187,783 men,II:death rates by cause.JAMA. 1958;166:1294-1308. 67. Cleary PD,Hitchcock JL,Semmer N,Flinchbaugh LJ,Pinney JM.Adolescent 7. Wynder EL,Graham EA.Tobacco smoking as a possible etiologic factor in bron- smoking:research and health policy.Milbank Q. 1988;66:185-235. chiogenic carcinoma:a study of 684 proved cases.JAMA.1950;143:329-396. 68. Tobacco Products Litigation Rptr.1991;61B:1-231. 8. Doll R, Hill AB. A study of the aetiology of carcinoma of the lung. BMJ. 69. Cavilia JW.Nonsmoker speaks up for freedom of choice.New York Times.April 1952;2:1271-1286. 10,1991;A:24. 9, Whiteside T.Selling Death:Cigarette Advertising and Public Health.New York, 70. Linn v United Plant Guard Workers,383 US 53,64(1966). NY:Liveright;1971. 71. Construction Workers v Laburnum Corp,347 US 656,663-664(1954). 10. Janson D. Cancer victim's husband testifies on her death.New York Times. 72. Huber PW.Liability:The Legal Revolution and Its Consequences.New York, April 7,1988. NY:Basic Books Inc Publishers;1988. 11. Kolata G.Smoking and cancer:what the cigarette companies really knew.New 73. Warner-Lambert Co v Federal Trade Commission,562 F2d 749,756(DC Cir York Times.June 17,1988. 1977),cert denied,435 US 950. 12. Wald ML.Using liability law to put tobacco on trial.New York Times.Febru- 74. Chapman S.Cigarette advertising and smoking:a review of the evidence.In: ary 14,1988. British Medical Association.Smoking Out the Barons:The Campaign Against the 13. Janson D.Data on smoking revealed at trial.New York Times.March 13,1988, Tobacco Industry:A Report of the British Medical Association on Public Affairs 14. Agins T,Freedman AM.Tobacco firms misled public,US judge says.Wall St Division.New York,NY:John Wiley&Sons Inc;1986:79-97. J.April 22,1988. 75. Warner KE.Selling Smoke:Cigarette Advertising and Public Health.Wash- 15.Janson D.Tobacco company is called blameless in a death.New York Times. ington,DC:American Public Health Association;1986. June 2,1988. 76. Cooper R,Simmons BE.Cigarette smoking and ill health among black Amer- 16. Brief for the Petitioner,in the Supreme Court of the United States,Cipollone icans.N Y State Med J.1985;85:344-349. v Liggett Group,No.90-1038. 77. Williams L.Tobacco companies target blacks with ads,donations,and festivals. 17. Cipollone v Liggett Group,593 F Supp 1146(D NJ 1984). Wall St J.October 6,1986. 18. Cipollone v Liggett Group,789 F2d 181,187(3rd Cir 1986),cert denied,479 US 78. Fiore MC,Novotny TE,Pierce JP,Hatziandreu EJ,Davis RM.Trends in cig- 1043(1987). arette smoking in the United States:the changing influence of gender and race. 19. Cipollone 11,893 F2d 541(3d Cir 1990). JAMA.1989;261:49-55. 20. Stephen v American Brands,825 F2d 312(11th Cir 1987). 79. Altman D.How an unhealthy product is sold:cigarette advertising in magazines, 21. Palmer v Liggett Group,825 F2d 620(1st Cir 1987),revg 633 F Supp 1171(D 1960-1985.J Commun.1987;37:95106. Mass 1986)(permitting such common law failure-to-warn claims would"disrupt ex- 80. Feldman v Lederle Laboratories,460 A2d 203,209(NJ Super.1983). cessively"the"carefully wrought balance of tuitional interests"struck by Congress). 81. McKenna A.The impact of product liability law on the development of an AIDS 22. Roysdon v RJ Reynolds Tobacco Co,849 F2d 230(6th Cir 1988),affg 623 F Supp vaccine. Univ Chicago Law Rev.1988;55:943-964. 1189(DC Tenn 1985). 82. Brown v Superior Court(Abbott Laboratories),751 Ptd 470,479(Cal 1988) 23. Pennington v Vistron Corp,876 F2d 414(5th Cir 1989)(failure-to-warn claim (before it was withdrawn because of the costs of liability,the price of Bendectin in- preempted,but not strict liability claim for manufacturing an unreasonably danger- creased by over 3000. ous product). 83. Chapman S. Tobacco excise and declining tobacco consumption: the case of 24. Gilboy v American Tobacco Co,582 Sold. Papua New Guinea.Am J Public Health.1990;80:537-540. 25. Dewey v RJ Reynolds Tobacco Co,577 AM 1239(NJ 1990). 84. Lewit E,Coate D,Grossman M.The effects of government regulation on teen- 26. Carlisle v Philip Morris Inc,805 SW2d 498(Tex Ct App 1991). age smoking.J Law Econ.1981;24:545569. 27. Forster v RJ Reynolds Tobacco Co,437 NW2d 655(Minn 1989). 85. Chaloupka F.Rational addictive behavior and cigarette smoking.J Polit Econ. 28. Hite v RJ Reynolds Tobacco Co,578 A2d 417(Pa Super Ct 1990)(defective de- 1991;99:722-742. sign claim not preempted,but was dismissed on substantive legal grounds). 86. Vicusi WK.Reforming Products Liability.Cambridge,Mass:Harvard Univer- 29. Rogers v RJ Reynolds Tobacco Co,557 NE2d 1045(Ind Ct App 1990). sity Press;1991. 30. 79 Stat 283(1965). 87. Ortho Pharmaceutical Corp v Wells,788 F2d 741(11th Cir 1986)($4.7 million 31. 84 Stat 88(1970). jury verdict against spermicide manufacturer for birth defects despite absence of 32. 15 USC 31333. scientific evidence of a causal connection). 33. In the Matter of Lorillard,80 FTC 455(1972). 88. Domenici PV,Koop CE.Sue the doctor?there's a better way.New York Times. 34. Cigarette Labeling and Advertising Act,15 USC 31334. June 6,1991:A25. 35. Article VI,cl 2. 89. Becker v Interstate Properties,569 F2d 1203,1209-1212(3rd Cir 1978). 36. Jones v Rath Packing Co,430 US 519,525(1977). 90. Knowles J.The responsibility of the individual.Daedalus.1977;106:57-80. 37. Rice v Santa Fe Elevator Corp,331 US 218(1947). 91. Surgeon General. The Health Consequences of Smoking:Nicotine Addiction: 38. Florida Lime and Avocado Growers v Paul,373 US 132,142-143(1963). Surgeon General's Report. Washington, DC: US Dept of Health and Human 39. Hines v Davidowitz,312 US 52(1941). Services,Public Health Service;1989.US Dept of Health and Human Services pub- 40. Maryland v Louisiana,451 US 725,746(1981). lication CDC 88-8406. 41. Hillsborough County v Automated Medical Laboratories,411 US 707, 719 92. Escobedo LG, Anda RF, Smith PF, Remington PL, Mast EE. Sociodemo- (1985) graphic characteristics of cigarette smoking initiation in the United States.JAMA. 42. Rice v Santa Fe Elevator,331 US 218,230(1947). 1990;264:1550-1555. 43. Taylor v General Motors Corp,875 F2d 816(11th Cir 1989). 93. Headen SW,Bauman KE,Deane GD,Koch GG.Are the correlates of cigarette 44. Silkwood v Kerr-McGee Corp,104 SCt 615(1984). smoking initiation different for black and white adolescents?Am J Public Health. 45. Pacific Gas and Electric Co v State Energy Resources Conservation and De- 1991;81:854-858. 3182 JAMA,December 11,1991-Vol 266,No.22 The Cipollone Case-Gostin et al Commentary The New Vital Sign Assessing and Documenting Smoking Status AT THE turn of the last century, the American medical vital signs, these distressingly low rates could have been community developed a standardized assessment to help ch- markedly improved. nicians confront the leading cause of death at that time, Most patients don't object to having their blood pressure infectious disease. This assessment, known as vital signs, measured during every clinic visit, and there are no data to included temperature,pulse rate,respiratory rate,and,later, suggest that they would object to the regular assessment of blood pressure.'-'Over time,the measurement of vital signs smoking status,particularly if presented in the context of rou- became an expected part of every clinic visit and an essential tine preventive care.',' In fact, assessing smoking status at component of the database physicians use to evaluate, diag- every visit might, by itself, motivate some smokers to con- nose, and treat patients. sider making a quit attempt. Making smoking status a vital sign would also promote the guidelines of the US Preventive See also pp 3139 and 3172. Services Task Forces that "tobacco cessation counseling should be offered on a regular basis to all patients who smoke cigarettes"and the National Health Promotion and Disease As we approach the next century, American medicine is Prevention Objectives for the Year 200010 to"increase to at challenged by a different cause of illness and death—tobacco least 75%the proportion of primary care and oral health care use. Cigarettes are now responsible for more than 430000 providers who routinely advise cessation and provide assis- deaths each year in the United States.' As with past epi- tance and follow-up for all of their tobacco-using patients." demics of this magnitude, institutional changes in the prac- The singularly devastating health impact of cigarette smok- tice of medicine must be adopted to overcome the enormous ing warrants the unique institutional response of expanding disease burden resulting from tobacco use. the vital signs.While not by itself sufficient,the elevation of Making smoking status the"new vital sign"is a simple way smoking assessment to the priority status of"a vital sign" to confront the chief avoidable cause of illness and death in would be an important first step in a coordinated action plan our society today. This small but fundamental change in for physicians to aggressively confront smoking. This action clinical practice will begin to address a current weakness in plan would include the following: the way we practice medicine—the failure to universally • Assess and document smoking status as part of the vital assess, document, and intervene with patients who smoke. signs for every patient during every clinical visit. Seventy Current standards of practice warrant appropriate documen- percent of smokers visit a physician each year,"and most of tation and intervention for elevated blood pressure,increased them are motivated to quit. Adding smoking status to the temperature, and arrhythmias. Adding smoking status as a vital signs assessment, an activity usually completed by a new vital sign will significantly increase the likelihood for nurse or medical assistant prior to the physician's encounter, intervention in this important area as well, will ensure that all smokers are identified.A model vital sign In the current issue of THE JOURNAL, Frank and col- stamp12 that includes smoking status assessment is shown in leagues' provide compelling evidence that this institutional the Figure. change is needed. These authors, in a large, well-conducted • Learn and use a brief intervention message to help pa- study,have convincingly documented that fewer than half of tients to quit.Physicians,once aware of the smoking status of all smokers reported that they had ever been advised by their their patients, can then focus on moving smokers from the physicians to quit or cut down.This disappointingly low rate stage of contemplating quitting to making a quit attempt.To of smoking status assessment was observed across most so- do this,physicians must become comfortable with a brief clin- ciodemographic subpopulations and was particularly low ical intervention to assist their patients who smoke.The Na- among young smokers and women who both smoked and used tional Cancer Institute has developed guidelines,based on the oral contraceptives. A meager 4%of smokers reported that results of a large series of clinical trials,for an effective,2-to their physician had helped them quit. These findings are 3-minute clinical intervention for physicians (How to Help regrettably similar to those of an earlier study of smokers in Your Patients Quit Smoking).13 These guidelines and the Na- Michigan, only 44% of whom reported that a physician had tional Cancer Institute's program to train 100000 physicians ever told them to quit.'In both of these studies,had smoking are summarized in an important report by Manley and col- status been assessed during the routine documentation of leagues"in this issue of THE JOURNAL. These authors have estimated that 3 million smokers would quit annually if this From the Department of Medicine and the Tobacco Research and Intervention intervention were universally adopted and succeeded with as Program,University of Wisconsin Medical School,Madison. few as 10%of American smokers.While physicians routinely Reprint requests to Tobacco Research and Intervention Program,University of advise and counsel patients with diabetes hypertension Or Wisconsin Medical School, 7275 Medical Sciences Center, 1300 University Ave, r Madison,WI 53706(Dr Fiore). even a sprained ankle,they have lagged in delivering a brief JAMA,December 11, 1991—Vol 266, No.22 Commentary 3183 legal product in our society today that, when used as in- VITAL SIGNS tended,results in illness and death.No matter how successful smoking cessation activities are in this country,they will be Blood Pressure rendered ineffective if the US tobacco industry continues to Pulse enlist 1 million new young people into the ranks of addicted Temperature smokers each year.21 At a minimum, physicians should sup- port a total ban on tobacco product advertising and promo- Respiratory Rate tion, enforcement of local and state ordinances that outlaw Smoking Status Current Former Never the sale of cigarettes to minors, a ban on cigarette vending Circle machines,and innovative preventive education efforts to con- vince young people never to start smoking. The article by Frank and colleagues5 provides compelling Vital signs stamp. evidence that clinicians are failing to adequately address the needs of their patients who smoke. Manley and colleagues'' but effective message to their patients who smoke. have outlined a simple and effective model to intervene with • Recognize cigarette smoking as a chronic disease and the patients. Adding smoking status as a new vital sign will provide appropriate long-term assistance to patients who provide the institutional framework by which the epidemic of smoke. As with hypertension, diabetes, or congestive heart tobacco use can be universally confronted. Through these failure, cigarette smoking can be considered a chronic dis- clinical and public health interventions, clinicians can have a ease,requiring ongoing attention and treatment.Since smok- positive impact on the national goal of a smoke-free society ers may be at various stages in the quitting process, physi- by the year 2000. cians must continue to assist their smoking patients, often Michael C. Fiore, MD, MPH through repeated contacts over many months or years. Re- cent quitters are at particularly high risk for relapse for 1. Seguin EC.The use of the thermometer in clinical medicine.Chicago Med several months after quitting. This group will particularly J.1866;23:193-201. benefit from follow-up visits for smoking cessation.15 2. Seguin E.Clinical thermometry.Med Rec. 1867;1:516-519. 3. Musher DM,Dominguez EA,Bar-Sela A.Edouard Seguin and the social • Expect modest success rates among patients who try to power of thermometry.N Engl J Med. 1987;316:115-117. quit smoking.As with other chronic diseases,the treatment 4. Centers for Disease Control.Smoking-attributable mortality and years of potential life lost:United States, 1988.MMWR. 1991;40:62-71. of cigarette smoking must continue even in the face of very 5. Frank E,Winkleby MA,Altman DG,Rockhill B,Fortmann SP.Predictors modest"cure"rates.A realistic smoking cessation rate of 5% of physician's smoking cessation advice.JAMA.1991;266:3139-3144. to 10%can be expected from a brief clinical intervention(such 6. Anda RF,Remington PL,Sienko DG,Davis RM.Are physicians advising smokers to quit?the patient's perspective.JAMA. 1987;257:1916-1919. as the one described by Manley and colleagues"').While most 7, Cohen SJ, Christen AG, Katz BP, et al. Counseling medical and dental physicians would be elated by a 10% remission rate with patients about cigarette smoking:the impact of nicotine gum and chart re- unmedicated hypertensive patients,many clinicians become minders.Am J Public Health. 1987;77:313-316. 8. Solberg LI,Maxwell PL,Kottke TE,Gepner GJ,Brekke ML.A systematic discouraged with smoking cessation interventions because of primary care office-based smoking cessation program.JFam Pract.1990;30:647- high relapse rates. If clinicians universally achieved a 5%to 654. 9. Michael Fisher, ed. Report of the U.S. Preventive Services Task Force: 10% successful cessation rate each year among all of their Guide to Clinical Preventive Services. Baltimore, Md:Williams&Wilkins; patients who smoke, the impact would be enormous. 1989:289-298. o Establish expertise in the diagnosis and pharmacologic 10. Healthy People 2000:National Health.Promotion and Disease Prevention Objectives.Washington,DC:US Dept of Health and Human Services;1991. treatment of nicotine addiction. Surgeon General Koop, in Publication PHS 91-50212. his landmark 1988 report on The Health Consequences of 11. Ockene JK.Smoking intervention:the expending role ofthe physician.Am Smoking,16 concluded that cigarettes and other forms of to- J Public Health. 1987;77:782-783. 12. Fiore MC,Pierce JP,Remington PL,Fiore BJ.Cigarette smoking:the bacco are addicting and that nicotine is the drug in tobacco clinician's role in cessation,prevention,and public health.Dis Mon.1990;35;181- that causes addiction. While this view is now widely ac- 242• cepted, physicians have not effectively used pharmacologic 13. Glynn T,Manley M.How to Help Your Patients Stop Smoking:ANational Cancer Institute Manual for Physicians.Bethesda,Md:National Institutes treatments to help smokers quit. This has been particularly of Health; 1989. US Dept of Health and Human Services publication NIH true with nicotine gum, the incorrect and insufficient use of 89-3064• 14. Manley M,Epps RP,Husten C,Glynn T,Shapland D.Clinical interven- which has been well documented."Understanding the role of tions in tobacco control: a National Cancer Institute training program for pharmacologic adjuvants in smoking cessation treatment will physicians.JAMA. 1991;266:3172-3173. become even more important over the next year,when a new 15. Wilson DM,Taylor DW,Gilbert JR,et al.A randomized trial of family physician intervention for smoking cessation.JAMA. 1988;260:1570-1574. product, the transdermal nicotine patch, is expected to be 16. Surgeon General. The Health Consequences of Smoking:Nicotine Ad- licensed by the Food and Drug Administration. The trans- diction:AReport ofthe surgeon General.Washington,DC:US Dept of Health and Human Services;1988.Publication CDC 88-8406. dermal patch has shown great promise as an aid to smoking 17. Cummings SR,Hansen B,Richard RJ,Stein MJ,Coates TJ. Internists cessation",19 and will likely be an important component of and nicotine gum.JAMA. 1988;260:1565-1569. smoking cessation treatment through the 1990s. 18. Hurt RD,Lauger GG,Offord KP,Kottke TE,Dale LC.Nicotine replace- ment therapy with use of a transdermal nicotine patch:a randomized,double- 0 Play a public health role in confronting cigarette smok- blind,placebo controlled trial.Mayo Clin Proc. 1990;65:1529-1537. ing. The power of physicians to influence health behavior in 19. Daughton DM,Heatley SA,Prendergast JJ,et al.Effect of transdermal this country is immense.In fact,physicians are one of the few nicotine delivery as an adjunct to low-intervention smoking cessation therapy. Arch Intern Med. 1991;151:749-752. groups that have the logistical and moral force to confront the 20. Samuels B, Glantz SA. The polities of local tobacco control. JAMA. pandemic of tobacco addiction,illness,and death.20 Through 1991;266:2110-2117. targeted actions, physicians can markedly change the social 21. Pierce JP,Fiore MC,Novotny TE,Hatziandreu EJ,Davis RM.Trends in cigarette smoking in the United States:projections to the year 2000.JAMA. climate that condones the aggressive promotion of the only 1989;261:61-65. 3184 JAMA,December 11, 1991-Vol 266,No.22 Commentary Editorials Tobacco Marketing Profiteering From Children In 1946,the RJ Reynolds Tobacco Co advertised that"More not be far behind," and cowboys on the open range (the Doctors Smoke Camels Than Any Other Cigarette."t Such a Marlboro image)have long been a part of childhood fantasy. message conveyed that it was medically safe to smoke, and Undoubtedly, some adults also respond to these campaigns, the tobacco companies targeted adults to receive that mes- but continued avoidance of the effects on children is willful sage. In 1991, tobacco companies don't use medical spokes- negligence. persons to sell their products. Few physicians would attest The translation of this induced interest (induced by ad- to the safety of smoking, and tobacco companies are not as vertising)into the purchase of cigarettes by children is also interested in adults anymore. clear. Despite restrictions on the sale of tobacco to minors in most states,an investigation by the Inspector General of the See also pp 3145' 3149, and 3154. US Department of Health and Human Services revealed that these restrictions are seldom enforced.' In fact, the 1989 Surgeon General's report on smoking concluded that the num- In today's cigarette advertisements, physicians in white ber of legal restrictions on children's access to tobacco prod- coats have been replaced by cartoon animals in bright, pre- ucts had decreased since 1964.1 Vending machines,free sam- school colors. With straight faces, RJ Reynolds and its in- ples, and lack of enforcement combine to negate virtually all dustry colleagues report that they've chosen figures like Old of the existing restrictions. Joe Camel because they believe that such figures will appeal Compounding the media encouragement and the easy ac- to adult smokers and encourage them to change brands.With cessibility of tobacco is the naivet6 of children. While adults the same straight faces, they will likely express shock that might read warning labels or might have experienced the children respond to this campaign by taking up smoking. It's death of a loved one from lung cancer or emphysema, few simple:first they wanted us to believe that smoking was safe. children have. As described by Fischer et a1,2 many 3- to Now they want us to believe that their advertising campaigns 6-year-old children were just as captivated by Old Joe Camel don't cause people to start smoking and that cartoons don't promoting cigarettes as they were by Mickey Mouse in ad- appeal to children. vertisements for The Disney Channel. Knowledge of health Not only are these statements unbelievable, but they are consequences might dissuade adults from starting to smoke, also immoral and dangerous. In light of what is known about but long-known data show that most smokers start in their the deleterious health effects of tobacco, cigarette advertis- teenage years. The study by Fischer et al suggests that nic- ing is the moral equivalent of a national campaign to"Drive otine addiction may have its roots in much earlier childhood. Drunk—Just for the Fun of It." The tobacco and advertising industries disclaim any re- None of this should be surprising. The success of the to- sponsibility or liability for youth appeal,saying that smoking bacco industry is dependent on recruiting people who don't is a custom for those mature enough to make choices and that believe that smoking kills. Enticing children, Third World they do not want children to smoke. Data reported in this populations,and disadvantaged members of our own society issue of THE JOURNAL should sound a warning to parents to smoke is the only way for tobacco companies to make up that exposure to cigarette advertising can place a child at risk for the number of smokers who quit or die. We should be of smoking, whatever the tobacco industry may say. ' especially alarmed at the tobacco industry's effort to recruit The real question is how to respond to youths'being tar- children to nicotine addiction. Young people are encouraged geted by tobacco advertising campaigns. Strict controls on to buy cigarettes, young people are able to buy cigarettes, advertising would be an appropriate first step, but tough but young people do not have the information and experience advertising control measures have found little else but rhe- to recognize the dangers of smoking cigarettes. Such a mar- torical support from within the Bush administration. Last ket niche is an advertiser's dream. year,legislation introduced by Rep Mike Synar(D,Okla)and The tobacco companies'success at targeting young people I would have prohibited the cartoonlike, youth-oriented im- is apparent from data reported in this issue of THE JOURNAL. agery characterized by the Camel advertising campaign. Re- Old Joe Camel has demonstrated appeal and recognition among grettably,that legislation failed to win a majority of votes in youth,"the new Kool-brand penguin,which was introduced the Subcommittee on Health and the Environment.' In ad- recently by the Brown&Williamson Tobacco Company,will dition to advertising restrictions, prohibiting vending ma- chine sales and free samples and strictly enforcing minimum- From the Subcommittee on Health and the Environment,US House of Represen- age laws are steps that can be accomplished at the state and tatives,Washington,DC.Mr Waxman is a Democratic representative from California. local levels.'Finally,federal regulators need to wake up and Reprint requests to Subcommittee on Health and the Environment,US House of put a stop to continuing violations of federal restrictions Representatives,300 New Jersey Ave SE,512 House Annex 1,Washington, DC 20515(Mr Waxman). against tobacco advertising on television. To its credit, the JAMA,December 11, 1991—Vol 266, No.22 Editorials 3185 Federal Trade Commission recently took action to stop the 1. Life Magazine.December 23, 1946. 2. Fischer PM, Schwartz MP, Richards JW Jr, Goldstein A0, Rojas TH. advertising of Red Man Chewing Tobacco through its spon- Brand logo recognition by 3-to 6-year-old children:Mickey Mouse and Old Joe sorship of televised sporting events.10 Ironically,the Justice the Camel.JAMA. 1991;266:3145-3148. Department, which has had jurisdiction over cigarette ad- 3. DiFranza JR,Richards JW Jr,Paulman PM,et al.RJR Nabisco's cartoon vertisin for 20 ears, has et t0 take similar action. camel promotes Camel cigarettes to children.JAMA. 1991;266:3149-3153. g y y 4. Pierce JP,Gilpin E,Burns DM,et al.Does tobacco advertising target young During the Nixon administration, Attorney General people to start smoking?evidence from California.JAMA.1991;266:3154-3158. John Mitchell was credited with the remark: "You'd be 5. Farhi P. Kool's penguin draws health officials' heat. Washington Post. October 23, 1991;sect C1:7. better informed if instead of listening to what we say, you 6. Hearings Before the Subcommittee on Health and the E tironment,Et- watch what we do."" This is good advice for evaluating ergs and Commerce Committee, 101st Cong,2nd Sess(19%)(testimony of the tobacco industry. If exposure to cigarette advertising Michael F.Mangano,Deputy Inspector General,Dept of Health and Human Services). is a risk factor for disease, it is incumbent on the public 7. Surgeon General.Reducing the Health Consequences of Smoking:25 Fears and elected officials to deal with it as we would the vector of Progress:A Report of the Surgeon General.Washington,DC:US Dept of Health and Human Services;1989.Publication CDC 89-8411. of any other pathogen. 8. HR 5041,Tobacco Control and Health Protection Act,introduced on June If we can protect our young people from nicotine addiction, 14,1990,and considered by the Subcommittee on Health and the Environment they are unlikely to die of smoking-related diseases as adults. of the House of Representative's Energy and Commerce Committee on August 3 and September 11, 1990. Tobacco use by young people should be treated with the same 9. Jason LA,Ji PY,Anes MD,Birkhead SH.Active enforcement of cigarette level of intolerance as the use of any other addictive drug. It controllaws_inthe prevention ofcigarettesales tominors.JAMA.1991;266:3159- is sobering to me as a legislator and recent grandfather that 31sT 10. FTC Settles Charges Against Pinkerton Tobacco Co in First Case h:- 6-year-olds now find Old Joe Camel as easy to recognize as volving Alleged Violations of the Statutory Ban on TV Advertising q f Snwke- Mickey Mouse. less Tobacco.Washington,DC:Federal Trade Commission;October 29,1991. Press release. Rep Henry Waxman 11. Watch what we do. Washington Post.July 7, 1%9:A22. Editorial. Reducing Youth Access to Tobacco To continue its longstanding commitment that smoking is more than four fifths of smokers born since 1935 started not for young people,the tobacco industry. . . is supporting smoking before the age of 21 years.'These sobering statistics state legislation to make it tougher for young people to buy have prompted two sorts of commentary:a rallying cry by the cigarettes. We are also working with retailers for strict com- health community for tighter controls over the promotion, pliance with state laws prohibiting sales of cigarettes to mi- sale,and distribution of tobacco to minors,and a fresh spate nors.—The Tobacco Institute, 1991 poster. of lip service from the tobacco industry and others who profit [W]e believe that tobacco distributors can play an important from tobacco sales. role in reinforcing retailer efforts to comply with state laws prohibiting tobacco sales to minors.Indeed,NA TD currently See also pp 3159 and 3168. is preparing a program to assist retailers in this regard.— Peter Strauss,president and chief executive officer,National Association of Tobacco Distributors.' Concern about the problem of youth access to tobacco has ebbed and flowed through the years.At one time,each state We are excited about NACS'responsible tobacco retailing had legislation prohibiting the sale or gift of cigarettes to program and will urge our members to adopt it as well as minors. Most of these laws were adopted on moral grounds company policies against illegal tobacco sales. We believe during the first half of this century. By 1964, when the first strongly that it will be successful.—Kerley LeBoeuf, pres- Surgeon General's report on smoking was released,48 states ident and chief executive officer, National Association of still had such laws on the books.5 During the next quarter Convenience Stores, letter to Louis W. Sullivan, MD, sec- century,that number fell to 43 as states continued to rescind retary,US Department of Health and Human Services,March these policies.This decline stands"in marked contrast to the 9, 1990. trends in virtually all other areas of smoking control policy," [Al 6-step self-regulation program designed to make sure as the Surgeon General noted in his 25th-anniversary report." cigarette vending machines are not a source of cigarettes for For most of the 1980s, little action occurred to stem the minors . . . has been vigorously followed by all members of distribution of tobacco to minors except for sporadic local our industry since it was first adopted in 1962.—Richard W. bans on free tobacco samples.I A turning point came with the Funk,chief counsel, National Automatic Merchandising As- publication of the Surgeon General's 1988 report on nicotine sociation.2 addiction.' In his preface to that report, former Surgeon General C. Everett Koop, MD, asked why we, as a society, More than 1 million Americans become new smokers each allow the sale of cigarettes through vending machines and the year—or about 3000 each day.'Most of these new smokers distribution of free tobacco samples through the mail and on are recruited from the ranks of children and adolescents; public property.Tobacco sales,he suggested,should be con- trolled at least as tightly as the sale of alcoholic beverages. His message, and the heavy press coverage it received, led From the Michigan Department Public Health,Lansing. to renewed interest in this issue among researchers and Reprint requests to Michigan Department of Public Health,3423 N Logan St,PO g Box 30195,Lansing,MI 48909(Dr Davis). tobacco control advocates. 3186 JAMA, December 11, 1991—Vol 266,No,22 Editorials Evidence of Easy Access Because most teenage smokers buy cigarettes over the Local"sting"operations have documented that teenagers counter, effective strategies must extend beyond vending have little difficulty in buying cigarettes where such sales are machines.Increasing merchant awareness of existing tobacco illegal.&14 National data confirm these findings. The 1987 access laws is an important first step. However, as several National Adolescent Student Health Survey15 showed that studies have shown, including that of Feighery et al in this 79%of eighth-graders and 92%of 10th-graders considered it issue of JAMA, merchant education alone will not lead to to be "very easy" or"fairly easy"to get cigarettes. long-term compliance.11,23.24 In 1990, the Inspector General of the US Department of During the past year, East Lansing, Mich, and five com- Health and Human Services surveyed the law enforcement munities in Minnesota adopted ordinances that require single- agencies and health departments of the 44 states that had package tobacco sales to be conducted from behind the counter. tobacco access laws at the time.Only five states could provide These laws grew out of concern that self-service promotional any statistical information on vendor violations of these laws, displays make tobacco packages inviting to young people and reporting a total of 32 violations for all of 1989.16 That same easy to steal(St Paul Pioneer Press. October 17, 1991:2D). year,almost 1 billion packs of cigarettes were sold to persons Secretary of Health and Human Services Louis Sullivan, under age 18 years.'' MD has proposed a comprehensive model bill"that would(1) To determine where youth obtain cigarettes,the Office on create a retail licensing system for tobacco similar to that Smoking and Health at the Centers for Disease Control in- used for alcoholic beverages; (2)rely primarily on state-ad- cluded several questions on this subject in its 1989 Teenage ministered civil penalties to avoid the costs and delays of the Attitudes and Practices Survey.A nationally representative court system;(3)establish a graduated schedule of penalties, sample of adolescents aged 12 to 18 years were interviewed.18 including monetary fines and license suspensions;(4)set the Among those who had smoked in the last 30 days,63%said minimum age of legal purchase of tobacco at 19 years;and(5) they usually bought their own cigarettes; these purchasers ban cigarette vending machines.Licensure provides a source (n=874) were asked how frequently they bought their cig- of revenue through license fees to fund enforcement(eg,the arettes from "a small store, such as a 7-Eleven or a gas hiring of inspectors to conduct sting operations) and allows station,"from"a large store,such as a supermarket,"or from for meaningful penalties for violations(license suspension or a vending machine. Small stores were the most common site revocation). of purchase(used"often"or"sometimes"by 86%), followed Several communities around the country have adopted a by large stores(51%)and vending machines(13%) (Gary A' tobacco licensing scheme, including King County, Washing- Giovino, PhD, Centers for Disease Control, oral communi- ton; East Lansing and Marquette County, Michigan; India- cation, October 25, 1991). napolis,Ind;Belmont,Brookline,and Leominster,Mass;and Similar results were obtained from a 1989 study commis- Woodridge, Ill. A law passed recently by Vermont incorpo- sioned by the National Automatic Merchandising Associa- rates licensure of tobacco retailers into its alcohol licensing tion, which represents more than 1000 companies that sell system and provides for administrative enforcement by the cigarettes through vending machines. Using the"mall-inter- liquor control board. cept"technique in 20 cities,the study assessed the cigarette- The article by Jason et a126 that also appears in this issue purchasing behavior of 1015 smokers between the ages of 13 of JAMA provides the first convincing evidence that these and 17 years. Eighty-two percent of respondents said they laws can work.With continued merchant education,repeated obtained their cigarettes by purchasing them. Of these,85% sting operations, and assessment of penalties (including li- purchased cigarettes "often"or"occasionally"from a sales- cense suspensions), the Woodridge, Ill, ordinance has re- person or store clerk, compared with 22%from vending ma- sulted in the virtual elimination of cigarette sales to minors. chines.Among those who bought cigarettes over the counter, Moreover,surveys of Woodridge students suggest that smok- 67%purchased them from convenience stores,55%from gas ing initiation may have declined after passage of the ordi- stations,39%from grocery stores and/or supermarkets,and nance. 22%from drug stores.19 The study by Feighery et a124 highlights the difficulty in using the criminal justice system to curtail tobacco sales to minors. Judges dismissed or reduced the fines imposed on Legislative Activity store clerks who sold cigarettes to minors. This leniency The findings of these studies have stimulated legislative supports administrative enforcement using civil penalties. action throughout the country. The number of states that Unfortunately, the study does not tell us how judges might have tobacco access laws has rebounded during the last few treat managers or owners of stores that sell tobacco to minors years from 43 to 46,and some existing laws have been made because proprietors have not been held liable under Califor- more restrictive.4,20-22 nia law. In October 1989,White Bear Lake,Minn(population 25 000), Both studies relied on police officers for enforcement. It became the first community in the nation to enact a complete seems unlikely, however, that law enforcement authorities ban on cigarette vending machines.During the next 7 months, can be persuaded to assume this responsibility throughout another 36 communities in the state(including Minneapolis, the country. A more practical approach is to vest enforce- St Paul,and Duluth)followed with vending machine restric- ment authority with local and state health departments,sim- tions of their own, half of which were total bans.14 Eight ilar to the roles they play in restaurant inspections. states21 and many cities outside Minnesota(eg, New York, The future of tobacco access legislation is promising. In- NY;Chicago, Ill; San Francisco, Calif;Houston, Tex;Pitts- terest in this issue is growing substantially.Two major foun- burgh, Pa;and Washington, DC)have adopted vending ma- dations—the Thrasher Research Fund and the Robert Wood chine controls. Johnson Foundation—are now supporting the organization JAMA, December 11, 1991—Vol 266, No.22 Editorials 3187 1 that has led the campaign to prevent the sale of tobacco to 4. Surgeon General.Reducing the Health Consequences of Smoking:25 Years minors Stop Teenage Addiction to Tobacco(STAT)(based in of Progress:AReport of the Surgeon General.Atlanta,Ga:Centers for Disease Control; 1989. US Dept of Health and Human Services publication CDC Springfield, Mass). 89-8411. When legislative controls are inadequate, an alternative 5. Hawkins CH.Legal restrictions on minors'smoking.Am J Public Health. strategy is litigation. In 1987 two young people sued a chain 1964'54`1741-1744' 6. Davis RM,Jason LA.The distribution of free cigarette samples to minors. of 109 convenience stores based in Waltham, Mass, alleging Am J Prey Med. 1988;4:21-26. that one of the chain stores sold them cigarettes as minors and 7. Surgeon General,The Health Consequences of Smoking:Nicotine Addiction: AReport of the Surgeon General.Atlanta,Ga:Centers for Disease Control;1988. thus aided in their becoming addicted(Kyte v Store 24). In a US Dept of Health and Human Services publication CDC 88-8406. settlement, the chain agreed to demand proof of age from 8. DiFranza JR, Norwood BD, Garner DW,Tye JB. Legislative efforts to anyone not clearly of legal age before selling them tobacco protect children from tobacco.JAMA. 1987;257:3387-3389. 9. Altman DG,Foster V,Rasenick-Douss L,Tye JB.Reducing the illegal sale (Wall Street Journal. June 18, 1991:B1, B11). of cigarettes to minors.JAMA. 1989;261:80-83. The ultimate question is whether reducing youth access to 10. Hoppock KC,Houston TP.Availability of tobacco products to minors.J Fam Pract. 1990;30:174-176. tobacco will lower future smoking prevalence and smoking- 11. Skretny MT, Cummings M,Sciandra R,Marshall J.An intervention to related disease. In theory,it should.Making it more difficult reduce the sale of cigarettes to minors.N Y State J Med. 1990;90:54-55. for young people to obtain tobacco should impede the tran- 12. Centers for Disease Control.Cigarette sales to minors-Colorado,1989. MMWR. 1990;39:794-795,801. sition from experimentation to regular use.Delayed onset of 13. Hearings Before the Subcommittee on Transportation and Hazardous smoking is associated with decreased mortality as well as Substances,House of Representatives Committee on Energy and Commerce, increased attempts to quit and smoking cessation during the 101th Cong,1st Sess(July 25 and Sept 13,1989),serial No.101-85,pp 68-74 atmp (testimony of Ronald M.Davis,director,Office on Smoking and Health). school years. 14. Forster JL,Hourigan M,Weigum J.The Movement to Restrict Children's Some have argued that tobacco access laws may be coun- Access to Tobacco in Minnesota. Presentation to the Federal Interagency to roductive because children may more aggressively Committee on Smoking and Health,Washington,DC,May 31,1990.Atlanta, rP Y gg Y P Ga:Centers for Disease Control. sue a "forbidden fruit."27 This argument suggests that we 15. Office of Disease Prevention and Health Promotion. The National Ado- should make all forbidden fruits-cars, beer, "adult"video- lescent Student Health Survey:A Report on the Health of America's Youth. Oakland,Calif:Third Party Publishing Co;1989. tapes-freely available to children. Others have pointed out 16. Office of Inspector General.Youth Access to Cigarettes.Washington,DC: that the purchase of alcoholic beverages by teenagers still US Dept of Health and Human Services;1990. occurs (or is increasing) despite licensure of retailers and 17. DiFranza JR,Tye JB.Who profits from tobacco sales to children?JAMA. g p 1990;263:2784-2787. other controls;however, lax enforcement of these policies2l 18. Centers for Disease Control. Cigarette smoking among youth-United undercuts the use of that argument against tobacco access States, 1989. MMWR. 1991;40:712-715. 19. Findings for the Study of Teenage Cigarette Smoking and Purchase Be- laWS. hatior.Chicago,Ill:Response Research Inc;1989. More research is needed to fully understand the effects of 20. Centers for Disease Control. State laws restricting minors'access to tobacco access laws.In the meantime,we should proceed"full tobacco.MMWR. 1990;39:349-353. 21. Centers for Disease Control.State tobacco prevention and control activ- steam ahead"with their adoption and enforcement. Preven- ities: results of the 1989-1990 Association of State and Territorial Health tion of smoking initiation requires a multifaceted approach Officials(ASTHO)Survey,final report.MMWR.1991:40(suppl RR-11):141. 22• Enrolled Act No. 34 (Senate) of the 51st Legislature of the State of that addresses the supply of tobacco,the demand for it, and Wyoming, 1991 General Session, Wyoming Statutes 14-3-301 through 14-3- the environment in which it is used. A smoke-free environ- 306. ment in schools helps to inculcate in our children the per- 23- Altman DG,Rasenick-Douss L,Foster V,Tye JB.Sustained effects of an educational program to reduce sales of cigarettes to minors. Am J Public ception of nonsmoking as the social norm. School health ed- Health. 1991;81:891-893. ucation and tobacco advertising restrictions are intended to 24. Feighery E,Altman DG,Shaffer G.The effects of combining education reduce the demand for tobacco among youth.Adding supply- and enforcement to reduce tobacco sales to minors:a study of four northern California communities.JAMA. 1991;266:3168-3171. side strategies to our armamentarium will increase our like- 25. Model Sale of Tobacco Products to Minors Control Act:A Model Law lih00d for success. Recommended for Adoption by States and Localities to Prevent the Sale of Ronald M. Davis, MD Tobacco Products to Minors.Washington,DC:US Dept of Health and Human Services;1990. 1. Hearing Before the Senate Committee on Labor and Human Resources, 26. Jason LA,Ji PY,Anes MD,$irkhead SH.Active enforcement of cigarette 101st Cong,2nd Sess(April 3, 1990), part 2, pp 67-74(testimony of Peter control laws in the prevention ofcigarettesales tominors.JAMA.1991;266:3159- Strauss,president,National Association of Tobacco Distributors). 3161. 2. Hearings Before the Subcommittee on Transportation and Hazardous Sub- 27. It Can Be Done:A Smoke-Free Europe:Report of the First European stances, House of Representatives Committee on Energy and Commerce, Conference on Tobacco Policy, Madrid,.7-11 November 1988. Copenhagen, 101st Cong,1st Sess(July 25 and Sept 13,1989),serial No.101-85,pp 237-296 Denmark:World Health Organization Regional Office for Europe;1990.WHO (testimony of Richard W.Funk,chief counsel,National Automatic Merchan- regional publications,European series,No.30. dising Association). 28. Office of Inspector General.Youth and Alcohol:Laws and Enforcement: 3. Pierce JP,Fiore MC,Novotny TE,et al.Trends in cigarette smoking in the Is the 21-Year-Old Drinking Age a Myth?Washington,DC:US Dept of Health United States:projections to the year 2000.JAMA. 1989;261:61-65. and Human Services;1991. Making Smoking Prevention a Reality Since the inception of the antitobacco movement in the late delayed an estimated three quarters of a million deaths from 1950s,national,state,and local public health campaigns have tobacco-related disease.' Although many lives have been saved,one in every five deaths in 1988 was smoking-related, From the Office of the Director,Centers for Disease Control,Atlanta,Ga. and smoking directly caused 434 000 deaths in 1988.23 It is Reprint requests to the Office on Smoking and Health,National Center for Chronic amazing to think that each and every day, 1200 Americans- Disease Prevention and Health Promotion,Centers for Disease Control,Mail Stop K-50,1600 Clifton Rd NE,Atlanta,GA 30333(Public Information Branch). 50 each hour-lose their lives from preventable smoking- 3188 JAMA,December 11,1991-Vol 266,No.22 Editorials Qo r ` relafed diseases such as lung cancer,heart disease,and chronic Numerous economic studies have demonstrated an inverse obstructive lung disease.4 relationship between the price of tobacco products and to- Tobacco,the plague of our time,brings forth the need for bacco consumption.' Increasing the price of cigarettes may us to band together-together,we can ultimately save thou- reduce the number of adolescent as well as adult smokers in sands of lives now held hostage by smoking.5 As a community this country; increasing state excise taxes would have the of health professionals, our mission is to support the year dual advantage of improving the health of citizens and gen- 2000 health objectives6 and to acknowledge tobacco as the erating state revenues. significant and far-reaching public health problem portrayed We cannot reach the year 2000 objectives without accel- in the objectives. If we are to reduce cigarette smoking to a erating cessation efforts.Physicians,dentists,and other health prevalence of no more than 15%by the year 2000,6 we will care providers play a crucial role in this strategy. All phy- need to double the current rate of decline in smoking prev- sicians and dentists should(1)ask all patients if they smoke, alence. To meet this objective, environmental, social, and (2) advise smoking patients to stop, (3) assist smoking pa- behavioral factors must all be included as part of any national tients with quitting (eg, set a target date), and (4) provide strategies. Any approach to solving this problem must rec- follow-up support.12 One key to prevention is to recommend ognize that these factors are intertwined. cessation for all patients who smoke,not just those exhibiting Even though 42 million Americans have quit smoking, smoking-related symptoms. nearly 50 million adults(28%of the population)in the United Making smoking prevention a reality will not be an easy States continue to smoke.'Nicotine addiction is difficult to task, as there are many obstacles to overcome. We must overcome,'and the decline in smoking has been substantially intensify our efforts and broaden our partnerships to cre- slower among women than men. Smoking prevalence re- ate a more efficient networking system. Many diverse mains disproportionately high among minorities, blue-collar organizations in our society have the potential to bring workers, and people with less education.' In addition, the about sweeping changes in public policies with regard to rate of smoking during the 1980s among our nation's teen- tobacco and health education efforts, minors' access to agers has not changed substantially compared with the rate tobacco products, smoking prevention programs in in the late 1970s and early 1980s.'The fact that about 80%of schools, tobacco advertising, and smoking cessation. Our all adult smokers become regular smokers before their 21st success in effectively controlling tobacco-a highly lethal birthday underscores the magnitude and pervasiveness of consumer product-requires a very real commitment by the problem. every one of us. The future health and well-being of our children and our William L. Roper, MD, MPH nation depend upon our ability to intensify efforts to dis- courage young people from starting to smoke. The coordi- 1. Department of Health and Human Services.Reducing the Health Conse- nation of supportive tobacco prevention and cessation pro- quences of Smoking:25 Years of Progress:A Report of the Surgeon General. Atlanta,Ga:Centers for Disease Control,National Center for Chronic Disease grams among health care agencies, businesses, p011Cymak- Prevention and Health Promotion,Office on Smoking and Health;1989. US ers, educational institutions, and other influential groups is Dept of Health and Human Services publication CDC 89-8411. also imperative.Once effective school-based prevention pro- 2. Centers for Disease Control.Smoking-attributable mortality and years of ams have been laced in all of our nation's schools potential life lost-United States, 1988.MMWR. 1991;40:62-63,69-71. P , young g 3. National Center for Health Statistics. Advance report of final mortality people will be better educated about the negative health statistics, 1988.Monthly Vital Stat Rep. 1990;39(7,suppl):2. consequences of tobacco use and about ways they can resist 4. Sullivan LW.Statement on the ASSIST Federal Smoking Control Project. Washington,DC:October 4, 1991;press release. the pressure to experiment with smoking. 5. Roper WL.Statement on the 1990 Surgeon General's Report:The Health Tobacco control policies represent important strategies for Benefits of Smoking Cessation.Washington,DC:September 25,1990;press limiting tobacco use and, therefore, preventing the onset of conference. 6. Department of Health and Human Services.Healthy People 2000:National smoking. One such policy prohibits the sale and distribution Health Promotion and Disease Prevention Objectives.Washington,DC:US of tobacco products to those younger than age 19.6 Other Dept of Health and Human Services; 1991. US Dept of Health and Human Services publication PH5 91-50212. components for reducing tobacco use-such as banning the 7. Centers for Disease Control. Cigarette smoking among adults-United use of vending machines, creating a licensing system for States, 1988.MMWR. 1991;40:757-759,765. vendors,and increasingthe frequency of enforcement-should 8. Department of Health and Human Services. The Health Consequences of Smoking:Nicotine Addiction:A Report of the Surgeon General.Atlanta,Ga: be included in minors'access legislation.The US Department Centers for Disease Control, Center for Health Promotion and Education, of Health and Human Services has urged the adoption of a Office on Smoking and Health;1988.US Dept of Health and Human Services model law, the Model Sale of Tobacco Products to Minors' gu Centers for Di ease6Control. Cigarette smoking among youth-United Control Act,'o in all 50 states.' States,1989.MMWR. 1991;40:712-715. Elimination or restriction of advertisements to which our 10. Department of Health and Human Services.Model Sale of Tobacco Prod- ucts to Minors Control Act:A Model Law Recommended for Adoption by young people are likely to be exposed is an objective worth States or Localities to Prevent the Sale of Tobacco Products to Minors.Wash- considerable effort.According to Louis W. Sullivan, MD,sec- ington,DC:US Dept of Health and Human Services;1990. retary of health and human services,"access to cigarettes is still 11. Sullivan LW.Youth smoking rates.Washington,DC:US Dept of Health and Human Services;October 17, 1991;press release. too easy for those who are under the legal age,and the tobacco 12. Glynn TJ, Manley MW. How to Help Your Patients Stop Smoking:A industry continues to advertise in a Shameless manner to sus- National cancer Institute Manual for Physicians.Washington,DC:National »1i Institutes of Health,National Cancer Institute,Smoking,Tobacco,and Can- ceptibleyoungpeople. Tobacco advertisements cloak a known cer Program;1990.US Dept of Health and Human Services publication NIH deleterious product in images that appeal to youth. 90-3064. JAMA,December 11, 1991-Vol 266, No.22 Editorials 3189 t�11y dis&vered ACTH-dependent adrenal adeno- and high-dose dexamethasone suppression.,Ann ing's syndrome:A prospective study of diagnostic �a presenting as`pre-Cushing's syndrome:Acta Intern Med 1982;9721-26. methods.Am J Med 1973;55:621-628. Endocrinol 1986;111:89-92. 13. Abboud CF:Endocrine laboratory tests in the 16. Gross MD,Wilton GP,Shapiro B,et al:Func- 10. Beyer HS, Doe RP:Cortisol secretion by an diagnosis of anterior pituitary disease in the adult, tional and scintigraphic evaluation of the silent ad- incidentally discovered nonfunctional adrenal ade- in Humburger HA, Batsakis JG (eds): Clinical renal mass.JNuct Med 1987;28:1401-1407. noma.JClin Endocrinol Metab 1986;62:1317-1321. Laboratory Annual, 1985. East Norwalk, Conn, 17. Gross MD,Shapiro B,Bouffard JA,et al:Dis- 11. Nugent CA,Nichols T,Tyler FH:Diagnosis of Appleton-Century-Croft,1985,vol 4,pp 199-254. tinguishing benign from malignant euadrenal Cushing's syndrome: Single-dose dexamethasone 14. Streetin DHP,Anderson GH Jr,Dalakos TG, masses.Ann Intern Med 1988;109:613-618. suppression test. Arch Intern Med 1965;116:172- et al:Normal and abnormal function of the hypotha- 18. Crapo L:Cushing's syndrome:A review of di- 176. lamio-pituitary-adrenocortical system in man.En- agnostic tests.Metabolism 1979;28:955.977. 12. Ashcraft MW,VanHerle AJ,Vener SL,et al: docrinol Rer 1984;5:371394. Serum cortisol in Cushing's syndrome after low- 15. Eddy RL,Jones AL,Gilliland PF,et al:Cush- Comparing Tobacco Cigarette Dependence With Other Drug Dependencies Greater or Equal `Difficulty Quitting' and `Urges to Use,' but Less Pleasure From Cigarettes Lynn T.Kozlowski,PhD;D.Adrian Wilkinson,DPhil;Wayne Skinner,MSW; Carl Kent,MSW;Tom Franklin,MSW;Marilyn Pope About 1000 persons seeking treatment for alcohol or drug dependence were term effects of nicotine, as usually con- asked, relative to cigarettes, about the difficulty of quitting the use of the sumed.' Most alcohol and drug users substance for which they were seeking treatment,the strength of their strongest undergoing treatment(80%to 95%)also urges to use,and the pleasure they derived from use. Fifty-seven percent said smoke cigarettes! Although clinical that cigarettes would be harder to quit using than their problem substance. lore in our institution holds that ciga- These ratings were related to the level of cigarette dependence and the per- rette dependence is considered by users ceived difficulty of quitting the use of the problem substance. The alcohol- of classic drugs as a very tenacious de endent persons were about four times more like) than the drug-dependent addiction,we could find little sTstemat- p p yis data regarding the issue. To evalu- persons to say that their strongest urges for cigarettes were at least as great as ate the addictiveness of cigarettes and their strongest urges for their problem substance. Cigarettes were generally to explore similarities and differences rated as less pleasurable than alcohol or other drugs.Thus,experiential experts among major forms of substance use,we on dependence judge cigarette dependence as at least as"addictive"as other asked"expert"alcohol and drug users to drug use, but not as pleasurable, indicating important similarities and differ- compare cigarettes with the use of the ences between cigarette dependence and other forms of dependence on substances for which they had sought psychoactive substances. treatment. (JAMA 1989;261:89&901) Alcohol and drug users who were seeking treatment for substance depen- dence were asked to rate(1)the relative difficulty of quitting the use of ciga- From the Clinical Institute,Addiction Research Foun- DESPITE scientific evidence that See also p 867. dation.Toronto(Drs Kozlowski and Wilkinson,Messrs nicotine is highly addictive,'the similar- Skinner.Kent.and Franklin,and Ms Pope);the Depart- ity between widely available, legal, ment of Preventive Medicine end Biostatistics,Universi- "domesticated" cigarettes and "hard" rettes vs their main problem substance, ty of Toronto(Dr Kozlowski); and the Department of difficult t0 accept.Psychology.York University,Toronto(Dr Wilkinson). can . Some (2)the relative strength of their stron- The opinions are our own and not necessarily those of of this skepticism comes from exagger- gest urges for cigarettes and for their the Clinical Institute,Addiction Research Foundation. ated but popular notions of the addict- main problem substance,and(3)the rel- Reprint requests to the Clinical Institute,Addiction s iveness of heroin and other drugs"and ative pleasure derived from cigarettes Research Foundation,33 Russell St,Toronto,Ontario, g g Canada M5S 2S1(Dr Kozlowski). also from the relatively tame short- and from their main problem substance. 898 JAMA,February 10,1989—Vol 261,No.6 Cigarette Dependence—Kozlowski at al I normal.Eight months after removal of a lateral lesions, and prominent asym- patients with obesity,acute and chroni�cc benign adrenal adenoma,the remaining metrical uptake(R>L)was noted in the stress, severe depression, and alcohol- adrenal gland was visualized by scintig- patient with bilateral masses. There ism and in those taking diphenylhydan- raphy and the abnormal test results had were no abnormalities of plasma or toin or potent estrogens. We reviewed reverted to normal. Although adrenal urine steroid hormone measurements, the medical records of 23 patients with insufficiency was not reported by these and overnight dexamethasone suppres- incidentally discovered adrenal masses. investigators,it is not clear whether the sion tests were normal. The authors Nine of the patients had overnight dexa- patient was treated with adrenal ste- performed bilateral adrenal vein cathe- methasone suppression tests. In addi- roids perioperatively. terization and demonstrated that corti- tion to case 2 reported herein,two pa- In a series of 58 patients with adreno- sol concentrations were greatest from tients had abnormal results. One cortical tumors,Bertagna and Orth'de- the side of the mass lesion in those with patient had adrenal cancer and elevated scribe two patients who secreted corti- unilateral masses and from the larger of basal steroid levels. The other patient sol in amounts insufficient to produce the two adrenal glands in the patient /probably had a false-positive test re- signs and symptoms of hypercortisolism with bilateral adrenal lesions. suit. The study was normal in the re- and who had normal levels of urinary In a subsequent article," these a maining six patients. The overnight steroids.'The first patient did not un- thors performed NP-59 scintigraphy/n dexamethasone suppression test has dergo suppression with dexamethasone 119 consecutive euadrenal patients yvlth not,to our knowledge,been systemati- administration but increased steroid ex- incidentally discovered adrenal n*sses. cally studied in patients with incidental- cretion no ally after giving metyra- In 76 patients,NP-59 uptake latgi alined ly discovered adrenal masses. The ex- pone and sponded dramatically to to the abnormal adrenal mass/seen on tensive experience, however, in normal ACTH infusio*The second patient was computed tomographic scans: In all of patients, patients with Cushing's syn- partially suppres$ible with dexametha- these patients,a diagnosis at benign ad- drome,"and case 2 herein suggests that sone administrat16 and responded nor- enoma was made by nee.}e biopsy, ad- the overnight dexamethasone suppres- mally to ACTH i ion. Her normal renalectomy, or repeated computed sion test is the most practical method for morning plasma cortin level was nor- tomographic scans. / detecting autonomous cortisol secretion mal,but there was no diiiipal variation. In summary, our tvi o patients indi- and, thus, for identifying patients at Bogner and coworkers'discovered an cate that apparently nonfunctional ad- risk for postoperative adrenal insuffi- adrenal tumor accidentally In patient renal adenomas may secrete cortisol. ciency. False-positive test results may without clinical signs of Cushy syn- The studies by Gross et al"' suggest result in perioperative steroid therapy, drome. The patient had an increased that this situati,D i'may be more common but this brief treatment is not likely to urinary cortisol excretion rate. Atthan previously" appreciated. Variable end in serious complications.Not all ad- though plasma and urinary cortisol leve amounts and durations of cortisol secre- renal masses are treated by operation. els were suppressed by dexamethasone tion probably-account for the difference Nonfunctional tumors below the recom- to the low-normal range, the authors test results observed in patients har- mended size for removal usually are fol- suggest that autonomous cortisol secre- bo ' g such tumors. Cortisol secretion lowed up by serial computed tomo- tion.probably was present. The patient by t tumor probably initially results grams.' If the natural history of was treated perioperatively with adre- in loss q diurnal variation.As ACTH is cortisol-secreting adenomas is to in- nal steroids. Four weeks after surgery, suppress to low levels, the adrenal crease progressively in function,follow- normal plasma cortisol levels without scan beco s characteristic and the up hormonal testing also can be substitution therapy were recorded. dexamethaso suppression test be- important. Beyer and Doe'provide detailed en- come progress' ely abnormal.&7.16 A This study was supported by a grant from the docrine studies in a middle-aged man cortisol-secreting enoma is respon- Gundexsen Medial Foundation,La Crosse,Wis. with an incidentally discovered adrenal sive to both exoge us and endoge- The authors thank Patti Bieber for expert secre- adenoma.The patient,who had no clini- nous ACTH.'Thus,if adrenal adeno- tarial assistance. cal manifestations of Cushing's syn- ma only recently sec ted enough drome,had normal plasma and urinary cortisol to replace normal enocorti- Rewences cortisol excretion but displayed abnor- cal secretion,a normal respon to me- 1. Prinz RA,Brooks MH,Churchill R,et al:Inci- mal circadian rhythm and failed to sup-,' tyrapone administration could ob- dental asymptomatic adrenal masses detected by press cortisol excretion with both low= served!Nonetheless, as demonst d computed tomographic scanning:Is operation re- and high-dose dexamethasone regi- b our second patient, adrenal ins 2.Gla JAMA WeymIM; as'7o1-703. g �Iee y p 2. Glazer HS,Weyman PJ,Segel SS,et al:Non- mens.Although his plasma ACTH le. 1 aency could occur when the suppressed ctioning adrenal masses: Incidental discoven- was suppressed, responses to adminis- adrenal cannot respond sufficiently af- computed tomography.AJR 1982;139:81-85. tration of metyrapone and injections of ter removal of the tumor. 3. peland PM:The incidentally discovered adre- nalACTH resulted in normal and ex � - Incidental adrenal adenomas are de- Ann Intern Med 1983; �g � 4. Bell A,Hussain S,Seltzer S say. SE,et al:Inci- ated hormonal responses, respectively. tected in approximately 0.6% of com- dentally di overed mass ofthe adrenal gland.Surg The patient did not experience adrenal puted tomograms of the abdomen.Since Gynecol obs 1986;163:203-208. insufficiency postoperatively despite basal levels of adrenal steroids may be 5. Beierwaltes ,Sturman MF,Ryo U:Imalpng the fact that steroid supplementation normal in patients with autonomous functional nodus¢of the adrenal glands with I- PP 19-iodocholesterot,JNucl Med 1974;15:246-251. was not provided. cortisol secretion,we agree with Cope- 6, Riga RA,Wahner HW,Spelsberg TC:Visual- Recently,Gross and coworkers18 per- land,'who suggests that patients with ization of nonfunctioning adrenal adenomas with formed iodine 131-60-iodomethyl-19- adrenal masses be screened with the iodocholesterol:Possible relationship to subcellular norcholesterol (NP-59) scintigraphy on overnight dexamethasone suppression distribution oftracer.JNuel Med 1978;19:458-463. seven patients with unilateral and one test. This test, extensive) used for di- 7. Charbonnel ad Charas JF,Ozanne s Does the Y corticoadrenal adenoma with `pre-Cushing's syn- patient with bilateral but asymmetric agnosing Cushing's syndrome,has a low drome'exist?J Nucl Med 1981;22:1059-1061. (R>L)adrenal masses discovered inci- frequency of false-negative results." S. Bertagna C,Orth DN:Clinical and laboratory dentally by computed tomography.The Furthermore,it is performed easily on aaddreenpatients with nocortical tumors admitted to gs and results of theraw in a8single medical NP-59 lateralized markedly to the side ambulatory patients and is inexpensive. center(1951 to 1978).Am JMed 1981;71:955-875. of the lesion in seven patients with uni- False-positive test results may occur in 9. Bogner v,Eggers U,Hensen S,et al:Inciden- JAMA,February 10,1989—Vol 261,No.6 Adrenal Insufficiency—Huiras et al 897 s Subjects and Methods much easier to give up cigarettes;2, a ness of fit to determine the most parsi- Three different surveys were con- little easier;3,about the same;4,a little monious model (set of variables) that ducted from 1986 through 1988 involv- harder; and 5, much harder [for logit could account for the variation in cell ing persons presenting for treatment at analysis (discussed later herein), three frequencies:high rather than lou P va1- the outpatient, young drug user, and scores were used: 1 indicates easier;2, ues are sought for such tests. Compari- drug therapy programs of the Clinical same; and 3, harder]); and "How hard sons are reported for the odds ratios Institute of the Addiction Research would it be for you to give up the drug or relating independent variables to a re- Foundation,7bronto.This clinic is a re- alcohol that brings you here?" (1 indi- coded,dependent variable dichotomy— search hospital offering a variety of cates not at all hard,to 4,very hard[for "cigarettes same or greater"vs'less." treatments for alcohol and drug prob- logit analysis, two scores were used: 1 The Wilcoxon signed-rank test was lems at no charge. (Following the Diag- indicates 1 to 3,and 2 indicates 4]). For used to compare within-subject ratings nostic and Statistical Manual of Men- the logit analysis,age was dichotomized of pleasure and urges; comparisons tal Disorders, third edition, revised,' as 30 years or under(1)and 31 years or across drug type used X'tests. Sample we consider both alcohol and"drugs"as older(2). sizes for particular analyses varied psychoactive substances but will contin- Comparing Strongest Urges: Sur- somewhat because of missing res- ue, with some reluctance, the common vey 2 (n=611).—The subjects were ponses. practice of maldng a distinction be- asked:"Compare your strongest desire Results tween nicotine, alcohol, and other or urge for cigarettes with your stron- drugs.)Across all three surveys,91%of gest desire or urge for your main Difficulty Quitting.—Seventy-four the subjects had smoked at one time and problem drug or alcohol. Which was percent reported that cigarettes would 84% were current, daily smokers. In stronger?(1)My strongest urge for cig- be at least as hard to give up as their surveys 2 and 3, we found that 76%of arettes was stronger than my strongest problem substance;57%said that ciga- the subjects were men, and their mean urge for my drug/alcohol. (2) My urge rettes would be harder to give up(Fig- age was 29 years (SEM=0.33 years). for cigarettes was about as strong. (3) ure). The multivariate logit analysis Our best estimate of the distribution of My urge for cigarettes was less strong." showed that the rated difficulty of giv- main problem substances comes from The next question was:"What problem ing up their problem substance and the survey 2 (see "Results" section). Our brings you here?(check one):(1)alcohol level of cigarette dependence had one-page, anonymous questionnaires (2) cannabis (hash, marijuana, etc) (3) strong"main effects"on ratings(L'for did not permit detailed characterization cocaine (4) heroin (5)prescription nar- goodness of fit[df=21,1.20;P=.55)and of drug habits; we could not match an- cotics(6)other(give name)." no interaction term was needed. Con- swers to full-assessment files or patient Comparing Urges and Pleasures: trolling for level of cigarette depen- records. Routine hospital records re- Survey 3(n=239).—The subjects were dence,those who rated the difficulty in garding this client population for 1987 asked:"Check each drug that you have giving up their problem substance as (n=1251) show that 44.5% have com- used regularly in the last six months:(1) less than "very hard" were more than pleted some secondary school, 18.7% alcohol,(2)cannabis(hash and marijua- twice as likely as those who said that have completed secondary school, and na),(3)cocaine,(4)heroin,(5)prescrip- their drug would be"very hard"to quit 27.9% have had at least some postsec- tion narcotics, (6) tranquilizers (Vali- to say that cigarettes would be the ondary education(7.3%have university tun) (7) cigarettes, (8) others (give "same"or"harder"to quit(odds ratio, degrees); 45.7% are single, 30.0% are names)." They then circled "the one 2.2; 95%confidence limits (CL), 1.5 to married,and 22.6%are divorced or sep- drug that has been the main reason for 3.3).Controlling for the difficulty of giv- arated; 49.9% are employed full-time, bringing you here" and were asked: ing up their problem substance,higher- 6.4%are employed part-time,32.2%are "Which drug that you have used regu- dependency smokers were nearly twice unemployed, 3.0% are students, and larly causes you the most pleasure when as likely as lower-dependency smokers 8.2%are not in the labor force. you use it?"(the same answer options as to say that cigarettes were at least as Questions and Procedures Common previously mentioned, omitting ciga- hard to give up as their problem drug to All Surveys.—Prospective clients rettes). Next, they were asked: "How (odds ratio,1.95;95%CL,1.3 to 2.9). were asked to fill out questionnaires does the pleasure from cigarettes com- Even among those who said that it during a break in intake assessment pare with the pleasure from this drug? would be "very hard" to give up their procedures. Less than 5%of the ques- (check here if you've never smoked):(1) problem substance (n=87), cigarettes tionnaires were spoiled and unusable; Cigarettes give me more pleasure than were rated as at least as hard to give up refusal to answer the survey was rare. this drug. (2)Cigarettes give about the by 57%(the same by 19.5%,a little hard- 7b assess heaviness of smoking,individ- same pleasure. (3)Cigarettes give less er by 13.8%, and much harder by uals were asked:"How many cigarettes pleasure."The final question asked for 24.1%). No information on type of pre- per day do you smoke at the present comparison of"strongest urges,"as in senting drug problem was available. time?"and "How long after waking do survey 2. 7b strictly compare the rat- Stronger of the Strongest Urges.— you wait before having your first ciga- ings of pleasure(from most pleasurable Many persons identified more than one rette of the day?"(in minutes). A ciga- substance)and urges(for main problem main problem substance,so a multiple- rette dependence score was calculated substance),the analyses were limited to drug—use category was needed. Main (moderate, <20 cigarettes per day or only those whose most pleasurable drug problem substances were alcohol the first cigarette of the day after more was also their main problem drug. (n=276), cannabis (n=81), cocaine than ten minutes of waking; and high, Analyses.—Multivariate logit yoga- (n=109), heroin (n=36), prescription the remaining smokers)." rithm of the odds) analyses were per- narcotics (n=45), other (n=45), and Difficulty Quitting: Survey 1 formed on surveys 1 and 210 to assess multiple(n=181).The logit analysis did (n=248).—The key questions were as joint effects of predictor variables on not support making a distinction in the follows:"Would it be easier or more dif- two trichotomized,dependent variables effects of several different types of drug ficult for you to quit smoking than to (relative difficulty in quitting and rela- use but did show that a model with give up the drug or alcohol that brings tive strength of urges).Likelihood-ratio `main effects"for drug type(alcohol vs you here for treatment?" (1 indicates X' tests (L') were conducted for good- drugs) and cigarette dependence pro- JAMA,February 10.1989—Vol 261,No.6 Cigarette Dependence—Kozlowski et al 899 1 50 sure was less strong. Of the cocaine t!5- ers,20.6%said their strongest urges for cigarettes were stronger than for co- caine,2.7%said their pleasure from cig- arettes was greater, 21.9% said that their urges were similar,1.4%said their 40 pleasure was similar, 57.5% said their urges were less strong, and 96% said H their pleasure was less strong. E ''Comment m Persons seeking treatment for drug CD 30 „ dependence tended to rate urges for cig- arettes and difficulty in quitting ciga- c) rettes as high as or higher than for their 0-0 main problem substances.This testifies Ln +1 powerfully to the"addictiveness"of cig- arettes.' Our classification of individ- uals as alcohol or drug dependent is probably meaningful, but certainly an c'n oversimplification, since "other drugs" 0 16 7 17 1 i differ importantly,P Y, on many sub- .......... stance-dependent pe are polydrug 10 dependent. Some terms commonly used to refer to dependence on various substances may obscure important differences.For example,a smoker who reports a"crav- ing"for a cigarette may be referring to a 0. different experience than some alcohol- Much Easier Little Easier Same Little Harder Much Harder ics.1 " For some drug users, the term Difficulty in Giving Up Cigarettes "craving" refers to any desires to use drugs,even weak ones,while for others Difficulty in giving up cigarettes t the term refers mainly to strong ones."Wsuch rug- comparison with giving up alcohol or drug use as reported by individuals e have tried to deal with h dto- n=248)presenting for treatment of alcohol or drug problems. drug and individual-to-individual ambi- guities by asking for direct comparisons of the strength of the urges associated vided an excellent fit with the observed ly than drug users to say that cigarettes with different substances. data W[df=10]=4.91,P=.90).Among were"stronger"vs"the same as"[95% Comparisons with cigarettes are use- the alcohol-dependent persons CL,1.2 to 2.5].) ful for indexing reactions to other (n=244), 50.2%(±7.9%, 95%CL)said Twenty-eight percent of the cocaine drugs,since cigarette dependence so of- their strongest cigarette urges were users said that cigarette urges were ten coexists with dependence on other stronger, 32.1% (±7%) said they were stronger;17%said they were about the substances. Our finding of low "plea- "about the same,"and 17.7% (±5.6%) same. Six percent of the heroin users sure" but high perceived "difficulty said they were less strong than for alto- reported cigarette urges were stronger; quitting" for cigarettes is consistent hol;in contrast,among the drug-depen- 31%said that the strongest urges were with other data that show that initial dent persons(n=467), the values were about the same. "liking" predicts subsequent addiction 25.4% (±4.8%), 27.2% (±4.9%), and Comparing Pleasure and Urges.— for many drugs but not for cigarettes." 47.4%(±5.7%),respectively. Although Overall, comparative ratings of cig- Hence,the pleasure derived from drug the young subjects were more likely to arettes and problem substances use may not be the most important de- use drugs than alcohol, no age effects (n=239)showed higher urges for ciga- terminant of difficulty in quitting or of were needed to account for the pattern rettes and less pleasure from cigarettes strength of urges to use; sheer "habit of comparative ratings. The odds ratio (Wilcoxon Z=8.32,P<.001).This same strength"may be important in this re- from the logit analysis showed that alto- pattern was found for alcohol users gard.'s The relative lack of pleasure hol users were about four times more (n=87) and cocaine users (n=73) from cigarettes may underlie some of likely than drug users to say that their (Z=4.86, P<.001, respectively). Both the skepticism about cigarette depen- strongest urges for cigarettes were at pleasure and urge ratings for cigarettes dente as a"true"addiction, and it may least the same as their strongest urges were lower for the cocaine users than mark important distinctions between for their problem drug(controlling for the alcohol users (X' [df=2]=30.4, nicotine dependence and some other cigarette dependence, odds ratio for P<.001; )e [df=2]=12.3, P<.005, re- forms of substance use. drug type,4.36;95%CL,2.96 to 6.43). spectively). Of the alcohol users,42.5% Our findings should not be construed Controlling for drug type,the more de- rated their strongest urges for ciga- as showing that smokers who are not pendent smokers were 2.2 times more rettes stronger, 13.8% said their plea- dependent on other drugs would per- likely than the less dependent smokers sure from cigarettes was greater,26.4% ceive quitting to be as difficult as judged to rate their urges for cigarettes as at said urges were similar, 28.7% said by our informants. "Clinical' samples least as strong (95% CL, 1.6 to 3.1). pleasure was similar, 31% said urges often show a greater dependence prob- (Alcohol users were 1.7 times more like- were less strong, and 57.5%said plea- lem than do more representative sam- 900 JAMA,February 10,1989—Vol 261,No.6 Cigarette Dependence—Kozlowski et al pies of substance users."" Most riga- portant questions on the Fagerstrom minutes. rette smokers give up smoking without scale appear to be the number of ciga- As cigarette smoking has become rar- formal treatment,"as do most heroin rettes smoked and the time to the first er in the general population, it seems users when giving up heroin.' General cigarette (within 30 minutes indicates not to have done so among other sub- knowledge of the former, but not the high dependence).8D For the general stance-dependent groups.`R'Even now, latter, fact may contribute to doubts public, the "within-30-minutes" cutoff patients who are heavy smokers should about the addictiveness of nicotine. In for dependent smoking may be ade- be recognized by physicians to be signif- both groups, repeated attempts to quit quate, but for persons dependent on icantly more likely than nonsmokers tend to pay off in the end and,therefore, other substances, we find that, even (especially than"never smokers")to be deserve encouragement.'B within the "within-30-minute" catego- at risk of another substance-depen- Our research supports other findings ry,faster smokers are more dependent. dence problem.'This association should that cigarette dependence can be diag- We advise that all smokers be asked the become stronger as cigarette use be- nosed with a few simple, direct ques- number of minutes it takes them to have comes rarer. tions. Killen et al"performed a simple their first cigarette of the day. Those diagnosis, using a five-question modifi- who smoke within ten minutes appear to We thank Roberta G.Ferrence,PhD,for advice cation of the eight-question Fagerstrom be more dependent(see survey 1)than and John Firth, MA, for providing demographic lblerance Questionnaire.'The most im- those who smoke within 11 to 30 data from hospital records. Refensness 1. The Health Consequences of Smoking:Nicotine Disorders,ed 3,revised.Washington,DC,Ameri- 15. Kozlowski LT,Mann RE,Wilkinson DA,et al: Addiction,The 1987 Report of the Surgeon Gener- can Psychiatric Association,1987. 'Cravings'are ambiguous:Ask about urges or de- al, publication DHHS (CDC) 88.8406. Dept of 8. Kozlowski LT,Director J,Harford MA:Tobacco gires.Addict Behav,in press. Health and Human Services,1988. dependence, restraint and time to the first ciga- 16. Robins LN:Estimating addiction rates and lo- t. Brecher LM: Licit and Illicit Drugs. Mount rette of the day.Addict Behar,1981;6213-219. cating target population:How decomposition into Vernon,NY,Consumers Union,1972. 9. Fagerstrom Ko:Measuring degree of physical stages helps,in Rittenhouse JD(ed):The Epidemi- 3. Kozlowski LT: Factors affecting widespread dependence to tobacco smoking with reference to ology of Heroin and Other Narcotics, research smoking cessation,in Ramstrom L,Raw M,Wood individualization of treatment. Addict Behan monograph 16.National Institute on Drug Abuse, M(eds):Guidelines on Smoking Cessation for the 1978;3:235-241. 1977,pp 25-39. Primary Health Care Team. Geneva, World 10. SPSSX Usere Guide,ed 2.Chicago,SPSS Inc, 17. Pechachek TF:Modification of smoking behav- Health Organization and International Union 1986. ior, in Krasnegor NA(ed): The Behatioural As- Against Cancer,1988,pp 14-18. 11. Haertzen CA,Kocher TR,Miyasato K:Rein- pecta of Smoking,research monograph 26.National 4. Johnson BD:The race,class,and irreversibility forcements from the first drug experience can pre- Institute on Drug Abuse,1979,pp 127-188. hypotheses:Myths and research about heroin,in dict late drug habits and/or addiction:Results with 18. Schachter S:Recidivism and self-cure of smok- Rittenhouse JD(ed):The Epidemiology of Heroin coffee,cigarettes,alcohol,barbiturates,minor and ing and obesity.Am Psychol 1982;37:436-444. and Other Narcotics,research monograph 16.Na- major tranquilizers,stimulants,marijuana and hal- 19. Killen JD,Fortmann SP,Telch MJ,et al:Are tional Institute on Drug Abuse,1977,pp 51-60. lucinogens,heroin,opioids and cocaine.Drug Alco- heavy smokers different from light smokers? 5. Kozlowski LT,Jelinek LC,Pope MA:Cigarette hol Depend 1983;11:147-165. JAMA 1988;260:1681-1585. smoking among alcohol abusers:A continuing and 12. Russell MAH:The smoking habit and its classi- 20. Lichtenstein E,Mermelstein R:Some method- neglected problem. Can J Public Health 1986;77: fication.Practitioner 1974;212:791-800. ological cautions in the use of the tolerance ques- 205.207. 13. Kozlowski LT,Willdnson DA:Use and misuse tionnaire.Addict Behav 1986;11:439-442. 6. Blumberg HH,Cohen SD,Dronfield BE,et al: of the concept of craving by alcohol,tobacco,and 21. Coambs RB,Kozlowski LT,Ferrence RG:The British opiate users:I.People approaching London drug researchers.Br J Addict 1987;82:31-36. future of tobacco use and smoking research,in Ney drug treatment centres.Int JAddict 1974;9:1-23. 14. Stockwell T:Is there a better word than'crav- T, Gale A(eds):Smoking and Human Behavior. 7. Diagnostic and Statistical Manual of Mental ing?Br JAddict 1987;82:44-45. Chichester,England,John Wiley&Sons,in press. JAMA,February 10,1989-Vol 261,No.6 Cigarette Dependence-Kozlowski et al 901 t Low and Medicine Lawrence O. Gostin, JD, Section Editor Criteria for Evaluating a Ban on the Advertisement of Cigarettes Balancing Public Health Benefits With Constitutional Burdens Lawrence O. Gostin, JD, Allan M. Brandt, PhD THE American Medical Association has long recommended sible in 1972 for the Court to uphold a congressional ban on a ban on promotional advertising of cigarettes,as has a broad tobacco advertising on radio and television.' range of public health organizations.' The tobacco industry, The Supreme Court held for the first time in 1975 that in an uneasy alliance with civil libertarians and the media, commercial advertising was entitled to First Amendment argues that a ban would violate the constitutional right to free protection:"The relationship of speech to the marketplace of expression. The debate has been bitter and vehement, im- products or services does not make it valueless in the mar- peding adequate examination of the complex public health ketplace of ideas."The Supreme Court has defended com- ped constitutional questions engendered by such a policy. mercial speech in cases involving advertisement of abortion In this article, we offer criteria for evaluating a national referral services'and the price of pharmaceuticals'The Su- ban on tobacco advertising. First, we explore modern case preme Court,however,has afforded commercial speech con- law to determine whether the Supreme Court,applying the siderably less protection than social or political discourse. commercial free speech doctrine,would uphold a ban.Second, The Court in two recent cases granted states significant we critically evaluate existing empirical claims about the leeway in regulating commercial speech. nature and impact of tobacco advertising on smoking behav- In Posadas de Puerto Rico v Tourism Company of Puerto ior.Third,we assess the unique aspects of the cigarette as a Rico,'the Court upheld a prohibition on advertising of legal consumer product to see whether it can be distinguished from gambling casinos to the residents of the commonwealth. other hazardous products. Fourth, we ask whether a ban Justice William Rehnquist said that because the legislature would constitute the least intrusive policy necessary to achieve could have prohibited gambling altogether it could take the the public health objective.Finally,we balance the potential lesser step of banning advertising: "It would surely be a public health benefits with the social costs and constitutional strange constitutional doctrine which would concede to the burdens that a ban would impose. legislature the authority to totally ban a product or activity, We conclude that a ban on cigarette advertising would be but deny to the legislature the authority to forbid the stim- an appropriate public health initiative.We build on previous ulation of demand through advertising." scholarship'by evaluating the competing First Amendment This language is highly apposite to advertising cigarettes. theories of a"free market of ideas"vs a"consumer protec- Since government unquestionably has the right to ban out- tion"vision and the distinctive historical features of tobacco right the sale of hazardous products,the Supreme Court is use and regulation. granting the government considerable discretion to reduce THE COMMERCIAL FREE SPEECH DOCTRINE demand for cigarettes through restrictions on advertising. The Court said so explicitly in Posadas: The Supreme Court has defined commercial speech as"ex- Legislative regulation of products or activities deemed harmful,such pression related solely to the economic interests of the speak- as cigarettes,alcoholic beverages,and prostitution,have varied from er and its audience."3 Advertisements by manufacturers of outright prohibition on the one hand...to legalization of the prod- cigarettes inviting consumers to buy their product are ex- uct or activity with restrictions on stimulation of its demand on the amples of commercial speech.The Court initially found that other hand. To rule out the latter, intermediate kind of response commercial speech deserved no First Amendment protec- would require more than we find in the First Amendment.' tion.'The failure to protect commercial speech made it pos- Notably,the Supreme Court cited cases upholding restric- tions on the advertisement of alcoholic beverages'and ciga- rettes'as precedents for its decision in Posadas. From the American Society of Law,Medicine,and Ethics(Mr Gostin),and the De- The Supreme Court in Board of Trustees of the State Uni- panment of Social Medicine,Harvard Medical School(Dr Brandt).Boston,Mass. versity of New York v Fox(Fox III)'concurred with a per- Reprint requests to the Office of the Executive Director,American Society of Law, missive approach to government regulation of commercial - Medicine,and Ethics,765 Commonwealth Ave, 16th Floor,Boston,MA 02215(Mr Gostin). speech.The Court refused to strike down a State University 904 JAMA,February 17, 1993—Vol 269,No.7 Ban on Cigarette Advertisements—Gostin&Brandt t .• 1 horticultural approaches should be con- derstand the effect of vitamin A as an This meta-analysis was supported in part by sidered. Vitamin A intervention strat- adjuvant to conventional therapy in oth- grant ROI HS-05936 from the Agency for health egies should be integrated into commu- er serious childhood illnesses such as Care P and Research'Rockville,MD,and by a r0 grant DAN-0045-G-SS-6067 from the Office of Nu- nit y programs dealing with other health diarrheal and lower respiratory infer- trition,US Agency for International Development, problems rather than implemented as a tions. Vitamin A supplements are also Washington,DC. vertical program. associated with a significant reduction we thank Walter Willett,MD,Mary C.Smith, MS,Chung-Cheng Hsieh,ScD,and Catherine Ber- CONCLUSIONS in mortality when given periodically to key,PhD,for their valuable insights.We also thank children at the community level. Fac- 'Bruce Kupelnick and Clarence E.Zachery for their Large doses of vitamin A are clearly tors that affect the bioavailability of assistance in the preparation of the manuscript. lifesaving when given to children with these large doses need to be studied measles.More research is needed to un- further. References 1. Underwood BA.Vitamin Aprophykuds programs mortality after a high dose of vitamin A in a high ships.In:Bauernfeind JC,ed.Vitamin A Deficien- in developing countries:past experiences and fu- risk population.BMJ.1992;304207-210. cy and Its Control.Orlando,Fla:Academic Press ture prospects.Nutr Rev. 1990;48265-274. 13. Muhilal, Penneisih D, Idjradinata YR, Mu- Inc;1986. 2. Ellison JB.Intensive vitamin therapy in mea- herdiyantiningsih,Karyadi D.Vitamin A-fortified 25. Sivakumar B,Reddy V.Absorption of labelled sles.BMJ.19322:708-711. monosodium glutamate and health,growth,and sur- vitamin A in children during infection.Br J Nutr. 3. Sommer A.Nutritional Blindness:Xerophthal- vival of children:a controlled field trial.Am J Clin 197227297304. mia and Keratomalacia.Oxford,England:Oxford Nutr.1988;48:1271-1276. 26. Mahalanabis D,Simpson TN,Chakraborty ML, University Press;1982. 14. Kothari G.The effect of vitamin A prophylaxis Ganguli C,Bhattacharjee AK,Mukherjee KL.Mal- o. Hussey GD,Klein M.A randomized,controlled on morbidity and mortality among children in ur- absorption of water miscible vitamin A in children trial of vitamin A in children with severe measles. ban slums in Bombay.J Trop Pediatr.1991;37:141, with giardiasis and ascariasis. Am J Clin Nutr. N Engl J Med.1990;323:160-164. 15. Chalmers TC, Smith H, Blackburn B, et al. 1979;32:313-318. 5, Barclay AJ,Foster A,Sommer A.Vitamin A A method for assessing the quality of a random- 27. Wolbach S,Howe P.Epithelial repair in recov- supplements and mortality related to measles: a ized control trial. Controlled Clin Trials. 19812: ery from vitamin A deficiency.J Exp Med, 1933; randomized clinical trial.BMJ. 1987294294-296. 31-49. 57:511-526. 6. Coutsoudis A,Broughton M,Coovadia HM.V]- 16. Detsky AS,Naylor CD,O'Rourke K,McGeer 28. Nauss KM.Influence of vitamin A status on the Lamin A supplementation reduces measles morbid- AJ, L'Abbe KA. Incorporating variations in the immune system.In:BauernfeindJC,ed.Vitamin A ity in young African children:a randomized,placebo- quality of individual randomized trials into meta- Defufency and Its Control.Orlando,Fla:Academic controlled,double-blind trial.Am J Clin Nutr.1991; analysis.J Clin Epidemiol.1992;45255-265. Press Inc;1986. 54:890.895. 17. Rothman KJ.Modern Epidemiology,Boston, 29. Abdeuaber MH,Monto AS,Tilden RL,Schork 7. West KP,Pokhrel RP,Katz J,et al.Efficacy of Mass:Little Brown&Co Inc;1986. MA,Tarwotjo I.The impact of vitamin A supple- vitamin A in reducing preschool child mortality in 18. DerSimonian R,Laird N.Meta-analysis in clin- mentation on morbidity.a randomized community Nepal.Lancet.1991;338:67-71. ical trials.Controlled Clin Trials.1986;7:177-188, intervention trial.Am J Public Health. 1991;81: & Herrera MG,Neste]P,El Amin A,Fawzi WW, 19. Coutsoudis A,Coovadia H,Broughton M,Sal- 1654-1656. Muhammad KA,Weld L.Vitamin A supplemen- isbury R,Elson I.Micronutrient utilisation during 30. Rahmathullah L,Underwood BA,Thulasiraj tation and child survival.lancet.1992;34026-1-271. measles treated with vitamin A or placebo.Int J RD,Milton RC.Diarrhea,respiratory infections, 9. Rahmathullah L,Underwood BA,Thuiasiraj RD, Vitam Nutr Res.1991;61:199-201. and growth are not affected by a weekly low-dose et al.Reduced mortality among children in south- 20. United Nations Children's Fund,The State of vitamin A supplement:a masked,controlled field ern India receiving a small weekly dose of vitamin the World's Children,1981.Oxford,England:Ox- trial in children in southern India.Am J Clin Nutr. A.N Engl J Med.1990;323:929-935. ford University Press;1991. 1991;54:568-577. 10. Sommer A,Tarwotjo I,Djunaedi E,et al.Im- 21. World Health Organization. Joint WHO/ 31. Arthur P,Kirkwood B,Ross D,et al.Impact of pact of vitamin A supplementation on childhood UNICEF statement on vitamin A formeasles.Wkly vitamin A supplementation on childhood morbidity mortality:a randomized controlled community tri- Epidemiol Rec.1987;62:133-134. in northern Ghana.Lancet.1992;339:361-362. al.Lancet.1986;1:1169-1173. 22. Frieden TR,Sowell AL,Henning KJ,Huff DL, 32. Arroyave G,Bauernfeind J,Olson J, Under- 11. Vijayaraghavan K,Radhaiah G,Prakasam B, Gunn RA.Vitamin A levels and severity of mea- wood B.Selection ofinterventionstrategies:IVACG Rameshwar KV,Reddy V.Effect of a massive dose sles:New York City.AJDC.I992;146:182-186. Task Force on Intervention.In:Guidelines for the of vitamin A on morbidity and mortality in Indian 23. Arrieta A,Zaleska M,Stutman H,Marks M. Eradication of Vitamin A Deficiency and Xe- children.Lancet. 1990;336:1342-1345. Vitamin A levels in children with measles.in Long rophthalmia:A Report of the Interwtional Vita- 12. Daulaire NM, Starbuck ES, Houston RM, Beach,California.J Pediatr.1992;121:75-78. min A Consultative Group(IVACG).New York, Church MS, Stukel TA, Pandey MR, Childhood 24. Mejia LA. Vitamin A-nutrient interrelation- NY:The Nutrition Foundation Inc;1977. JAMA,February 17,1993-Vol 269,No.7 Vitamin A Supplementatior---Fawzi et al 903 of New York prohibition on private commercial enterprises Advertisements that portray unlawful activities such as from operating on campus.The Court reiterated:"We have smoking by children would not receive constitutional pro- been loath to second-guess the Government's judgment,"and tection.Targeting children with advertisements would sim- "It is up to the legislature to decide ... so long as its judg- ilarly be unlawful.The challenge for antismoking campaign- ment is reasonable." ers is to prove that advertisements that present images at- Many have seen Posadas and Fox 111 as the evisceration tractive to children,such as cartoon characters,are intended of First Amendment protection of commercial speech. Two to promote smoking by children. antithetical visions emerge from the commercial free speech Manufacturers cannot lawfully deceive or defraud the pub- doctrine.The"free market of ideas"vision sees individuals as he with unsubstantiated health claims.Advertisements, for free agents with the ability to assess commercial as well as example,cannot claim that cigarettes have no effect on breath- health messages and to make decisions in their own interests. ing or athletic abilities,or that they pose little risk to health. The"consumer protection"vision,which underlies much so- While the line between misleading claims and healthful visual cial welfare legislation,points to a need for state regulations images is not always clear,the courts rarely find advertise- that protect individuals from the hazards of a free market.10,11 ments misleading provided they do not make any specifically The Supreme Court is clearly leaving the decision whether to false claims. ban the promotion of dangerous goods and services to the Future courts, of course, could find cigarette advertise- political organs of government. ments deceptive even in the absence of the portrayal of un- THE CONSTITUTIONALITY OF REGULATION lawful activities or unsubstantiated health claims. In a legal OF ADVERTISING opinion for the American Medical Association,Vincent Blasi and Henry Paul Monaghan conclude: The deferential posture of the Supreme Court on govern- Given what the cigarette advertising does portray,what it fails to say, ment regulation of commercial speech makes it likely that it and the vast public ignorance of the dangers and addictive quality of would uphold a ban on advertising of cigarettes.12 Most federal smoking,particularly among young persons,it is plain to us that this and state court cases involving the regulation of the adver- kind of advertising can be proscribed as deceptive and misleading. tisement of hazardous products have involved alcoholic bev- The Supreme Court in Cipollonez'left the door open to tort erages, since the ban on cigarette advertising on television litigation by smokers based on claims that cigarette manu- and radio was upheld in 1972.Alcohol cases provide a reason- facturers engaged in fraudulent misrepresentation or con- able but imperfect model for assessing the constitutionality of spiracy to misrepresent or conceal material facts. a ban on cigarette advertising. It is conceivable that courts would afford states greater leeway to regulate alcoholic bev- erage advertisements because of the added"presumption of 2.Is the Government Interest Substantial? validity"'conferred by the Twenty-first Amendment(which Given the strong epidemiologic evidence associating smok- repealed Prohibition).A recent review of all alcoholic bever- ing with lung cancer,heart disease,and other causes of mor- age cases for the National Institute on Alcohol Abuse and Al- bidity and mortality,26 no court would deny that the state has coholism found nearly unanimous judicial support for adver- a compelling interest in reducing smoking. The state has a tising restrictions.13 Three federal courts of appeals,'1•1'and direct paren.s patriae interest in protecting individuals from three state supreme courts1619 all upheld state regulation of harming themselves. The Supreme Court has endorsed the alcoholic beverage advertising,rejecting claims that the reg- concept of"rational paternalism"in state regulation for the ulations violated the commercial free speech doctrine.While health of its citizens.'As evidence mounts concerning the each case involved a different set of facts,some upheld total health hazards of environmental exposure to cigarette smoke, bans on advertising."'In only two cases did state courts find the state may also exercise its police powers to protect non- regulations to be constitutionally impermissible.'0z7 These smokers. courts appear to have taken minority positions. Since they were both lower state courts and decided before Posadas and Fox 111,they have weak precedential value. 3. Does the Regulation Directly Benefit Most advertising cases have followed the standard four- the Public Health? part test for assessing commercial speech in Central Hud- The third prong of the Central Hudson test should be a 8011.=' First, for commercial speech to be protected by the significant hurdle in cigarette advertising cases.Central Hud- First Amendment,it must concern a law-fuI activity and not son requires that the regulation clearly achieve the public be misleading.Second,the government interest asserted must health goal and not be an"ineffective"or"remote"method. be substantial. Third, the regulation of commercial speech Before Posadas and Fox III,lower courts carefully searched must directly advance the governmental interest asserted. for empirical evidence showing a link between alcoholic bev- Fourth,the regulation must be no more extensive than nec- erage consumption and advertising.Virtually all courts found essary to serve the government's interest. such a link despite the absence of scientifically rigorous stud- ies. The courts appear to have accepted a reasonable legis- 1.Is Cigarette Advertising Lawful and Not Deceptive lative belief and have not required an objective scientific or Fraudulent? assessment.One federal court of appeals,15 for example,found Most cigarette advertisements would receive constitution- it not "constitutionally unreasonable for the State of Okla- al protection because they concern a lawful activity and are homa to believe that advertising will increase sales ... of not deceptive or fraudulent.To be sure,many advertisements alcoholic beverages."The Supreme Court has fostered this glamorize smoking by associating it with adventuresome, acceptance of legislative opinion over scientific evidence in athletic,sexual,or creative activity.Yet,the Supreme Court's Posadas. The Court refused to review the strength of the concern with"inherently misleading"advertising is directed legislature's findings,but accepted its"belief'that there was toward messages that encourage fraud, overreaching, or a direct and immediate connection between gambling adver- confusion 7;Glamorous advertising,projecting images of smok- tising and an increase in crime,prostitution,and corruption. ers as"successful,fun-loving people,without warning of the As described herein,objective evidence points to an associ- dangers"are not legally considered to be synonymous with ation between cigarette advertising and smoking.Even if this misleading advertisements.15 evidence does not(and perhaps cannot)reach the level of a JAMA.February 17, 1993--Vol 269,No. 7 Ban on Cigarette Advertisement"ostin&Brandt 905 causal connection, the Court would be likely to accept the cigarette advertising. The tobacco industry argues teat the belief of the legislature that advertising increases tobacco impact of advertising on the decision to start or continua consumption. smoking is relatively limited."They claim that advertising is directed to encouraging brand loyalty and brand switching. 4.Is the Regulation of Advertising More Restrictive Antismoking advocates,however,emphasize the appeal and Than Necessary? force of modern advertising.They suggest that the industry The government must not merely show that its regulation would not spend more than$3 billion annually to advertise if directly achieves a public health purpose, but also that the returns were so modest.Cigarettes remain the second most means used are no more extensive than necessary to achieve heavily advertised product in the United States after auto- that purpose.Even before Posadas and Fox 111 courts tend- mobiles 3°Both sides cite studies to support their respective ed to accept legislative judgments that restrictions on alco- positions. holic beverage advertisements were no broader than neces- The effect of a ban on advertising is, in fact, difficult to sary to pursue the public health goal of impeding stimulation assess clearly because of the diverse determinants of smoking of consumer demand',16 The Supreme Court in Fox 111 made behaviors,methodologic problems with existing studies,and clear that the regulation of commercial speech need not be the strong disagreement among researchers about the validity of least restrictive to achieve the desired end,but only a"rea- the studies.Existing studies do not adequately account for a sonable fit"between the means and ends. full range of possible variables that can affect smoking,such Assuming that a ban on cigarette advertising would im- as media health messages,cultural perceptions,taxation,and pede demand, the question remains whether other reason- labeling requirements. able means could be used to achieve the goal as well or better. Research on the relationship between advertising and smok- Manufacturers and broadcasters point to other less extensive ing includes econometric studies,"cross-cultural studies,33.31 time, place, and manner restrictions, such as disclaimers, and the measurement of advertisement recognition 4041 Each stronger warnings, or counteradvertising. The response of category of research has its own technical problems for eval- the courts has been that the"state's concern is not that the uation of an advertising ban.Econometric research has only public is unaware of the dangers of alcohol ... the concern studied the impact of marginal changes in advertising ex- instead is that advertising will unduly-promote alcohol con- penditures on smoking behavior over short periods of time. sumption despite known dangers."' Posadas again merely Not surprisingly,the studies have found,at best,only mar- deferred to government beliefs saying that it is "up to the ginal changes in consumption.This research,moreover,often legislature to decide whether or not ... a `counter-speech' does not take account of the time lag that would be expected policy would be effective in reducing demand." to occur between the media campaign and subsequent chang- es in consumption.'Cross-cultural studies have questionable validity because of the difficulty in drawing conclusions about The fact that a ban on advertising cigarettes would prob- advertising from one country to another due to differences in ably be held constitutional does not mean it would represent the cultural,social,and regulatory environment.Finally,it is sound public policy. Herein we present and analyze the cri- difficult to extrapolate studies of cigarette brand recognition teria necessary for careful evaluation of such a significant to future consumption of cigarettes. Most of these experi- social policy. mental studies do not purport to draw a causal connection Would a Ban Achieve the Desired between advertising and smoking,"although some do43 Given the difficulty of any final resolution to our under- Public Health Purpose? standing of the role of cigarette advertising,any evaluation The impact of smoking on the health of the population and of a ban must rest on other than purely empirical grounds. the health care system is unparalleled in modern America.21 Public policy,however,is almost always fashioned in a con- Smoking remains the leading cause of avoidable death in text of some uncertainty about its ultimate effect.Research America, so pervasive that it is responsible for about one is unlikely to resolve the scientific issue of the relative ef- sixth of deaths from all causes.21 Cigarettes are responsible fectiveness of a ban on advertising. The question emerges for more than 400 000 deaths each year.30 Smoking causes whether it is reasonable to demand certainty in the causal extensive morbidity and mortality from a broad range of relationship between advertising and smoking before imple- diseases, including coronary heart disease, cerebrovascular menting a major public health policy. disease, and many forms of cancer.31 Despite the maze of empirical research about advertising, The cost of diseases relating to smoking is conservatively sufficient evidence exists to suggest that a ban would have estimated to be on the order of$22 billion for health care. significant benefits in further reducing the prevalence of Indirect costs for lost productivity and earnings from excess smoking over time. The impact of an end to a $3 billion morbidity,disability,and premature death are estimated at campaign that has permeated American culture could result more than$43 billion."Economic costs are borne not only by in the following public health benefits: a reduction in the individuals and families, but by the health care system in initiation of young people to tobacco by removing attractive undercompensated care, the government in Medicare and environmental associations with pleasure and vitality;an in- Medicaid expenditures, and the insurance industry. crease in success rates for those wishing to quit by removing The sheer magnitude of the health problem is not in itself the reassurance of commercial messages;and a reduction in a justification for an advertising ban.The ban must be rea- the daily consumption of cigarettes by removing the rein- sonably likely to achieve its goal.The unprecedented health forcement and cues provided to smokers by advertisements. hazard posed by smoking,however,does justify more active The impact of advertising on children and adolescents is of governmental regulation than would be the case with a public concern,particularly because nearly half of all smokers begin health danger that is less certain and pervasive.Additionally, when less than the age of 18 years,S2 and young people may given the fact that some 50 million Americans smoke; even be more susceptible to subtle commercial messages.43•41 Re- relatively small changes in behavior can achieve significant cent studies demonstrating high levels of recognition of RJ health benefits. Reynolds'Joe the Camel must be examined in the context of Sharp disagreements exist between the tobacco industry the sharp increase in sales of the brand; Camel cigarettes' and antismoking advocates about the nature and purpose of market share among persons less than 18 years of age has 906 JAMA,February 17,1993—Vol 269.No.7 Ban on Cigarette Advertisements—Gostin&Brandt M � i increased from 0.5%to 32.8%since 19$8."Other studies show sumption,always recognizing that the product was legal and •the disproportionate percentage of the tobacco advertising that public health interventions had to confront the positive budget devoted to magazines and sporting events seen pre- cultural images historically associated with its use.`*4 The dominantly by adolescents,particularly young women.41 This cigarette,in its popularity,danger,and lack of regulation,is may have a dual negative effect by both increasing exposure an unusual product,requiring unusual public interventions. of these populations to commercial cigarette messages and The unique features of the cigarette suggest that efforts to decreasing exposure to news and features about the health ban its promotion through advertising are unlikely to lead to hazards of smoking.Some data exist to show that reliance on other abridgments of commercial speech. cigarette advertising revenues has chilled the print media from adequately covering smoking and disease.","" A sec- Would a Ban Represent an Incremental Approach ondary benefit of an advertising ban,therefore,would be to to Regulation? eliminate the financial incentive for magazines and newspa- Proponents for a ban on cigarette advertising must do more pers to downplay the health effects of tobacco. than show that it has a reasonable potential for reducing Historically,the rise of modern smoking during the course consumption.The ban must be shown to be no more intrusive of the 20th century rested fundamentally on the extensive of First Amendment values than necessary to achieve the marketing of the product. In the past two decades profound public health objective. Two decades of regulatory history changes have occurred in the meaning of the cigarette from suggest, however,that a spectrum of less restrictive alter- s positive icon of the consumer culture to its reidentification natives have already been employed,and an advertising ban as a dangerous and addictive product. Tobacco advertising, represents the next incremental public health measure. however,continues to serve as a powerful force maintaining The extensive history of partial regulation of cigarette attractive images of smoking and smokers. The touting of advertising has included substantial efforts by the Federal lower tar and nicotine content suggests to the consumer that Trade Commission to curtail deception";the application by cigarettes are safer.We draw the intuitive conclusion,sup- the Federal Communications Commission of the fairness doc- ported by a range of quantitative and qualitative evidence— trine, requiring a reasonable amount of airtime devoted to however inconclusive—that in an environment free of tobac- public health messages 17;a ban on advertising on radio and co advertising, consumption would decline. television; increasingly strict warning label requirements"; and indirect regulation through tobacco liability litigation." Do the Unique Characteristics of Cigarettes Justify The Surgeon General has sought,without success,voluntary Special Treatment? self-regulation from the industry to withdraw current cam- A further rationale for our recommendation for a ban on paigns that solicit young smokers.fi0 Governmental interven- cigarette advertising rests on the unique features of the tion has also included regulation of smokers themselves cigarette.There are five aspects of the cigarette which lead through significant restrictions on smoking in public places." to this conclusion: First,cigarettes remain---despite signifi- While these regulatory efforts have had considerable suc- cant declines in consumption—an enormously popular prod- cess in identifying cigarettes as a substantial health risk, uct,with some 50 million Americans still smoking regularly. contributing to a decline in consumption, smoking remains Indeed,the cigarette is one of the most popular and successful the principal preventable cause of death.The spending power products of the 20th century.Therefore,the health impact of and creativity of the tobacco and advertising industries to the cigarette, as noted herein, is significant. promote cigarettes in the face of partial regulation create an Second,the cigarette is one of the most well-studied con- obligation on government to consider seriously the next in- sumer products in American history. Epidemiological,clini- cremental step. cal,and scientific evaluations conducted since the 1940s have A ban on advertising would not represent a precipitous and uniformly documented the range of serious harms induced by overly restrictive policy, mindless of the impact on First smoking.There is no longer any empirical uncertainty about Amendment freedoms. Rather, it would be viewed as an the dangers of smoking.31.51 incremental measure taken only after extensive partial reg- Third,cigarettes have idiosyncratically been virtually free ulation, calls for voluntary cooperation, numerous congres- of all governmental regulation,including food and drug reg- sional hearings, and considered scholarly examination.14 ulation"Given current standards of product safety promul- gated by Congress and the courts,it would be virtually im- possible to bring a similar product to market today. Indeed Outweigh the Public Health Benefits? no other product presents comparable risks to health when Few people in the public health community would lament used as intended.If a product were developed today that had the silencing of the tobacco industry. Yet, a thoughtful ex- similar risks(such as a pharmaceutical),it would be banned, amination of free speech claims should not be abandoned even if it offered substantial utility. when the speaker is disliked and the message is considered Fourth, no other product of comparable danger is so vig- simply wrong. We begin our inquiry by stating the consid- orously promoted through the popular media. In the course erable respect we have for First Amendment ideals and the of the 20th century,the rise of smoking was strongly tied to consequent social costs inherent in a ban on advertising.We advertising media 53.11 Tobacco manufacturers have contin- conclude,however,that cigarette advertising does not merit ued to promote their product vigorously in the 30 years since a high level of constitutional protection because of its com- the risks of smoking were fully documented. mercial character and the negligible informational value to Finally,the regulation of tobacco advertising can only be consumers. understood in a broad historical and cultural context that The appeal to First Amendment values by the industry,the recognizes the popularity of the product and the evolution of media, and civil libertarians is premised on a belief in the the industry that produces it.A popular and prevalently used intellectual integrity and right to self-determination of the product was discovered to have enormous risks for eonsum- individual."First Amendment theorists see advertising re- ers. It was not a new product being tested, but rather a strictions as paternalistic so that "the only way [the state] long-standing product with a large clientele and a powerful could enable its citizens to find their self-interest is to deny industry with significant economic force. Since the 1960s,a them information that is neither false nor misleading."3 Jus- number of interventions have been attempted to reduce con- tice Harry Blackmun called for an alternative approach to a JAMA.February 17, 1993—Vol 269,No.7 Ban on Cigarette Advertisements—Gostin&Brandt 907 t ban on commercial speech:to assume that"information is not age smoking and,in any case,would not be lost in a ban.Ijather, in itself harmful, that people will perceive their own best health education campaigns(financed by government,an added. interests....and that the best means to that end is to open tax on cigarette sales,or by the industry itself)could provide the channels of communication rather than to close them."'' objective health information to the public-41 This"marketplace of ideas"concept assumes the existence A ban on advertising could also affect the economic via- of an equal and free flow of information.The sizable financial bility of certain magazines,advertising agencies,and sports investment of the tobacco industry in advertising realistically or cultural events that have become dependent on tobacco cannot be matched by health education programs, and the advertising revenue.While some of these activities would be campaigns may subtly misinform consumers. First Amend- impaired,it would be possible to cultivate alternative spon- ment theorists respond by pointing out that the receipt of sorship or financial support that would not rely on promoting information is the key,not the communicator's economic pow- products that are harmful. er or motivation."Equally troublesome problems of dispar- We do not underestimate the constitutional and social val- ate economic power and profit motivation are evident through- ue of free expression. Yet,we suggest the impact of a ciga- out political,social,and commercial debate ranging from lob- rette ban on overall First Amendment ideals would be lim- byists and political action committees to industrial polluters. ited, and the potential public health benefits would be suf- Justice Rehnquist's observation in Posadas that the power ficiently substantial to justify a ban at this time in the de- to ban the manufacture of hazardous products includes the velopment of smoking policy. lesser power to ban advertising is also rejected by First Amendment theorists.If Justice Rehnquist were to be taken CONCLUSION literally,the state could conceivably ban advertising of any A ban on cigarette advertising is more likely to achieve its product or service that it could constitutionally prohibit.Civil goals if it is implemented as part of a comprehensive public libertarians suggest that it is the restriction of speech,not the health strategy.An advertising ban,for example,could sharp- prohibition of the activity that is the greater power. ly reduce the cost of doing business for cigarette manufac- We ground our arguments in favor of a cigarette adver- turers with a potential drop in price. Since the cigarette tising ban, not on a fundamental disagreement with First market exhibits some price elasticity,particularly for young- Amendment principle,but on the narrow constitutional pro- er,poorer smokers,"','significant increases in taxes may be tection that should be afforded to tobacco advertising. Ad- necessary in order to maintain or increase the price of cig- vertising cigarettes is commercial speech,not discourse about arettes.Revenue created by increased taxation could be used society, politics, or health. The Supreme Court's jurispru- for health education. smoking prevention, and cessation dence has emphasized that"commercial free speech[enjoys] programs.66,67 a limited measure of protection,commensurate with its sub- We view a ban on advertising as an enhancement, not a ordinate position on the scale of First Amendment values" diminution of autonomy. By eliminating existing powerful and is subject to modes of regulation that might be imper- and pervasive commercial messages encouraging smoking, missible in the realm of noncommercial expression.'-" individuals will be able to make choices about smoking based The right to restrict commercial speech is particularly com- on more relevant and accurate information.Our assumption pelling when the product has the potential for serious harm to is that this will lead to a deep and sustained reduction in the public.By analogy,the Food and Drug Administration re smoking over time, resulting in a significant benefit to the tains considerable authority to dictate the commercial claims of health of Americans. pharmaceutical manufacturers even for approved drugs.Cig- arette advertising deserves a meager level of constitutional Mr Gostin is grateful for the support for research on this article from the National Institute on Alcohol Abuse and Alcoholism of the US Public Health protection even when compared with other commercial speech. Service,Washington,DC. The courts have upheld commercial advertising because it con- nces veys some information clearly beneficial to consumers,such as 1. American 3 1. American MedicalAssociation.M e ting-,.American Medical Association House i.Delegates the price of a pharmaceutical or professional services or the Resolution 96.Interim Meeting;December 3-6,1989;Honolulu,Hawaii. availability of reproductive services.'There is negligible infor- 2.warner KE,Emster VL,Holbrook JH,et al.Promotion of tobacco products:is- mational value to consumers in seductive,healthful,and beau- sues and policy options.J Health Polit Policy Law.1986;11:367.392. & Virginia State Board of Pharmacy v Virginia Citizen's Consumer Council,425 tiful images of smoking. US 748,762(1976). It is important to emphasize that a ban on advertising 4• Valentine v Chrestensen,316 US 52(1942),rev'd,Schaumburg v Citizens for set- would not extend to noncommercial speech b the indust ter Environment,444 US 620(1980). p 3 ry. 5. Capitol Broadcasting Co v Acting Attorney General,405 US 1000(1972).(mem) A tobacco company could,for example,argue in a newspaper ajf^d sub nom,Capitol Broadcasting Co v Mitchell,333 F Supp 582(DDC 1971). or journal that insufficient scientific evidence exists to link 7.Bigelowos ;aVirginia, icco v Tourism 421 US C Company of Puerto Rico,478 US 328(1986). smoking to heart disease.The public health community would & Dunagin v City of Oxford,Miss,718 F2d 738,751(5th Cir 1983)(en bane),cert be unwise to support a policy that allowed the government to dente,467 Us 1259(1984). 9. Board of Trustees of the State University of New York v Fox,493 US 887(1989), suppress communications that contradict established scien- 10. Leading cases.Harvard Law Rev.1%6;100:172.182. tific orthodoxies of the day,for those orthodoxies may sub- 11.Alcoholic beverage advertising on the airwaves:alternatives to a ban or coun- se sequently be dis roved teT'adve�ing•UCLA Law Rev.1987;34:1139-1193.Comment. 9 J p 12. Ile ML,Kroll LA.Tobacco advertising and the First Amendment.JAMA.1990; The fears expressed by some civil libertarians that a ban 264:159-1594. L.Restrictions on the advertising of alcoholic beverages:a review of the on cigarette advertising would erode the First Amendment law.Un published report re ' p po prepared for National Institute on Alcohol Abuse and Al- in other spheres are exaggerated. There is no reason to coholism.19W draft. believe that government would use a ban on cigarette ad- 14.Actmedia Inc v Stroh,830 Fed 957(9th Cir 1986). 15. Oklahoma Telecasters Assn v Crisp,699 F2d 490(10th Cir 1983),rev'd on other vertising as a precedent for restricting other commercial grounds sub nom,Capital Cities Cable v Crisp,467 US 691(1984). speech. Similar "slippery slope" scenarios failed to materi- 1& Queensgate Invest Co v Liquor control cam,433 NE2d 138(SCt Ohio 1982),cert alize in the more than 20 ears following the ban on cigarette sensed,456 Us 807 arse). 3r g � 17.S&S Liquor Market Jae v Pastore,997 AW 729(SCt AI 1985), advertising over the airwaves. 18. Rhode Island Liquor Stores Assn v The Evening Call Pub,497 A2d 331(SCt RI A broad ban,of course,might produce some undesirable ef- 1985). 19. Oklahoma Alcoholic Beverage Control Bd v Heublein Wines Intl,566 Ptd 1158 fects. Cigarette manufacturers sometimes present positive (SCt Okla 1977). messages, such as "young people should not smoke." These 20• Michigan Beer&Wine Wholesalers Asan v Attorney General of Michigan,370 A'W2d 328(Mich App 1985),cert denied,479 US 939(1986). • messages,however,are rarely effective in discouraging under- 21.Brooks v stare,through the Alcohol Beverage Control Commission,442 A2d 93 908 JAMA,February 17,1993—Vol 269,No.7 Ban on Cigarette Advertisements—Gostin&Brandt I (Del Super Ct 1981). 45. Aitken PP,Eadie DR.Reinforcing effects of cigarette advertising on under-age 22. Cerjjral Hudson Gas d Electric Corp r Publ c Service Commission of Neu•York, smoking.Br J Addict.1990;85:399.412, 447 US 557(1980). 46. Aitken PP,Leather DS,O'Hagan FJ.Children's perceptions of advertisements 23. Ohrolik r Ohio State Bar Assoc.436 US 447,462(1978). for cigarettes.Sac Sri Med.1985;21:785.797. 24. Blasi C,Monaghan HP,The First Amendment and cigarette advertising.JAMA. 47. Klitzner M,Gruenew•ald PJ,Bamberger E.Cigarette advertising and adolescent 1956256:502.509. experimentation with smoking.Br J Addict.1991;86:287-298. 25. Cipollone v Liggett Group Inc,112 SCt 260S(1992). 48.Albright CL.Altman DG,Slater MD,Maccoby N.Cigarette advertisements in 26. US Dept of Health and Human Services.Reducing the Health Consequences of magazines:evidence for a differential focus on women's and youth magazines.Health Smoking:25 fears of Progress:A Report of the Surgeon General,Washington,DC: Edur Q.1988;15:225 233. US Dept of Health and Human Services;1989.Publication 10 89-3411. 49. Warner KE.A ban on the promotion of tobacco products.N Engi J Med.1987: 27. Walters v National Association of Radiotion Survivors,105 SCt 3180,3190 316:745-747. (1985)• 50. Warner KE,Goldenhar LM.The cigarette advertising broadcast ban and mag- 28. US Dept of Health and Human Senices,Public Health Service,Centers for Dis- azm coverage of smoking and health.J Public Health Policy.1989;10:32-02. ease Control.The Health Consequence,of Involuntary Smoking:A Report of the 51, Warner KE,Goldenhar LM,McLaughlin CG.Cigarette advertising and maga- Surgeon General.Washington.DC:US Dept of Health and Human Services;1986. zine coverage of the hazards of smoking:a statistical study.N Engl J Med.1992;326: 29. Davis RAI.Current trends in cigarette ad vertising and marketing.NEngt JMed. 905-309. 1987;316:725-732• 52. Neuberger MB.Smoke Screen:Tobacco and the Public Welfare. Englewood 30. Death toll from smoking is worsening.New fork Times.February 1,199LA9, Cliffs,NJ:Prentice-Hall International Inc;1963:49-51. A14. 53. Whelan E.A Smoking Gur.Philadelphia,Pa:George F Stickley;1984. 31. Fielding JE.Smoking:health effects and control.N Engt J Med.1985;313:491- 54.Warner KE.Selling Smoke:Cigarette Advertising and Public Health.Wash- 498. ington,DC:American Public Health Association;1986. 32. US Dept of Health and Human Services.Smoking and Health:A National Sta- 55. Brandt AM.The cigarette,risk,and American culture.Daedalus.1990:119:155- tus Report.2nd ed.Rockville,Md:US Dept of Health and Human Services.Office 176. on Smoking and Health;1990.US Dept of Health and Human Services publication 56.Trade Regulation for the Prevention of Unfair or Deceptive Advertising and CDC 87-8396. Labelling of Cigarettes in Relation to the Health Hazards of Smoking and Accom- 33. Boddew yn JJ.There is no convincing evidence for a relationship between ciga- panying Statement of Basis and Purpose of Rule 99-113(1964). rette advertising and consumption.Br J.addict.1489:84:1255-1261,1263-1265. 57. Ban-haft,Federal Communications Commission,405 F2d 1082(DC Cir 1968). 34. Centers for Disease Control. Cigarette advertising: United States, 1985. subnom,TobaceolnstitutelnevFederal Communications Comm ission,eertdenied, MAfli'R.1990;39:261-265. 3%US 842(1969). .35. Schneider LM, Klein B. Murphy KM. Governmental regulation of cigarette 58. Fritschler AL.Smoking and Politics:Polirymaking and the Federal Bureot+- health information.J Lou-Econ.1981;24:575-612. crncy.New York,NY:Appleton-Century-Crofts;1969. 36, Seldon BJ:Dordoodian K.A simultaneous model of cigarette advertising:effects 59. Gostin LO,Brandt AM,Cleary PD.Tobacco liability and public health policy. on demand and industry response to public policy.Rrr Econ Stat.1989:71:673-677- ✓A31A.1991266:3178-3182. 37, Levit EM,Coate D.Grossman M.The effects of government regulation on teen- 60. Elliot S.Top health official demands abolition of'Joe Camel'ads.New York age smoking.J Law Ecun.198124:545-561. Times,March 10,1992:D1,D21. 38, Hamilton JL.The effect of cigarette advertising bans on cigarette consumption. 61. Rogotti NA,Pashos CL.No-smoking laws in the United States:an analysis of In:Proceedings of the Third World Cor lc varve on Smuking and Health.Washing- state and city actions to limit smoking in public places and workplaces.JA.41A.1991; ton.DC:US Dept of Health,Education,and Welfare;1975. 266:3162-3167, 39. Reuijl JC-On the Deter ination of Advertising Effectiveness:An Empirical 62. Hearings Before the Subcommittee on Health and the Environment,House Study afthe Germmn Cigarette Market.B,-ston,Mass:Kluwer-Nijhoff;1982. Committee on Energy and Commerce on Proposals to Ban or Restrict Tobacco 40. Goldstein AO,Fischer PM.Richards JW,Creten D.Relationship between high Product Advertising(1987)(statement of Martin H.Redish). school student smoking and recognition of cigarette advertisement.J Pediatr.1987; 63. Redish M.Product health claims and the First Amendment:scientific expression 110:485-491. and the twilight zone of commercial speech.Vanderbilt Lou-Rev.1990;43:1433-1461. 41. DiFranza JR,Richards JR',Pauiman PM,et al.RJR Nabisco's cartoon camel 64. Chapman S.Tobacco excise and declining tobacco consumption:the case of Pap- promotes Came]cigarettes to children.JA IIA-1991;266:3149-3153. ua New Guinea.Am J Public Health.1990;80:537-540. 42. Fischer PM,Schwartz MP,Richards J�\,Goldstein A0,Rojas TH.Brand logo 65. Peterson DE,Zeger SL,Remington PL,Anderson HA.The effect of state cig- recognition by children aged 3 to 6 years:Siickey Mouse and Old Joe the Camel. arette excise tax increases on cigarette sales,1955 to 1988.Am J Public Health, JA.11A.1991;266:3145-3148. 1992;82:94-96. 43. Pierce JP.Gilpin E,Burns Dhi,et al.Does tobacco advertising target young 66. Bal DG,Kizer 104%Felten PG,Mozar HN,Niemever D.Reducing tobacco eon- people to start smoking'.evidence from Ca'ifornia.J.-UlA.1991;266:3154-3158, sumption in California: development of a statewide anti-tobacco use campaign. 44. Aitken PP,Eadie DR,Hastings GB.H.:rwood Al.Predisposing effects of cig- JA31A.1990264:1570-1574. arette advertising on children's intentions t smoke when older.Br J Addict.1991: 67. Erickson AC,McKenna JW,Romano RM.Past lessons and new uses of the mass 86,383-3St4- media in reducing tobacco consumption.Public Health Rep.1990;105:239-234. JAMA.February 17, 1993-Vol 269, No. 7 Ban on Cigarette Advertisements-Gostin&Brandt 909 $pepial Communications Legislative Efforts to Protect Children From Tobacco Joseph R. DiFranza, MD; Billy D. Norwood; Donald W. Garner, JD; Joe B.Tye, MBA Public health laws intended to prevent children from smoking have been enacted enact legislation that could effectively in many states. We surveyed the relevant laws in all states and the District of prevent children from using tobacco. Columbia. The efficacy of one such law prohibiting the sale of tobacco to METHODS individuals under the age of 18 years was assessed with the cooperation of an 11- Testing the Efficacy of year-old girl.She was successful in 75 of 100 attempts to purchase cigarettes.On Massachusetts Gen Law the basis of this experience and a review of existing laws, we have made Chap 270,§§6 and 7 recommendations for a model law.These include a prohibition of the possession One hundred business establish- of tobacco by minors,a prohibition of the sale of tobacco to minors,a requirement ments in nine communities in central for a warning sign at the point of sale,a ban on cigarette vending machines,and a Massachusetts were identified where reward for individuals reporting violators of vending laws. cigarettes were sold. These included (JAMA 1987;257:3387-3389) variety stores, pharmacies, gas sta- tions, restaurants, and supermarkets. In some locations, cigarettes were sold over the counter, and at others, they THE USE of tobacco is the most com- quit, more than half die of the sequelae were available from a vending machine. mon form of drug addiction and is re- of smoking.' Since approximately 4000 The study was conducted with the sponsible for one of every four deaths in American children become smokers assistance of an ll-year-old girl who was the United States.'' Nicotine is the daily, nicotine addiction is the most told to enter each establishment unac- psychoactive and addictive phar- common lethal condition of childhood.' companied,look for a posting of the law, macologic component of tobacco. When Preventing children from becoming and either request cigarettes from the measured by the percentage of users addicted to nicotine is such a logical step cashier or attempt to purchase ciga- who lose control of their substance in- in combating the epidemic of tobacco- rettes from a vending machine.If asked take,nicotine is six to eight times more induced disease that it seems surprising whom the cigarettes were for, she was addictive than alcohol.' The addictive that there has never been a coordinated instructed to say they were for an older power of nicotine is further manifest in national effort to prevent children from relative. The child looked her age, and the fact that 90% of current smokers obtaining tobacco. no attempt was made to make her look want to quit and have failed to do so.' Most states have enacted laws in- older. Before this study, she had never Nicotine addiction typically begins tended to prevent children from using attempted to purchase cigarettes. during childhood,when the average age tobacco. As these interventions are in- We recorded the name and address of for the first use of cigarettes is 13 years, tended to reduce the incidence of a each establishment and whether the and for snuff,10 years.'In a survey of 15- particular disease,ie,nicotine addiction tobacco law was posted,cigarettes were year-old children smoking five or more in the pediatric population, it is appro- available over the counter or from a cigarettes per day, 51% had failed to priate for the medical community to machine, our subject was asked for stop smoking when they tried,and 27% monitor the results, compare the proof of her age, and whether she was thought they could not stop no matter efficacy of various methods, and make able to purchase cigarettes. how hard they tried.I Nicotine addiction recommendations as to how more effec- All 100 establishments were studied is resistant to treatment. Long-term tive prevention might be achieved. during a one-week period in July 1986. abstinence rates achieved by smoking In 1985, Massachusetts enacted Gen During the next two weeks,an attempt cessation programs rarely exceed 25%, Law chap 270,§§6 and 7,under the title was made to contact a representative of similar to abstinence rates for the treat- "Crimes Against the Public Health." each establishment by telephone. The ment of alcoholism.6,7 This law increased the minimum age of caller identified herself as a representa- Many individuals become hooked on persons to whom tobacco may be legally tive of the American Lung Association tobacco as children and, despite re- sold to 18 years, required a copy of the doing a survey. The person answering peated attempts to stop, continue to law to be posted conspicuously on the the telephone was asked the following: smoke throughout their lifetime. premises, and deleted a previous re- (1) Do you have a policy about how old Among those unwilling or unable to quirement that the Department of Pub- someone has to be to buy cigarettes in he Health distribute copies of the law to your store?(2)Are you aware of any laws From the Department of Family and Community towns.Violation of the law is punishable about selling cigarettes to children?(3) Medicine,university of Massachusetts Medical School, by fines of$100,$200,and$300 for first, Do you have a sign about this law in your Worcester(Dr DiFranza);Less Tobacco for Children in Maryland,Greenbelt(Mr Norwood);the College of Law, second,and third offenses,respectively. store?Those who claimed to be aware of Southern Illinois University(Mr Garner);and Stop Teen- The purpose of this study was to test the law were asked to identify the legal age Addiction to Tobacco,Palo Alto,Calif(Mr Tye). the efficacy of this law in preventing age for purchasing cigarettes. Only in- The opinions expressed herein are those of the children from purchasing cigarettes,authors and do not necessarily represent the positions p g g rettes,to dividuals who could correctly identify of the American Lung Association. review relevant laws in all 50 states and the legal age were considered to be Reprint requests to Family Practice Residency Pro- the District of Columbia, and, on this informed about the law. gram, Universityof Massachusetts Coordinated Pro gram 2,47 Ashby-,State Rd,Fitchburg,MA 01420(Dr- basis, o suggestmolaw to del lprovisions We did not conduct the interview in DiFranza). that could focus a national effort to person because many of the stores sam- JAMA,June 26,1987—Vol 257,No.24 Children and Tobacco—DiFranza et al 3387 Table 1.—Results of 100 Attempts to Illegally Pur- terviewed where no signs were posted, tion and compliance.A posting of the law chase Cigarettes 38 (93%) sold the child cigarettes is an effective method of infor>>�g vC*•- No. (P<.005). Of the 36 stores with posted dors of the law,serving as a reminder for Total at sue- signs, only one respondent was un- employees and demonstrating that man- No.of cessful familiar with the law. agement is concerned about complying Attempts Attempts w with the law. In this study, compliance overall 100 75 (75) ... National Survey Results was greatest among those establish- Sign posted 36 21 (58) <,005 No sign posted 41 38 (93) All 50 states and the District of Co- ments where the law was posted. Informed of law 49 36 (73) NS lumbia responded to the survey.A sum- Even among those who were informed Not informed of law 28 24 (86) Available from a dark 93 s9 (63) NS mary of the laws restricting the access about the law, compliance was poor. + Available from a machine 7 6 (86) of children to tobacco as of January 1986 This suggests that either there was lit- •computed by x2 is presented in Table 2. Since then, at tle fear of prosecution,or,if prosecution least three changes have occurred.New was a likely possibility, the penalties pled were part of convenience store Hampshire has outlawed the sale of provided by the law were not stringent chains and we were concerned that man- tobacco to children under the age of 18 enough to act as a deterrent. Enforce- agers of other area stores might be years; Virginia has enacted a law that ment of this law has been almost nonex- warned of our activities, thus biasing prohibits sales to children under the age istent, and this lack of enforcement is the results. of 16 years; and South Dakota has out- likely responsible,in large part,for the lawed the sale of smokeless tobacco to failure of this law to achieve its purpose. National Survey of Laws children under the age of 18 years. Comparisons of the efficacy of various Concerning the Access of laws and enforcement procedures would COMMENT Minors to Tobacco be facilitated by the availability of up-to- From November 1985 through Febru- In a state where the law prohibits the date data on smoking prevalence among ary 1986, questionnaires requesting in- sale of tobacco to persons under the age children of various ages on a state-by- formation on laws concerning tobacco Of 18 years,it was shockingly easy for an state basis. However,as the availability and minors were sent to legislative 11-year-old child to purchase cigarettes. of tobacco is only one of many factors libraries, departments of education, More effective laws are clearly needed. contributing to such prevalence rates,a and law enforcement agencies in each Several reasons for the failure of this more direct measure of the efficacy of state and the District of Columbia. particular law can be identified. One these laws would be preferable. An third of the individuals selling ciga- effort to collect such data is under way RESULTS rettes did not know about the law through two nonprofit health organiza- Efficacy of Chap 270 Educating vendors about tobacco tions, Stop Teenage Addiction to To- Of the 100 establishments sampled,75 laws is crucial to obtainingtheir coopera- bacco and Doctors Ought to Care. sold cigarettes to an U-year-old girl (Table 1). She purchased cigarettes Table 2.—Summary of State Laws on Tobacco Access by Minors as of January 1986 from a clerk in 59(63%)of 93 attempts NO of states States and bought from a machine in six(86%) of seven attempts. Whenever she was (y)below which distribution of tobacco fo minors is illegal P 15 1 Hawaii refused cigarettes,she was told she was 16 9 Alaska,Connecticut,District of Columbia,Indiana,Maryland,New too young.She was never asked her age. Jersey,Pennsylvania,Rhode Island,Texas Only four of the 100 stores had a copy of 17 3 Delaware,North Carolina,Vermont the law posted where it was visible to 18 24 Arizona,Arkansas,California,Florida,Idaho,Illinois,Iowa,Kansas, customers. Maine,Massachusetts,Michigan,Minnesota,Missisllppi, Nebraska,Nevada,New York,North Dakota,Ohio,Oklahoma, We contacted 80 representatives of Oregon,south Carolina,Tennessee,Washington,West Virginia the establishments by telephone; 77 19 2 Alabama,Utah agreed to answer the three questions. Legal to sell 12 Colorado,Georgia,Kentucky,Louisiana,Missouri,Montana,New Many businesses had unlisted tele- at any age Hampshire,New Mexico,South Dakota,Virginia,Wisconsin, phone numbers.Forty-nine(64%)knew Wyoming of the law, while 28 (36%) either were Use,purchase,or possession of tobacco by a Arizona, Idaho,Illinois,Kansas,Michigan,Minnesota,Nebraska, unaware of the law or could not cor- North Dakota,Rhode Island,Tennessee,Utah,West Virginia rectly identify the legal age for purchas- blaming required at point of sale ing cigarettes.Of the 49 establishments Copy of law 2 California,Vermont where the respondent was informed Warning sign 6 Illinois,Indiana,Massachusetts,New York,Ohio,Tennessee about the law, 36 (73%) had sold ciga- On vending machine 6 Indiana,Maryland,Minnesota,New York,Tennessee,Utah rettes to the child.Of the 28 who did not Penalties Loss of license 4 Hawaii,Nebraska,Nevada,Tennessee(loss of license to do any know the law, 24 (86%) had sold ciga- business) rettes. Monetary fine 28 Alabama,Connecticut,District of Columbia,Florida,Hawaii,Idaho, In addition to the four establishments Illinois.Indiana,Kansas,Maine,Maryland,Massachusetts, Michigan,Mississippi,Nevada,New Jersey,North Carolina,New where copies of the law were visible, York,Ohio,Oklahoma,Pennsylvania,Rhode Island,South another 32 representatives of the stores Carolina,Tennessee,Texas,Utah,Vermont,West Virginia contacted by telephone claimed to have Imprisonment 15 Alabama,District of Columbia,Florida,Idaho,Kansas,Maine, a CO of the law posted where it was Maryland,Massachusetts,Michigan,Mississippi,Nevada,New copy p Jersey,North Carolina,New York,Ohio,Oklahoma, visible to employees.(The law states the Pennsylvania,Rhode Island,South Carolina,Tennessee,Texas, sign must be posted in a conspicuous Utah,Vermont,West Virginia location but does not specify t0 whom It Not specified 9 Arkansas,Alaska,California,Delaware,Iowa,Minnesota,Nebraska, P Y Oregon,Washington should be conspicuous.) Of these 36 Vending machines must be inaccessible to minors establishments with posted signs, 21 2 Idaho,Utah (58%)sold cigarettes to the child.Of the Portion of fine goes to informer 2 Oklahoma,South Carolina 41 representatives of establishments in- 3388 JAMA,June 26,1987—Vol 257,No.24 Children and Tobacco—DiFranza et al Unfortunately, from a public health vending machines provide easy access fine against a merchant caught selling Lawpolnt, these laws must be enacted for minors.The logic for the elimination tobacco to a minor will go to the in- on a state-by-state basis. However,this of cigarette vending machines was de- former who notified authorities. Ten- does not prevent a coordinated national scribed by the Seventh Circuit Court of nessee provides that it is not entrap- effort by the medical community to Appeals in 1937 in upholding such a ban ment for law enforcement authorities to promote effective legislation. enacted by the City of Chicago: use a minor to purchase tobacco for Proposed Model The evil sought to be reached by forbidding purposes of monitoring compliance with the sale of cigarettes in automatic vending the law. A hybrid of these two ap- Law Characteristics >3the g roaches would allow civic or aniza- machines was the purchase of cigarettes by P g 1. Possession of tobacco by children immature minors. Automatic vending ma- tions to monitor compliance with the should be prohibited. This is now the chines in order to achieve their purpose, law by utilizing our study technique. case in only 12 states. When teaching namely,dispensing with salesmen and mak- Volunteer health organizations, parent about smoking in grade schools, one of ing facile the purchase of goods without the teacher associations, and youth groups us P.R.D.)has been asked repeatedly: intervention of human service,are placed in (eg, the Boy Scouts and Girl Scouts of "If smoking is so bad for you,why aren't localities easily accessible to the public,are America)could earn money while help- cigarettes illegal?" The fact that chil- inanimate and automatic, and respond ing to enforce the law.This money could dren can legally smoke reinforces the equally efficiently to coins placed therein by be used to support other community a boy or a girl as to coins inserted by an message of tobacco advertisements: adult.8 health projects such as smoking cessa- "Smoking can't be as bad as they say." tion programs or school health educa- Children violating this law might be Our experience confirms the wisdom of tion. required to attend a smoking cessation this statement. There is no excuse,given all we know program for children and their parents. 7. All free distributions of tobacco today about the addictive and dan- t. Possession or use of tobacco by should be prohibited. 7bbacco company gerous properties of tobacco, for states students should be specifically prohib- employees distributing free samples to to fail to enforce their laws against the ited on school property.Such bans are in children recalls the image of the drug sale of tobacco to children,and for nine effect in five states. Smoking by school pusher giving out samples in the states to have no law whatsoever that personnel should also be banned be- schoolyard to get the children hooked prohibits the sale of tobacco to minors. cause they serve as role models. and generate future business. Of the 39 The wisdom of enforcing these laws is 3. The sale of all tobacco products to states banning the sale of cigarettes to also compelling. If the decision about individuals under the age of 21 years minors(as ofJanuary 1986),all but three smoking can be delayed until adulthood, should be prohibited. Nine states now (Indiana, New York, and Oregon) also choosing to become a smoker is unusual. allow the sale of tobacco to children of prohibit providing free tobacco to mi- Given that only 10%of current smokers any age. Standard identification should nors. However, children frequently ob- began as an adult, enforcement of mi- be required for proof of age, and laws tain free samples in violation of such nors' tobacco access laws presents a should not provide that a child may buy laws and the tobacco industry's volun- unique opportunity to deal a mortal tobacco with a parental note, as notes tary code.Legislation was introduced in blow to this fatal addiction. Efforts to are easily forged. 1986 in Maine; Michigan, Minnesota, enact comprehensive and enforceable With the threat of withholding high- Ohio, and Pennsylvania to ban all free laws, such as those outlined herein, way funds, the federal government is sampling or to ban sampling where chil- should receive the full and enthusiastic encouraging establishment of a uniform dren might be present. support of the health care community. legal drinking age of 21 years.The mor- 8. All vendors of tobacco products bidity and mortality due to tobacco should be licensed annually. Annual li- This study was supported in part by a grant from dwarfs that due to alcohol. It seems tensing will provide an opportunity to the hues terican Lung Association of Central Massa- s. reasonable to apply a similar standard educate vendors about current laws re- We thank our 11-year-old assistant,Rachael Dus- to tobacco. garding tobacco sales, and allow public sault,who is in the sixth grade at St Patrick School, 4. Signs warning that it is illegal for officials to prevent repeated offenders Jafey,NH. minors to buy tobacco, or for stores to from selling tobacco. sell it to them, should be conspicuously 9. Penalties for the sale of tobacco to m"� visible to both employees and custom- minors must be stringent enough to 1. Ravenholt RT:Tobacco's impact on 20th-century ers wherever tobacco products are sold, deter would-be offenders in light of the US mortality patterns.Am J Prev Med 1985;1:4- including vending machines. Ten states profit that can be made by selling to- 17. 2. Ravenholt RT: Addiction mortality in the now require signs stating that it is il- bacco to children. United States, 1980:'Tobacco, alcohol, and other legal to sell tobacco to minors. This Pressed to decide between condoning substances.Poput Dev Rte„1984;10:697-724. study demonstrates the effectiveness of smoking by children and enacting laws 3. Pollin W:The role of the addictive process as a such signs in increasing compliance considered by some to be unenforcea- key step in causation of all tobacco related diseases. with the law. ble state le islatures have decided in 40 JAMA son L 52:2874. s � 4. Johnson LD,O'Malley PM,Bachman JG:Use of 5. State law should require public states and the District of Columbia to Licit and Illicit Drugs by Americas High School schools to provide education about the outlaw tobacco sales to minors. Effec- Students:1975-1984. National Institute on Drug health effects of tobacco use. Such a tive enforcement of these laws is criti- Abuse, 1985. reuirement is already m effect in it cal however,if the are to achieve their 6. ce of J,Drury CG: in assessment of the inei- q Y � , r Y dente of cigarette smoking in fourth year school states. Public health departments or purpose. Enforcement is occurring in children and factors leading to its establishment. other appropriate governmental agen- some communities,but in the absence of Public Health 1980;94:243-260. cies should be required to make signifi- community concern, law enforcement 6. Ockene JK, Benfari RC, Nuttall RL, et al: cant efforts to educate the public and officials are unlikely to expend si fi- behavior change g psychological factors to smoking P y p gm behavior change in an intervention program.Prey business community about the law and cant resources on the enforcement of Med 1982;11:13-28. its underlying rationale. these laws. 7. Adesso VJ:Some correlates between cigarette 6. Cigarette vending machines Several tactics can be used to improve Z7Sking and alcohol use.Addict Behav 1979;4:269- should be prohibited as are alcohol enforcement.Oklahoma and South Car- 8. Illinois Cigarette Service Co v City of Chicago, • vending machines. Poorly supervised olina provide a bounty:a portion of any 89 Fed 610(7th Cir 1937). JAMA,June 26,1987—Vol 257,No.24 Children and Tobacco—DiFranza et al 3389 xµ ContemporaryThreats to Adolescent Health in the United States Robert Blum, MD, PhD Over the past 30 years,adolescents have been the only population in the United lems (Fig 1). By the end of the decade, States who have not experienced improvement in their health status. Violence the number of youths began to fall;the has replaced communicable diseases as the primary cause of juvenile mortality reduction will continue through the and,currently,over 77%of adolescent deaths are caused by accidents,suicide, 1980s. and homicide. Increasingly, poverty, life-style, and risk-taking behaviors are Second,when compared with whites, the persistently younger age of death influencing the morbidities of youth, with associated sequelae of trauma, and the consistently higher birth rates adolescent pregnancy,substance abuse,physical/sexual abuse,and most other of Hispanic and nonwhite populations in major health problems of adolescents. Clinicians seem to be interested in the United States mean that minorities addressing the shifting nature of adolescent morbidity, yet are insufficiently are clustering in younger age groups. trained to be effective. The physician must go beyond traditional clinical Concurrent with this trend is the rising medicine and address the social, environmental, and behavioral factors that tide of poverty in the juvenile popula- underlie current morbidity. tion. By 1990, it is estimated that one (JAMA 1987;257:3390-3395) third of the population under 20 years of age in the United States will be His- panic or nonwhite' and one fifth of all youths will live at,or below,the poverty THOSE who provide adolescent health early 1970s, the number of adolescent level(S. Rosenbaum,JD, and B. Star- care daily witness the mortality and minors 12 to 17 years of age reached field, MD, MPH, unpublished data, morbidity that are the result of social unprecedented numbers in excess of 25 April 1986). Juvenile poverty is dispro- factors and health-risk behaviors of million, thereby increasing the preva- portionately found among nonwhite youth. It is the intent of this report to lence of many adolescent health prob- populations. outline the major health problems of teenagers from the multiple perspec- tives of mortality and morbidity data Fig 1—Number of adolescent minors 12 to 17 years of age in the United States,by year.(From Kovar MG: and of youths themselves.What I hope Better Health for Our Children:A National Strategy.US Dept of Health and Human Services,1981,vol 3.) will emerge is an awareness that for health care providers to address the 2s physical morbidities without address- ing the social etiologies is, in the long �/' (25.1) ' (24.1) run, going to do little to improve the 24 (242) (24.9) (25.0) 6(24.1) health status of youth. , w (22.7) i THE SHIFTING DEMOGRAPHY s_` i OF YOUTH IN AMERICA 2 22 (21•7) r(22.0) Two major trends are having and will (zt.$) continue to have significant impact on 020.9) ;r how adolescent health problems are 5 20 viewed in the United States. Between 6 1954 and 1964, there were 4 million births each year, reaching a peak in 18 1961.' The consequence is that in the From the Adolescent Health Program,Departments of Pediatrics and Matemal and Child Health,University O ry b 6 0 Z 5 h of Minnesota,Minneapolis. �O� ,O� Cb Nq -z Reprint requests to the Adolescent Health Program, Box 721, University of Minnesota Hospital and Clinic, Year Harvard Street at East River Road, Minneapolis, MN 55455(Dr Blum). 3390 JAMA,June 26,1987—Vol 257,No.24 Adolescent Health—Blum JUnmentary Removing the Incentive to Sell Kids Tobacco A Proposal THE TOBACCO companies have a strong financial incentive smoke and revised their marketing and distribution systems to recruit children as new smokers.Every day 1200 smokers to reduce consumption by.children. die and another 3500 stop smoking.Once a smoker is addicted, There are several ways they could do this.The Marlboro he or she can be expected to generate a daily revenue stream and Camel campaigns3,1,'clearly appeal to children. The to- for the tobacco industry for 20 or more years.Since very few bacco companies could drop these campaigns. Advertising people begin smoking after they reach adulthood,the tobacco could be moved away from schools and publications read by industry must recruit children. Tobacco advertising should teens. Sponsorship of events that appeal to kids could be not be viewed as a current expenditure but as a long-term ended.The companies could include language in their distri- investment. It is worth the high cost of advertising because bution agreements canceling the agreements if distributors it generates future profits.',' As a result, these children- are caught,by company representatives,selling to kids.The smokers are crucial to the industry's long-term health. The tobacco companies already enforce many restrictions on their tobacco companies regularly and emphatically assert that vendors; this simple step would put the tobacco industry's "We don't want kids to smoke," yet they spend billions on considerable economic clout behind reducing access to ciga- advertising campaigns featuring cowboys and cartoon char- rettes by minors, which remains a continuing serious acters.Not surprisingly,Marlboro and Camel are the brands problem.' most commonly smoked by kids;in California,59%of 12-to This assessment would be determined by the states based 17-year-old smokers smoke Marlboros and 23%smoke Cam- on sales to kids. National surveys issued by the Office of els 3 These figures are typical. Smoking and Health or by the individual states would be used Arrayed against this potent economic force are state laws to determine the prevalence of teen smoking and the quan- making sales of tobacco to youth illegal.These laws are weak tity,the brand,and the revenues generated by tobacco sales. and poorly enforced; 60% to 90% of children report buying These surveys would be conducted annually,and the tobacco their own cigarettes."' While improving enforcement does manufacturers would be sent a bill,much as the property tax have some short-term benefits,it is expensive and requires assessor sends homeowners and business operators a bill. continuous vigilance.A better approach would be to remove There are about 5 million teenaged smokers in the United the economic incentive that the tobacco companies have to States,8 and they each smoke about 0.6 pack per day or about see children smoking. 1.1 billion packs of cigarettes a year.At$1.60 per pack,this I suggest an assessment that eliminates the economic in- amounts to about$1.7 billion in annual sales. centive to get children to smoke or use other forms of tobacco. The assessment would be$3.4 billion,allocated according This assessment on tobacco companies would be twice the to teenagers' smoking patterns. The $3.4 billion is a lot of revenues generated by sales to children younger than 18 money but is less than the$3.6 billion the cigarette companies years of age. This amount would be about equal to the dis- spent on advertising and promotion and less than half their counted present value of the long-term profits generated by profits in 1992. new smokers 2 The assessment would not be on smokers but For example,in 1990 in California,3 Philip Morris sold 61% the tobacco companies.Prices would only have to go up if the of all cigarettes smoked by kids(virtually all Marlboros),RJ tobacco companies refused to pay the assessment out of prof- Reynolds,25%(virtually all Camels),and Lorillard,8%(New- its, which have increased rapidly in recent years to exceed port),with the remaining 5%distributed among other brands. $8.3 billion in the first 9 months of 1992.More important,this If these numbers are typical nationwide, Philip Morris'an- assessment would be avoidable if the tobacco companies sim- nual bill would be$2 billion,RJ Reynolds, $850 million, Lo- ply made good on their claim that they do not want kids to rillard,$272 million,and the others,a mere$170 million. These assessments could be levied at the national or state level, where industry performance could be closely moni- From the Department of Medicine,Cardiology Division,and the Institute for Health tored.In California,for example,it would yield$400 million Policy Studies and Cardiovascular Research Institute,University of California,San for the state. Francisco.Dr Glantz has testified as an expert witness on the health effects of en- vironmental tobacco smoke. This assessment has several advantages:The tobacco com- Reprint requests to Department of Medicine,Cardiology Division,University of California,San Francisco,Moffitt Hospital,Room 1186,San Francisco,CA 94143- panes would have a difficult time objecting to the tax in 0124(Dr Glantz). principle because of their public posture that they don't want JAMA,February 10, 1993—Vol 269,No.6 Commentary 793 kids to smoke; if kids don't smoke, they don't pay the as- the increasing price of_cigarettes, when, in fact, P t i sessment.It splits the tobacco companies from the smokers, creases by the tobacco companies themselves (to i±�creal since the smokers will only have to bear the assessment if the profits)have increased more than three times faster than the tobacco companies refuse to absorb it out of profits. The tax on cigarettes. By presenting the tobacco tax as a tax on assessment is based on actual smoking behavior of kids rather smokers,the tobacco industry is able to mobilize its victims than where they buy the tobacco;this procedure avoids prob- to keep taxes down,making more room for managed price and lems of interstate purchases and smuggling, which are po- profit increases by the industry itself.The proposed assess- tential concerns when there are large differences in state ment would bypass this argument by charging the tobacco excise taxes. Such an assessment provides an economic in- companies based on consumption by children at the same centive for the tobacco companies to reduce marketing to time they are taking a public posture of being opposed to children. The annual public release of statistics on brand smoking by children.This refocusing of the issue of tobacco preference by children would put public pressure on the taxation might be enough to make it politically feasible. tobacco companies to change their marketing efforts,much as Most important,this assessment may create a situation in the recent controversy about RJ Reynolds'Joe Camel cam- which the tobacco companies really do not want kids to smoke. paign has.The survey used to assess the tax will also provide Stanton A.Glantz,PhD ongoing surveillance of tobacco use by youth,which can guide This work was supported,in part,by funds provided by the Cigarette and planning and implementation of tobacco control progress. Tobacco Surtax Fund of the State of California through the Tobacco and Re- Assessing the tobacco companies based on sales price avoids lated Disease Research Program of the University of California,San Francisco, the need for complicated accounting rules based on profits.It under award IRT520. is easy and inexpensive to implement, and it will generate 1. DiFranza JR,Tye JB.Who benefits from tobacco sales to children?JAMA.1990; increasingrevenues as the tobacco companies continue to 2. TyeJ-2787. 11 2. Tye JB,Warner KE,Glantz SA.Tobacco advertising and consumption:evidence increase the price of cigarettes.Should the tobacco companies of a causal relationship.J Public Health Policy.1987;8:492-508. the rice on to their victims the rice increases will 3. Pierce JP,Gilpin E,Burns DM,et al.Does tobacco advertising target young peo- pass P r P ple to start smoking.evidence from California.JAMA.1991;266:3154-3158. discourage some of them from smoking.' 4. Slade J.Learning to fight Nicotiana tobacum.N J Med.1988;85:102-106. 5. Pierce JP,Mills SL,Marcus SE.Accessibility of cigarettes to youths aged 12-17 Enacting this proposal will not be easy. Despite the.-fact years—United States,1989.MMWR.1992;41:485-488. that most states are desperate for funds and looking for new 6. Fischer PM,Schwartz MP,Richards JW Jr,Goldstein AO,Rojas TH.Brand logo recognition by children aged 3 to 6 years:Mickey Mouse and Old Joe the Camel. sources of revenue,the tobacco industry has generally been JAMA.1991;266:3145-3148. successful in keeping excise taxes low in the United States. 7. DiFranza JR,Richards JW Jr,Paulman PM,et al.RJR Nabisco's cartoon camel Il g promotes Camel cigarettes to children.JAMA.1991;266:3149-3154. (By world standards tobacco taxes in the United States are 8. US General Accounting Office. Teenage Smoking:Higher Excise Tax Should very low-11] Canada a pack Of cigarettes costs about US Significantly Reduce the Number of Smokers. Washington, DC: US General rJr P g Accounting Office;1989.General Accounting Office document GAO/HRD-89-119. $5.40,compared with only about$1.60 in the United States.9) 9. Sweanor DT.The tax burden on tobacco:an analysis of tobacco taxation policy in the United States.Ottawa,Ontario:Nonsmokers'Rights Association of Canada; The tobacco industry has kept taxes low through campaign November 1991. contributions to politieians10 and the industry's successful 10. Begay M,Glantz S.Political Expenditures by the Tobacco Industry in Califor- nia State Politics from.1976 to 1991.San Francisco:University of California Insti- public relations campaign to blame government and taxes for tute for Health Policy Studies Monograph Series;September 1991. Editorial Editorials represent the opinions of the authors and THE JOURNAL and not those of the American Medical Association. Toward Fewer Procedures and Better Outcomes In 1987 and 1988 the Health Program at RAND,Santa Mon- When the criteria sets were applied to actual patient records, ica,Calif,published a series of articles on the appropriateness it appeared that a significant percentage of the procedures of various medical procedures that had been performed were unnecessary. The medical and general public pounced on these reports.At last there was evidence that substantial See also pp 753, 761, and 766. savings might be achieved in health care;and since the sav- ings would derive from the elimination of unnecessary care, there should actually be a concomitant increase in the quality between 1979 and 1982.1-3 The RAND investigators had de- of care. veloped appropriateness criteria by a specific method in which In this issue of THE JOURNAL there are three more articles' panels of nine experts reviewed relevant medical literature using the same general method with criteria sets for cardiac and then rated a series of clinical scenarios (indications). procedures that have been updated to 1990.These have been applied to a single state,New York,and suggest that very little of the current care in that state is inappropriate. From the Harvard Community Health Plan Inc,Brookline,Mass. It is tempting to try to draw lessons from both series of Reprint requests to Harvard Community Health Plan Inc,10 Brookline Place W, Brookline,MA 02146(Dr Schoenbaum). studies despite the fact that purists will argue that they are 794 JAMA,February 10, 1993—Vol 269,No.6 Commentary/Editorial of coMparable:Winslow et al,3 who found that 14%and 30% The RAND method includes careful gathering of the ex- f bypass surgeries were performed for inappropriate and isting literature for the expert groups. It would be an enor- uncertain reasons,respectively,studied patients in only three mous amount of additional work to do all the possible meta- hospitals in a western state with a high prevalence of pro- analyses to synthesize the existing literature.Thus,one ad- cedures;whereas Leape et al,'who found that approximately vantage to convening groups of experts is that they do the 2% were inappropriate and 7% had uncertain indications, synthesis in their-heads,rapidly, without a lot of hardware studied a sample of all procedures in an eastern state with a and software. But are physician-experts unbiased? The re- relatively low prevalence of procedures and strict limitations cently published work of McClellan and Brook10 demonstrates on the number of sites at which invasive cardiac procedures that for at least one set of criteria the experts were biased: are performed. It is regrettable that the current studies do In the case of carotid endarterectomy, the experts overly not allow us to compare apples with apples,but undoubtedly assessed appropriateness when their decisions were com- this will become a subject for future research. pared with an approach incorporating outcomes data. To put these studies into perspective some other facts must A determination of appropriateness does not mean that a be taken into account: Between 1980 and 1990 the annual procedure should be performed. In the present series of ar- incidence of bypass surgery has more than doubled,and at the ticles the RAND investigators have added a category of same time there has been more than a ninefold increase in rating procedures as"necessary"or"crucial'for a scenario. percutaneous transluminal coronary angioplasty(PTCA)pro- To be considered necessary the expert group must determine cedures.'If the savings that were projected from the earlier that the procedure is appropriate without disagreement and studies have actually occurred, they must have been over- that"it would be improper care not to provide this service for taken by the extension of angiography,PTCA,and bypass to most patients."Lomas et al"have shown that consensus on previously underserved populations. One might expect that appropriateness, using the RAND method, is much more this .phenomenon would be accompanied by a marked im- likely to occur for patient scenarios for which there is relevant provement in outcome for the American population. Yet; a scientific evidence.This notwithstanding,it is difficult to see recent study by Udvarhelyi et all indicates that among pa- how a determination of necessity can be made by any group tients with myocardial infarction,which is one potential set of experts unless there are sufficient comparative outcomes or"chapter"of indications for these procedures,certain pop- data so that one could be sure that virtually all competent ulation groups, particularly women and blacks, were less patients presented with the data would choose the procedure. likely to undergo the procedures but had at least as good a Is it really necessary to do all the procedures that we are chance or a better chance of survival. currently doing? If not, how might we do fewer? One way Perhaps the outcome driving the increase in procedures is would be to adopt strict payment rules. For example, pay- not survival but quality of life. Data from the Coronary Ar- ment would be made only for care that was supported by tery Surgery Study indicate that some parameters of quality evidence or that met a set of criteria in which a large group of life may be better for patients who have bypass procedures of experts agreed that the benefits were likely to outweigh as compared with medical therapy;however,there were no the risks.When experts cannot agree that a procedure should better results throughout a 10-year follow-up period for rec- be beneficial, should it be done? Does"primum non nocere" reational status,employment status,frequency of heart fail- translate narrowly into"do nothing inappropriate"?Or does ure,use of other medications,or hospitalization frequency.' it include the broader concept of doing only that for which the The RAND method determines that a procedure is appro- benefits appear to outweigh the risks? priate,ie,the benefits outweigh the risks,if the median rating Another way to do fewer procedures would be to bring the of the group of nine experts is within a specified range and patient into the decision-making process.This would partic- if no more than two of the nine rated the procedure inap- ularly apply to the procedures that the RAND method clas- propriate. Nevertheless, a procedure could be rated appro- sifies as uncertain.These are the procedures for which there priate for a given patient scenario if one or two of the experts is disagreement or for which the experts could not decide felt,even after the consensus process,that the procedure was whether the benefits would exceed the risks. When these definitely not of benefit for that type of patient(ie,inappro- opinions are coupled with the possibility of a bias toward an priate)or of uncertain benefit.This,and the notion that the assessment of appropriateness by physicians,it seems fairly benefits must merely exceed the risks,makes a determina- obvious that the decision'should not be made by the physician. tion of appropriateness a relatively low hurdle to clear. The work of Wennberg,Mulley,and colleagues on shared Why then did the RAND investigators develop a method decision making suggests that patients,when presented with that obtains consensus from experts rather than basing rat- evidence about the benefits and risks of procedures choose ings of appropriateness solely on outcomes data?There are differently from their physicians.12,11 How else might we ex- several reasons: One is that there are so many potential plain why patients who were exposed to an interactive video- patient scenarios that.it would be virtually impossible to disk program about prostatectomy as well as to the advice of perform all of the outcomes studies that would be needed to their surgeon chose the procedure considerably less often ground all appropriateness assessments in outcomes data. than patients who only had the advice of the surgeon? It is Another is that these technologies are already being used, hard to believe that there would have been as many angiog- and decisions on their application should be made with the raphies or PTCAs in New York State(20%and 38%uncertain best available information.It can even be argued that by the indications,respectively)had the patients undergone an in- time all of the outcomes studies were done on existing tech- formed consent process in which they learned that for per- nologies new ones would have been introduced. This argu- sons with their condition a group of experts could not de- ment,however,assumes that it is good medicine to continue termine from current medical evidence that the benefits of to introduce new technologies without outcomes studies. the procedure exceeded the risks or that the experts could JAMA,February 10, 1993—Vol 269,No.6 Editorial 795 ' not agree whether it was appropriate to do the procedure. Are we now capable of looking at ourselves critic*y a There are ways to perform fewer procedures that do not discovering the nature of the clothes we are wearing?If require case-by-case examination of the appropriateness of public discovers we are naked before we do,embarrassment the procedure.We and others have shown that it is possible will be the least of our worries.The fabric of our world seems to change the antecedent processes of care,such as the con- to be opinion and uncertainty. If we are going to do what is duct of labor,and to observe marked differences in the con- really appropriate,we need to make an orderly transition to comitant use of procedures, such as cesarean section.","In a world based on evidence, carefully designed processes of addition,prepaid group practices with salaried or capitated care,and truly informed consent.Although it is heartening physicians generally have lower hospital admissions and low- that only a small percentage of cardiac procedures in New er use of discretionary procedures than fee-for-service-based York State meet criteria for inappropriateness,our challenge provision systems.Physicians in prepaid group practices also is to perform fewer procedures and achieve better aggregate appear to use necessary processes of care at least as often, outcomes. and with as good or better outcomes than do fee-for-service Stephen C.Schoenbaum,MD,MPH physicians.1fi-19 Restriction of care to that which is known to be appropriate I. Chassin MR, Kosecoff J, Park RE, et al. Does inappropriate use explain or to that which an informed patient consents and redesign geographic variations in the use of health care services?a study of three procedures. of the processes of care or the provision systems are ap- JAMA 1987;258:2533-2537. 2. Chassin MR,Kosecoff J,Solomon DH,Brook RH.How coronary angiography is proaches that do not require rationing of health care. Al- used:clinical determinants of appropriateness.JAMA.1987;258:2543-2547. though each could lead to fewer procedures being done,none 3. Winslow CM,Kosecoff JB,Chassin M,Kanouse DE,Brook RH.The appropriate- ness of performing coronary artery bypass surgery.JAMA.1988;260:505-509. requires withholding care that should be given. 4.Leape LL,Hilborne LH,Park RE,et al.The appropriateness of use of coronary It should be disturbing to us as a profession that we have artery bypass graft surgery in New York State.JAMA.1993;269:753-760. g Il 5. Hilborne LH,Leape LL,Bernstein SJ,et al.The appropriateness of use of per- so few outcomes data and use so few in our practices. Most cutaneous transluminal coronary angioplasty in New York State.JAMA.1993;269: of us do not learn enough in our training to collect or analyze 6. Be 5. g g' y 6. Bernstein SJ,Hilborne LH,Leape LL,et al.The appropriateness of use of cor- our own data or to interpret consistently the work of others 20 onary angiography in New York State.JAMA.1993;269:766-769. 7. Graboys TB, Biegelsen B, Lampert S, Blatt CM, Lown B. Results of a Early in our training,however,we learn to live and deal-with second-opinion trial among patients recommended for coronary angiography.JAMA. z1 199268:2537-2540. uncertainty ;and although we talk about primum non nocere, 8, Udvarhelyi IS,Gatsonis C,Epstein AM,Pashos CL,Newhouse JP,McNeil BJ. we clearly have a bias toward action.It is hard to escape the Acute myocardial infarction in the Medicare population:process of care and clinical rejoice with our patients and their families outcomes.JAMA.gin 2682530-2536. fact that we can re jP 9. Rogers WJ,Coggin CJ,Gersh BJ,et al.Ten-year follow-up of quality of life in pa- when they have good outcomes,empathize with them when tients randomized to receive medical therapy or coronary artery bypass graft sur- gery:poor personal gratification the Coronary Artery Surgery Study(CASs).Circulation.1990;82:1647-1658. 3' IP ,g P 10. McClellan M,Brook RH.Appropriateness of care:a comparison of global and interaction and our work in both instances, and not trouble outcome methods to set standards.Med Care.1992;30:565-586. 11. Lomas J,Anderson G,Enkin M,Vayda E,Roberts R,MacKinnon B.The role them or ourselves with hard data.In short,we learn to enjoy of evidence in the consensus process:results from a Canadian consensus exercise. playing the game of medical care without hard evidence,and JAMA.1988;259:3001-3005. 12.Winslow R.Videos,questionnaires aim to expand role of patients in treatment the outcome---often unknown in statistical terms-is Of see- decisions.Wall Street Journal.February 25,1992:B1,B3. onda importance to the rOCBSS. 13. Weinstein MM.Dr Video:how best to decide what patients need.New York P P Times.December 13,1992:E16. Major industries have grown up to support Our practices. 14. Schoenbaum SC.An attempt to manage variation in obstetrical practice.In: supporting angiography,PTCA and b ass Heitkoff HA,Lohr KN,eds.Effectiveness and Outcomes in Health Care.Washing- The industries su PP g r ton,DC:National Academy Press;1990:190-200. surgery range from medical equipment companies to the 15. Myers SA,Gleicher N.A successful program to lower cesarean section rates.N Engl J Med.1988;319:1511-1516. educational apparatus that trains all the levels of staff en- 16.Wells KB,Manning WG Jr,Valdez RB.The effects of a prepaid group practice gaged in the decision making and execution of these proce- on mental health outcomes.Health Sero Res.1990;25:615-625. 17. Udvarhelyi IS,Jennison K,Phillips RS,.Epstein AM.Comparison of the quality dures to the hospitals and surgicenters whose bottom lines of ambulatory care for fee-for-service and prepaid patients.Ann Intern Med.1991; depend on this production.The are formidable forces with 115:394 400. I1 P �' 18.Greenwald HP,Henke CJ.HMO membership,treatment,and mortality risk which to contend. It has been difficult for us to desist from among prostatic cancer patients.Am J Public Health.1992;82:1099-1104. 19. Carlisle DM,Siu AL,Keeler EB,et al.HMO vs fee-for-service care of older per- using practices that have been demonstrated to be ineffica- sons with acute myocardial infarction.Am J Public Health.1992;82:1626-1630. cious (eg, intermittent positive pressure breathing') or for 20. Evidence-Based Medicine Working Group.Evidence-based medicine:a new ap- proach to teaching the practice of medicine.JAMA.1992;268:2420-2425. which the benefits no longer outweigh the risk(eg,the slow 21. Light D Jr.Uncertainty and control in professional training.J Health Soc Be decline in the practice of smallpox vaccination in this country hay.1979;20:310-322. 22.Duffy SQ,Farley DE.The protracted demise of medical technology:the case of in the 1970s). intermittent positive pressure breathing.Med Care.1992;30:718-736. 7% JAMA,February 10, 1993-Vol 269,No.6 Editorial ■ Spacial Communication The Economic Implications of Tobacco Product Sales in a Nontobacco State Kenneth E. Warner, PhD, George A. Fulton, PhD Objective.—The tobacco industry claims that tobacco makes a significant eco- nontobacco states. Economic depen- nomic contribution to each of the states in the United States.We sought to deter- dence on tobacco is a less obvious con- mine whether Michigan,a nontobacco state,would reap more economic benefits cern in the nontobacco states than in the from the presence or absence of tobacco product sales. half dozen southeastern tobacco states. Design.—Computer simulation of the Michigan economy, with and without The industry's estimates greatly ex- tobacco product sales,for the years 1992 through 2005. aggerate tobacco's economic importance. intervention.—We simulated Michigan's economy with tobacco expenditures Implicitly,they treat the resources de- g y p voted to tobacco product production and eliminated or reduced and the equivalent spending redistributed to other goods and distribution as disappearing if sales de- services,according to consumers'normal spending patterns. We compare these cline or cease altogether.However,the results with baseline forecasts of the economy. amount of economic activity associated Main Outcome Measures.—Numbers of jobs and tax revenues. with tobacco product sales would not Results.--Michigan would have had 5600 more jobs in 1992 had there been no disappear if consumers decreased or expenditure on tobacco products, with equivalent spending redistributed to other even ceased their spending on tobacco goods and services(and to other taxes,to replace half of lost cigarette excise tax products.Rather,it would be redistrib- revenues),By the year 2005,a tobacco-free Michigan would still have almost 1500 uted as consumers used the same money er- more jobs than it will have if sales trends for tobacco products continue(ie,gradu- to purchase alternative goods and . vices.Just like spending on tobacco,this this ally declining over time). If, instead of tobacco expenditures ceasing,the contem- alternative spending would generate em- porary rate of decline in tobacco consumption had doubled,the state would have ployment and tax revenues associated had over 300 more jobs in 1992 and would have nearly 800 more in 2005.With with the production, distribution, and cigarette excise tax revenues not replaced,the elimination of spending on tobacco sale of purchased goods and services. products would have decreased Michigan's tax revenues by$254 million in 1992. In most or possibly all nontobacco The more realistic doubling of the expected decline in smoking would have states, the reallocated spending might decreased revenues by$14 million that year. be expected to produce more employ- Conclusions,--Reducing or eliminating tobacco product spending in Michigan ment than does tobacco.A sizable pro- will increase employment in the state, as well as health. portion of the dollars consumers spend (JAMA.1994;271:771-776) on tobacco products are "exported"to the tobacco states—half in the case of our state, Michigan. If reallocated, a SINCE the late 1970s, the tobacco in- sures that reduce smoking and hence larger proportion of the spending might dustry has relied increasingly on an eco- cigarette sales,such as clean indoor air be on goods and services indigenous to nomic argument as a public relations laws, will increase unemployment and the nontobacco state.As a consequence, defense against proposed tobacco-con- government deficits.'(In the case of one a higher percentage of the dollars would trol measures. The industry has at- policy measure of contemporary inter- be recycled within the state's economy, tempted to convince legislators, jour- est,cigarette excise taxation,the indus- thereby generating greater local eco- nalists,and the general public that,re- try's argument relates only to employ- nomic activity than did the expenditures gardless of its effects on physical health, ment,2 since higher taxes clearly wiII on tobacco. tobacco plays a critical role in their com- increase tax revenues 3) To evaluate the net effect of this re- munities' economic health, generating To develop its economic argument, allocation in a nontobacco state,we em- employment and contributing vital rev- the industry has commissioned consul- ployed a state-of-the-art macroeconomic enues to government coffers. The in- tants to estimate the economic contri- model to simulate the consequences of dustry contends that regulatory mea- bution of tobacco to the nation as a whole consumers' reducing their tobacco ex- and to the individual states, including penditures in Michigan,with the same the employment generated by produc- level of spending redistributed to other From the Department of Public Health Policy and tion, distribution, and sale of tobacco goods and services according to con- Administration, School of Public Health(Dr Warner), products;compensation associated with sumers'normal spending patterns.We and Department of Economics and Institute of Labor the em to ent;and overnment rev- considered two cases:the complete and and Industrial Relations(Dr Fulton),University of Michi- p � g Ii gan,Ann Arbor.Dr Warner has served as a consultant enues from taxes on the sale of tobacco instantaneous cessation of tobacco ex to numerous voluntary and government organizations products and on associated employment penditures and an acceleration in the concerned with tobacco policy issues.He has received numerous grants for tobacco policy research from and business income.` Findings from current rate of decline in tobacco con- government agencies and private foundations. these studies are transmitted to legis- sumption in Michigan. Reprint requests to Department of Public Health lators,journalists,and other interested Obviously, Michigan's becoming in- Policy and Administration, School of Public Health, University of Michigan,1420 Washington Heights.Ann parties When legislative proposals stantly"tobacco free.>is highly unreal- Arbor-Ml 48109-2029(Dr Warner). threaten cigarette sales,particularly in istic. Nevertheless, this hypothetical JAMA,March 9, 1994—Vol 271,No. 10 Economic Implications of Tobacco Product Sales—Warner&Fulton 771 11 situation directly addresses the issue of Estimation of Gross Vs Net in which tobacco spending has keen re' the net contribution of tobacco to the Economic Impact allocated.The process for estimating the economy of a nontobacco state, in con- net consequences of realistic reductions trast to the industry's evaluation of the The model calculates a baseline fore- in tobacco spending was identical, ex- gross impact.Analysis of this situation cast,which projects the Michigan economy cept that a correspondingly smaller therefore offers a more meaningful view for a number of years into the future as amount of expenditure was removed of the true contribution of tobacco to the it is expected to be,without changes in from tobacco spending and reallocated economy of Michigan. tobacco consumption trends.The model to other goods and services. Because this all-or-nothing proposi- then calculates an alternative forecast, tion is unrealistic,however,we also ex- based on changes in tobacco consumption Assumptions amined the more realistic situation of we introduce into the model and on our Consumption of tobacco products de- gradual reductions in tobacco spending assumptions of how the government will clined gradually in Michigan each year associated with tobacco-control policy react to losing cigarette excise tax rev- through the 1980s and into the beginning measures;specifically,we analyzed the enues.The baseline forecast is subtracted of the present decade (through 1992).'' economic impact of a doubling of the from the alternative, and the resulting To simulate the Michigan economy for rate of decline in tobacco product con- difference is the effect of the simulated the period 1992 through 2005,the base- sumption in Michigan observed over the change. Since the most recent year for line forecast(ie,the expected economic past decade. If Michigan is representa- which we have a complete set of histori- outcomes without any additional tobacco- tive of nontobacco states, our findings cal data is 1991, the initial year for the control measures)must include specific for both scenarios will generalize quali- projections is 1992 (ie, we assume that assumptions about the expected trend in tatively to the other nontobacco states. 1992 is the year in which new tobacco- tobacco expenditures. We therefore as- Neither our study nor the industry's control policy measures first affect the sumed that per capita tobacco consump- analysis considers tobacco-related eco- level of tobacco consumption). tion would continue to decline gradually, nomic activity in the health care sector, The baseline forecast depends on in- according to the recent historical trends, namely, the provision of services to puts for the US economy, calibrated without any additional interventions. Americans suffering from diseases that through 1995 to a forecast provided by Were these normal downward trends not result frorTf smoking. Greater in mag- the Research Seminar in Quantitative included in the analysis,we would have nitude than the direct expenditure on Economics at the University of Michi- overestimated the impact of the cessa- tobacco products,'tobacco-related health gan.Beyond 1995,the baseline forecast tion of tobacco expenditures. care, like tobacco farming, represents is calibrated to the long-term forecast In simulating the economy with specific economic activity jobs, in- generated by the federal Bureau of La- greater reductions in tobacco expendi- comes,and tax revenues—attributable bor Statistics, which runs through the tures,including the scenario of complete to the consumption oftobacco products." year 2005.These inputs to the baseline elimination of such expenditures,we had Its omission from the formal analyses forecast determine the time horizon for to deal with cigarette excise taxes,Since therefore results in an underestimation our projections in this article (ie, 1992 these are not specifically identified as of the gross economic"contribution"of through 2005). By studying effects be- such in the REMI model.Based on our tobacco, as the industry defines it. We yond the initial year of the hypothetical projections of cigarette consumption and omitted this factor to make our analysis change in tobacco consumption,we were excise tax rates,we determined the mag- directly comparable to those produced able to examine the economic irnplica- nitude of state and federal cigarette ex- for the industry. We consider the im- tions of the change over time. cise tax revenues from the sale of ciga- plications of this omission in the"Com- We evaluated two economic outcomes, rettes in Michigan. The state govern- ment"section. employment and tax revenues.To ana- ment would respond to the loss of its METHODS lyze the gross contribution of tobacco, excise tax revenues with an approach analogous to the work of the industry's somewhere between two extremes.One This section describes the basic meth- consultants,we removed from the model extreme would be to compensate com- ods of the study. Details on the com- estimates of personal consumption ex- pletely for lost revenues by increasing puter model, specific assumptions and penditures on tobacco products for each other taxes thereby maintaining the procedures,and sensitivity analyses are of the years from 1992 through 2005 and level of government expenditures(and presented in an appendix available from compared economic outcomes with those hence government employment). The the National Auxiliary Publications Ser- from model forecasts that included ex- other would be not to compensate at all vice (NAPS). pected tobacco expenditures(the base- for the loss of revenues by increasing line forecast).The resultant decrease in other taxes, in which case we assume Macroeconomic Model employment and tax revenues consti- that government expenditures would be For our simulations, we employed a tutes the gross contribution of tobacco reduced by the same amount.Given the macroeconomic model of the Michigan product sales to the Michigan economy. range of possible responses,we adopted economy produced by Regional Eco- To estimate the net economic cone- an intermediate position in performing nomic Models,Inc(REMI),and adapted quences of Michigan's having no expen- our simulations, namely, that the gov- and operated by University of Michigan diture on tobacco products,we removed errinient chooses to replace precisely half economists for a wide variety of state all such spending, as above, and then of lost excise tax revenues.The two ex- economic forecasting and policy analy- reallocated the identical amount of treme responses are evaluated in sen- sis purposes.Outside Michigan,state or money among other goods and services, sitivity analyses,discussed later. substate versions of the REMI model so that the relative consumptions of Analytically,to model the first of the are in use in overhalf the states.Unlike these other products among Michigan extreme positions (complete replace- many other macroeconomic models,the residents remained unchanged,We then ment of lost excise taxes),we compen- REMI model's methodology is fully subtracted the model's baseline forecast sated for the lost excise tax revenues by documented in the professional litera- for the Michigan economy over the same raising personal income tax payments ture and has been subject to peer re- years(ie,including expected tobacco con- in the model by the equivalent amount. view.10,11 sumption)from the simulated economy In the case of the other extreme (no 772 JAMA,March 9, 1994—Vol 271,No. 10 Economic Implications of Tobacco Product Sales—Warner&Fulton replacement of lost excise taxes), we spending on alternative goods and ser- tobacco spending is reallocated accord- diverted the excise tax share of whole- vices. It is possible that a portion of the ing to the general pattern of spending in sale trade(where state excise taxes are reallocated resources would go to saving Michigan. If tobacco users in Michigan collected)to the RE MI model's final de- rather than spending. To examine the are not typical of Michigan consumers, mand vectors for state and local gov- implications of this possibility,we tested it is possible that their alternative spend- ernment.In this event,all of the former the effect of allocating 5% of the redis- ing patterns would differ somewhat from spending on tobacco products was re- tributed moneys to saving.This is a rea- those of the average Michigan consum- directed to nontobacco consumption,but sonable assumption,since the US saving ers.We have no basis,however,for sug- state government expenditures(includ- rate has been below 5%in 6 of the past 7 gesting an alternative spending pattern. ing wage payments) were cut back to years(and was only 5.3%in the seventh Furthermore, any alternative pattern match the loss in excise tax revenue. year). Furthermore,we treated the ad- will produce the same qualitative eco- For intermediate positions,such as our ditional saving as not producing any ad- nomic impact as long as it does not in- base-case assumption of half replace- ditional investment in Michigan and hence volve a dramatically different mix of ment of revenues,shares were handled generating no future employment gains spending on indigenous and out-of-state proportionately. All situations involv- in the state. This highly unrealistic as- goods and services. There is no logical ing an increase in income taxes reduce sumption—analogous to assuming that consideration that would make such a the disposable income otherwise avail- all of the additional saving was stuffed difference plausible. able for purchases in the private sector. into mattresses (or investments were In effect, the excise tax share of the made exclusively out ofstate)--introduces RESULTS former tobacco spending is diverted to a still more conservative bias into this other tax revenue to sustain the level of sensitivity analysis. Gross Impact of Tobacco Sales government spending.In all situations, 3. Reductions in tobacco product sales on Employment and Taxes of course,disposable income for nonto- in Michigan result in revenue declines in in Michigan, 1992 bacco purchases increases significantly. tobacco states. Some small portion of We estimated that, in 1992, tobacco The loss of federal excise tax revenues this loss returns to Michigan in the form product sales in Michigan generated 7843 was treated as increasing the federal of reduced purchases of Michigan prod- "direct employment"jobs(those direct- budget deficit, rather than the federal ucts by the tobacco states, a phenom- ly related to the distribution and sale of government's reducing its expenditures. enon not accounted for in our base-case tobacco products within the state), With this assumption,Michigan's share analysis.We examined the employment concentrated heavily (over 90%)in the of federal revenues is not affected by its impact of assuming that Michigan's share retail and wholesale trade sectors. In diminished contribution of cigarette ex- of national tobacco manufacturing pur- addition to direct employment, we es- cise tax revenues. We deemed this a chases equaled 5%,although the state's timated that tobacco spending was re- reasonable assumption because sales of share of national economic activity in sponsible for 11284 spin-off jobs in the cigarettes in Michigan account for only general is only about 4%.In addition,we state: 2159 "indirect" jobs (generated a minuscule fraction of total federal rev- assumed that all of the direct expendi- by increased purchases from Michigan enues (0.02%) and because we find it ture losses in Michigan were in auto- suppliers)and 9125"induced"jobs (re- implausible that the decrease in ciga- motive products, the state's dominant sulting from spending by people who rette excise tax revenues that results industry.Each of these assumptions in- receive income due to tobacco product from successful tobacco-control policies troduces a conservative bias into the sales);the induced employment includes would be seen by the federal govern- sensitivity analysis.(Note that some of 5333 government jobs. Thus, when we ment as reason to decrease a state's fed- the former tobacco spending that is re- combined direct, indirect, and induced eral grants. distributed to other goods and services employment, tobacco product sales in leaks outside Michigan to other states, Michigan were associated with a total of Sensitivity Analyses but that some of the related economic 19127 jobs in the state in 1992, or just Where our assumptions might affect activity returns to Michigan.Quite pos- under 0.5% of total state employment. the findings of the analysis, we tested sibly, the returning dollars would be According to official state figures,ciga- the impact of employing alternative as- much greater than those lost because of rette sales generated$245 million in ex- sumptions. Specifically, we performed reduced activity in tobacco states.If all cise tax revenues in Michigan in calen- four sensitivity analyses: such fine-tuning measures could be in- dar year 1992, approximately $60 mil- 1. In the absence of any additional troduced correctly into the model, the lion in general sales taxes, and much tobacco-control measures,we assumed net employment effect of the sensitivity smaller amounts in related personal and that per capita tobacco consumption test might be positive.) business income taxes. would continue to decline at a rate con- 4.As discussed in the preceding sec- sistent with recent historical experience tion, we do not know how Michigan's Net Impact on Employment (2.75%per person per year).Were con- state government would react to losing of Eliminating Tobacco Expenditures sumption to decline more rapidly inde- cigarette excise tax revenues.Thus,in in Michigan, 1992 Through 2005 pendent of additional policy measures, addition to our base-case assumption Table 1 presents our estimates of the our model would overestimate net ad- that government replaces half of lost gross and net impacts on employment of ditions to employment because tobacco excise taxes,we evaluated economic out- completely eliminating tobacco product would be a smaller economic activity in comes for the two extreme possibilities: sales in Michigan for selected years from the state prior to implementation of new complete replacement of lost revenues 1992 through 2005 (columns 1 through tobacco-control measures.Thus,we also and no replacement. 3)and the redistribution of jobs by sec- tested the case in which the control fore- There are other assumptions that for in a Michigan economy with no to- cast's assumed rate of decline is 3.5%,a could affect quantitative findings some- bacco product sales(columns 4 through conservative estimate designed to chal- what but would not affect qualitative 7).The first column indicates the gross lenge our findings. conclusions, and the direction of the number of tobacco-related jobs that 2. In our base-case simulations,we re- quantitative impact is unclear. For ex- would have been lost in the Michigan allocated all former tobacco spending to ample, we have assumed that former economy if tobacco expenditures had JAMA, March 9, 1994—Vol 271, No. 10 Economic Implications of Tobacco Product Sales—Warner&Fulton 773 I Table 1.—Gross and Net Impact on Employment in Michigan of Eliminating Personal Consumption Expenditures on Tobacco,by Major Sector,1992 Through 2005* Breakdown of Net Employment by Sector(No.of Jobs)$ (1) (2) Private Sector Gross Gross (3) (4) Tobacco-Related Replacement Net State and Local (5) (6) (7) Year Jobs Lost Jobs Employmentt Government Retail Wholesale Other§ 1992 –19127 24735 5608 –2445 –617 –364 9034 1995 –17492 21314 3822 –2123 –970 –441 7356 2000 –15005 17 330 2325 –1773 –1058 –475 5631 2005 –14091 15 569 1478 –1567 –1020 –457 4522 *Per capita consumption falls 2.75%annually in the absence of any additional tobacco-control interventions.When tobacco expenditures decrease due to additional interventions,half of lost cigarette excise tax revenues are replaced by increasing personal income tax payments.See text and NAPS deposit for additional details. tColumn 3=column 1+column 2. $Sum of columns 4 through 7=column 3. §Services,manufacturing,transportation,communication,public utilities,finance,construction,and mining. ceased completely and instantly in 1992. of$38 million attributed to higher earn- Net Impacts on Employment and Tax The figure for 1992 is thus the gross ings among existing workers and $188 Revenues of Doubling the Rate employment contribution in Michigan of million earned by the 5608 additional of Decline in Tobacco Purchases, in-state tobacco sales, reported in the workers.By 2005,according to our pre- 1992 Through 2005 preceding section.After accounting for dictions, earnings will be $112 million Table 2 presents the results from the a continuation of the expected"normal" greater without tobacco (in 1992 dol- simulation of the more realistic scenario: decline in tobacco expenditures, the lars).This gain includes$61 million for a doubling of the expected rate of de- gross number of tobacco-related jobs higher earnings of existing workers and cline in tobacco product purchases,from that would be lost due to cessation of $51 million for earnings of the additional 2.75%per person per year to 5.5%.The tobacco expenditures would fall gradu- 1478 workers. (In the tobacco-free gloss number of tobacco-related jobs ally to just over 14 000 by the year 2005. economy,existing workers have higher that would be lost as a result of this Column 2 of Table 1 shows our esti- earnings in total for two reasons:there eater than expected rate of decrease mates of the gross employment that is a net change in the mix of industries ranges from 1066 in 1992 to 4842 in the would replace that generated by tobacco in the economy from lower-paying to year 2005(column 1).Gross replacement sales, were such sales ended and the higher-paying industries, and competi- jobs,reflecting alternative spending of equivalent amount of money redirected tive pressures drive up earnings because this money,are given in column 2.The to consumption expenditures on other of increases in demand in the stronger net impact on employment in the state goods and services.Column 3 then pre- economy.) associated with the increased rate of sents the sum of the gross employment decline in smoking is seen in column 3: impacts (with and without tobacco), Net Impact on Tax Revenues of 312 more jobs in 1992,rising to 791 more yielding the net change in employment Eliminating Tobacco Expenditures jobs by 2005. The distribution of job in Michigan that would result from the Assuming that government replaced losses and gains among industries, col- i elimination of spending on tobacco prod- half of lost cigarette excise tax revenues umns 4 through 7,reflects the same gen- ucts: 5608 more jobs in 1992 without by increasing other taxes(our base-case eral pattern (albeit with much smaller tobacco,falling to 1478 more jobs in the assumption), we estimated that if to- numbers)as in the case of the complete year 2005 (again, after accounting for bacco product sales had suddenly ceased, elimination of tobacco product spend- the expected continuation of the "nor- government revenues would have fallen ing. mal" decline in tobacco expenditures). by$126 million in 1992, 1.3%of discre- Replacement of half of lost cigarette Columns 4 through 7 show the distri- tionary state revenues for the year(ie, excise tax revenues would have resulted bution of these additional jobs: small excluding earmarked revenues). Much in a net loss of$7 million in government losses in employment in the sectors most of other tax revenues associated with revenues in Michigan in 1992(less than dependent on tobacco product sales— tobacco product sales, considerably 0.1% of discretionary state revenues), retail trade,wholesale trade,and state smaller in amount than excise tax rev- growing to $29.5 million in 2005. Had and local government (reflecting lost enues,would have been replaced auto- the state made no attempt to compen- cigarette excise tax revenues)—and,as matically by revenues from sales of al- sate for cigarette excise tax losses,gov- shown in column 7, more substantial ternative purchases(ie,the state's sales ernment revenues would have fallen by gains in employment in all other indus- tax)and taxes on related personal and $14 million in 1992(0.14%of state rev- tries(services,manufacturing,transpor- business income.By the year 2005,with enues)and $59 million in 2005. tation, communication, public utilities, no explicit program to replace more than finance, construction, and mining). half of cigarette taxes,the loss of gov- Sensitivity Analyses Not shown in the table is that,in ad- ernment revenue would fall to$86 mil- Results of the sensitivity analyses are dition to the increase in jobs without lion (also in 1992 dollars). The reduced presented in Table 3.In each of the four tobacco sales, employment-related in- losses would be a result of the declining analyses, and for every year studied, come—labor and proprietors'income— trend in tobacco purchases,along with elimination or reduction of tobacco ex- cises as well. ("Labor and proprietors' a small contribution from the expanding penditures increases employment within income," or earnings, is the closest economy associated with nontobacco ex- the state of Michigan, without excep- equivalent in official personal income sta- penditures. tion.This indicates that the qualitative tistics to "employment compensation," Had the state chosen not to replace findings from the base-case analyses are the term used in the tobacco industry's any lost excise tax, government rev- unaffected when more conservative as- analyses.) Earnings are estimated to enues would have fallen by$254 million, sumptions are introduced.The patterns have been $226 million greater in the or 2.6%of state government revenues. for the specific quantitative estimates tobacco-free economy in 1992 than in By 2005,the loss would have fallen to presented in Table 3 are discussed in the baseline forecast.This gain consisted $167 million. the appendix available from NAPS. 774 JAMA,March 9, 1994—Vol 271,No. 10 Economic Implications of Tobacco Product Sales—Warner&Fulton Table 2.—Gross and Net Impact on Employment in Michigan of Reducing Personal Consumption Expenditures on Tobacco,by Major Sector,1992 Through 2005* i Breakdown of Net Employment by Sector(No.of Jobs)$ (1) (2) Private Sector Gross Gross (3) (4) Tobacco-Related Replacement Net State and Local (5) (6) (7) Year Jobs Lost Jobs Employmentt Government Retail Wholesale Other§ 1992 –1066 1378 312 –137 –34 –20 503 1995 –2368 2919 551 –290 –122 –56 1019 2000 –3856 4622 766 –451 –226 –106 1549 2005 –4842 5633 791 –539 –301 –144 1775 *Per capita consumption falls 2.75%annually in the absence of any additional tobacco-control interventions.When tobacco expenditures decrease due to additional interventions,half of lost cigarette excise tax revenues are replaced by increasing personal income tax payments.See text and NAPS deposit for additional details. tColumn 3=column 1+column 2. $Sum of columns 4 through 7=column 3. §Services,manufacturing,transportation,communication,public utilities,finance,construction,and mining. Table 3.—Net Increase in Employment in Michigan With Tobacco Expenditures Eliminated or Reduced,1992 Through 2005:Sensitivity Analyses Sensitivity Analysest Base-Case Year Assumptions* (1) (2) (3) (4a) (4b) Tobacco Expenditures Eliminated 1992 5608 5523 4475 5213 7712 3506 1995 3822 3666 2824 3491 5188 2458 2000 2325 2094 1480 2085 2750 1898 2005 1478 1223 687 1278 1308 1648 Tobacco Expenditures Reduced 1992 312 310 249 290 431 196 1995 551 534 415 506 759 343 2000 766 707 547 704 996 536 2005 791 691 517 725 944 634 *Per capita consumption falls 2.75%annually in the absence of any additional tobacco-control interventions.When tobacco expenditures decrease due to additional interventions,half of lost cigarette excise tax revenues are replaced by increasing personal income tax payments.See text and NAPS deposit for additional details. t(1)Annual decrease in tobacco consumption would have been 3.5%.(2)Five percent of former tobacco spending redistributed to saving(with no in-state investment).(3) Out-of-state purchases of motor vehicles reduced by 5%of tobacco manufacturing expenditure.(4a)Complete replacement of lost cigarette excise tax revenues.(4b)No replacement of lost cigarette excise tax revenues.See text and NAPS deposit for additional details. COMMENT Furthermore, the contractor said that tion per se.It is worth observing,how- We estimated both the gross and net compensatory responses to the absence ever, that each such policy measure economic implications of tobacco for the of tobacco spending"that would occur would itself have economic implications state of Michigan. Our estimate of the automatically within the economy and for the state's economy by virtue of re- gross number of direct jobs,associated within[the firm's]Macroeconomic Mod- distributing resources from other uses primarily with tobacco product whole- el...were constrained from taking place to implementation of the policy. Typi- saling and retailing,is very close to the within [the firm's] analysis."' cally,these impacts would be very small, tobacco industry's most recent estimate: . Our analysis was designed to account reflecting both the relatively small 7843 jobs in 1992 according to our analy- precisely for such compensatory re- amounts of expenditure involved and sis, 7724 in 1990 according to the Price sponses.We found that,in a nontobacco the fact that the relevant redistribu- Waterhouse analysis for the industry.' state such as Michigan,expenditures on tions would not necessarily alter signifi- The proximity of the two estimates sug- tobacco products actually decrease em- cantly the mix of in-state and out-of- gests that both studies are measuring ployment.The important corollary is that state expenditures,as does the shift from the same phenomenon. (Our estimates tobacco-control policies that succeed in tobacco to other expenditures. The is- of indirect and induced employment,and reducing smoking in a nontobacco state sue of the differential economic effects hence total employment, are not com- may have a salutary effect on employ- of various policies is one that we are parable to those of Price Waterhouse, ment within the state,in addition to the pursuing in ongoing research. since the latter include the impact of obvious health benefits. There is one category of economic cost tobacco spending nationwide on employ- Throughout this article,we have men- potentially associated with reductions ment in Michigan,including,for example, tioned tobacco-control policies in a non- in tobacco product consumption that we the effects on Michigan employment of specific manner.As noted earlier,there have not considered in this analysis:the car buying by recipients of tobacco is a wide range of such policies,covering costs of making the transition toward an spending in the other 49 states.) everything from antismoking media economy less dependent on tobacco. In Unlike our analysis, none of the in- campaigns to laws restricting smoking the hypothetical situation in which a non- dustry-sponsored studies evaluated the in public places.With the exception of a tobacco state instantly ceases spending net economic impact of tobacco. How- cigarette excise tax increase, which on tobacco products, transitional costs ever, one of the contractors acknowl- would increase government revenues at would occur because a subset of tobacco edged that money not spent on tobacco the same time that it reduced consump- product retailers and wholesalers and products would be reallocated to other tion,our analysis applies equally to any government employees would lose their spending,and that nationwide(combin- policy that succeeds in reducing tobacco jobs,at least temporarily,while a larger ing tobacco and nontobacco states),the consumption by a given amount, since number of new jobs were created in the economic results with and without to- the analysis focuses on the economic ef- invigorated economy.The disruption for bacco"would be substantially the same." fects of the reduced tobacco consump- the newly unemployed would constitute JAMA, March 9, 1994—Vol 271,No. 10 Economic Implications of Tobacco Product Sales—Warner&Fulton 775 a genuine cost,although even a Tobacco not create as many more jobs as would ployment at the same time that they Institute spokesperson has acknowl- the direct reduction in spending on to- improve the public's health. edged that"Lilf the industry would van- bacco, however, since proportionately I., ish tomorrow,most would find alterna- less health care spending in Michigan is We are grateful to David DiGiuseppe and Peter Live work."13 exported to other states. Nicolas,MPP,for research assistance and to Paul Courant,PhD,Donald Grimes,MA,Jeffrey Harris, This hypothetical scenario is unreal- • The eventual absence of tobacco- MD,PhD,Saul Hymans,PhD,Eugene LeAit,PhD, istic,however.In the more realistic situ- related disease implies that,on average, Jacqueline Murray, and George Treyz, PhD, for ation in which tobacco consumption de- Michigan residents would not contract helpful comments on previous drafts of the manu- ipt. creases rather than ceasing,transitional serious illnesses until later in their se rSee NAPS document No.05096 for 10 pages of costs would be minimal because,as seen (longer) lives, Therefore, some of the supplementary material. Order from NAPS c/o in Table 2,tobacco-related job loss would reduction in health care services real- Microfiche Publications,PO Box 3513,Grand Cen- be very small(less than 0.03%of state ized in a tobacco-free state would trans- tral Station,New York,NY 10163-3513.Remit in employment). late into other kinds of health care ser- advance,in US funds only,$7.75 for photocopies or ' $4 for microfiche.There is a$15 invoicing charge on vices later on (possibly more geriatric all orders filled before payment. Outside the Economic Implications of medicine, for example), mitigating the United States and in some parts of Canada,add Tobacco-Related Health Care diminution of the health care sector im- postage of$4.50 for the first 20 pages and$1 for 10 $1.75 Neither the tobacco industry's analy- plied above) each On balance,however,ag- the firspages of material thereafter,or t microfiche and 50c for each microfiche for iche sis nor ours includes economic activity gregate tobacco-related health care ex- thereafter, in the health care sector in estimating penditures would be expected to fall," References the economic contribution of tobacco. freeing resources to be reallocated to To be consistent with the industry's stud- other uses. 1. Warner KE.Health and economic implications ies, we adopted their implicit assump- ofa tobacco-free society.JAMA.1997;2 :2080-2086, tion that the health effects of smoking The Economic Contribution of 2. Economic Losses to(the(IS Economy or a Spe- cific State)From Increasing the Federal Cigarette are not relevant to the analysis.Clearly, Tobacco in Context Tax Prom 24 Cents to(48 Cents,$1.24,or$2.24)per this is an unreasonable assumption,pro- Ignoring the health care expenditures Pack.Washington,DC:The Tobacco Institute;1993, ducing a substantial underestimation of related to tobacco consumption,we es- Series of processed documents. 3. Saving Lives and Raising Revenue:The Cost, the gross impact of tobacco on employ- timate that tobacco product sales in for Major Increases in State and Federal Tobacco ment.In Michigan,spending on tobacco- Michigan were associated with approxi- 'Taxes.Washington,DC:Coalition on Smoking or related health care was,conservatively, mately 19 000 jobs in the state in 1992, Health;1993. close to $2 billion in 1992,11.1,5 similar in as well as more than$300 million in gov- 4. A Study of the Tobacco Industry's Economic Contribution to the Nation,Its Fifty States,and the magnitude to direct spending on tobacco ernment revenues.Out of context,these District q Columbia. Philadelphia, Pa:Wharton .f rt products of approximately$1.8 billion." are impressive-sounding numbers, es- Applied Research Center and Wharton Economet- Thus,by producing illness,tobacco gen- pecially to bureaucrats and legislators ric Forecasting Associates Inc,University ofPenn- erated economic activity (jobs and in- who must grapple with the difficult twinSylvania;1979. 5. The Economic Impact of the Tobacco Industry comes) in the health care industry in challenges of unemployment and per- an the United States Econo'?ny in 1.983.Bala Cyn- Michigan, just as it did in tobacco in- petually strained government budgets. wyd,Pa:Chase Econometrics;1985;1:chap V,p V-3. dustry retailing and wholesaling.Indeed, In context, however, the numbers are 6. The Economic Impact o the Tobacco Industry�f tobacco-related health care may have less impressive:they mask the fact that, on the United States Economy. Update of 1985 study.Price Waterhouse;1990. accounted for more jobs in Michigan than on balance, expenditures on tobacco 7. The Economic Impact of the Tobacco Indu-stry did tobacco product sales per se, since products actually reduce employment on the United States Economy. Update of 1990 half of the expenditures on the latter in the state. If consumers spent their study.Price Waterhouse;1992. "leaked" out of state. money on goods and services other than 8. Public Heari,?zgsonthe Financing Provisions of Since the tobacco industry's economic tobacco products, the state would gain the Administration's Health Security Act Before the House Committee on Ways and Means,103rd argument asserts that a state will suffer jobs. Cong,1st Sess(November 18,1993)(testimony of economically if tobacco consumption de- The truly important consideration, J.E.Harris,MD,PhD). clines, a complete examination of the however, is the one that places tobac- 9. Schelling TC. Economies and cigarettes. Prev Med. 1986;15:549-560. issue would include consideration of the co's economic role into a proper social 10. TreyzGI.Regional EconondeModeling.-A Sys- economic implications of the eventual context. Annually, tobacco claims the ternatic Approach to Economic Forecasting and elimination of tobacco-related disease. lives of more than 15 000 Michigan citi- Policy Analysis.Boston,Mass:Kluwer Academic While we have not estimated this im- zens.17 For the nation as a whole, the Publishers;1993. pact quantitatively,we can offer the fol- 1990 death toll was 419 000 Americans11. Treyz G., economic-GI, Rickman DS, Shao demographic forecasting anThe REM' d simulation lov6ng qualitative conclusions: each of whom lost an average of 12 years model.Int Regional Sci Rev.1992;14:221-253, * Even if tobacco expenditures could of life expectancy"; millions of others 12. The Tax Burden on Tobacco-Historical Com- cease instantaneously, tobacco-related suffer illness and disability as a direct pilation,1992.Washington,DC:The Tobacco In- stitute;1993. health care expenditures would decline consequence of smoking.Surely,any rea- 13. City Newspaper[Rochester,NY].October 16, only gradually, since current tobacco- sonable accounting of tobacco's"contri- 1986:1,6-8. related disease reflects the cumulative bution"would emphasize the enormity 14. Smoking and Health:A National Status Re- effects of past decades of smoking. of the health toll,With the economic con- port:A Report to Congress,2nd ed.Rockville,Md: Public Health'Service,Dept of Health and Human • Just as in the case of tobacco prod- sequences constituting no more than a Services;1990:48(Table 6).DHHS publication CDC uct expenditures per se,reduced spend- distant secondary consideration. 87-8396(revised February 1990). ing on tobacco-related health care would When that secondary consideration is 15. Council of Economic Advisors.Economic Re- be redistributed to other spending within raised,however,it is essential that it be port of the President. Washington, DC: US Gov- ernment Printing Office;19.92. the economy that would generate "re- correctly interpreted:as this study has 16. Hodgson TA.Cigarette smoking and lifetime placement"employment. Whether this demonstrated,the industry's economic medical expenditures.Milbank Q.1992;7081-125. alternative spending pattern would cre- argument is misleading.In Michigan and, 17. Michigan ASSIST Project Site Analysis.Lan- ate more or fewer jobs depends on a by extension, in other, perhaps all, of sing:Michigan Dept of Public Health;1992. 18. Cigarette smoking-attributable mortality and number of interindustry factors.The al- the more than 40 nontobacco states,to- years of potential life lost-United States, 1990. ternative spending pattern likely Would bacco-control policies can increase em- MAM7R Morb Mortal Wkly Rep.I993;42:645-649. 775 JAMA,March 9, 1994-Vol 271,No. 10 Economic Implications of Tobacco Product Sales-Warner&Fulton t u � Frbm the Centers for Disease Control teads From the Morbidity and Mortality Weekly Report Atlanta, Ga Comparison of the Cigarette Brand Preferences of Adult and Teenaged Smokers— United States, 1989, and 10 US Communities, 1988 and 1990 MMWR. 1992;41:169-173,179-181 those who could not be reached by tele- ana, Iowa, Kansas, Michigan, Minne- TOBACCO USE is the single most pre- phone were mailed a questionnaire.Dur- sota, Missouri, Nebraska, North Da- ventable cause of death in the United ing September-December 1989, the kota, Ohio,South Dakota, and Wiscon- States.' CATI interviews were conducted; be- sin], South[Alabama, Arkansas, Dela- Approximately three fourths of adult cause only persons reached by telephone ware, District of Columbia, Florida, regular smokers tried their first ciga- were asked what brand they usually pur- Georgia, Kentucky, Louisiana, Mary- rette before the age of 18 years (Na- chased, the data for this report were land,Mississippi,North Carolina,Okla- tional Institute on Drug Abuse[NIDA], obtained from 9135 CATI respondents homa, South Carolina, Tennessee, 1988 NIDA Household Survey,unpub- (79%of 11609 adolescents with known Texas,Virginia,and West Virginia],and lished data);about half had become reg- telephone numbers and 76% of 12 097 West[Alaska,Arizona,California,Col- ular smokers before their 18th birth- adolescents in the total sample). orado,Hawaii,Idaho,Montana,Nevada, day.'Knowingwhat brands youngsmok- These data were weighted to provide New Mexico, Oregon, Utah, Washing- ers prefer may suggest what encour- national estimates.Confidence intervals ton, and Wyoming], Marlboro was the ages them to smoke and may suggest (CIs)were calculated by using the stan- most popular brand. smoking-prevention or smoking-cessa- dard errors estimated by the Software Newport was second in the North- tion strategies." for Survey Data Analysis(SUDAAN).8 east(16%),and Camel was second in the To determine brand preferences of Adolescent current smokers (adoles- West(18%).Among white adolescents, smokers, data were reviewed from cents who reported smoking cigarettes Newport was more popular in the North- CDC's National Center for Health Sta- on 1 or more of the 30 days preceding east(14% [95%CI= ± 5.0%])and the tistics'1989 Teenage Attitudes and Prac- the survey)were asked if they usually Midwest(7% [95%CI= ± 3.5%])than tices Survey (TAPS) and the National bought their own cigarettes and, if so, in the South (1% [95% CI= ± 1.2%]) Cancer Institute surveys of adults in which brand they usually bought. and the West(1%[95%CI= ± 1.3%]). 1988 and 9th-grade students in 1990 in Of the 1396 current smokers, 865 COMMIT 10 U.S.communities(four of the 10 com- (62%)reported that they usually bought munities surveyed are located in the their own cigarettes. Smokers aged 16- For the COMMIT study,data on the Northeast[Fitchburg/Leominster,Mas- 18 years were more likely to buy their adult preferences for cigarette brands sachusetts;Paterson, New Jersey;and own cigarettes(71%[95%CI= ±2.9%)) were obtained from telephone surveys Utica and Yonkers, New York]; three than were smokers aged 12-15 years conducted during January-April 1988 of in the West [Vallejo, California; Med- (45%[95%CI= ±4.9%]).Marlboro was random samples of 15415 adult current ford/Ashland,Oregon;and Bellingham, the most commonly purchased brand for smokers(adults who answered"yes"to Washington];and one each in the South both male(69%)and female(68%)ado- the question"Have you smoked at least [Raleigh, North Carolina], Southwest lescents.Camel was preferred more of- 100 cigarettes in your entire life?"and [Santa Fe, New Mexico], and Midwest ten by males(11%)than by females(5%). then answered"yes"to the question"Do [Cedar Rapids, Iowa])participating in Although Marlboro was the most pop- you smoke cigarettes now?")aged 25-64 the Community Intervention Trial for ular brand among white(71%)and His- years in the 10 communities. The sur- Smoking Cessation(COMMIT)evalua- panic (61%) adolescents, black adoles- vey was conducted in two stages: 1)an tion.6 This report examines the findings cents preferred the mentholated brands adult household member reported the of these surveys on the cigarette brand of Newport(61%), Kool(11%), and Sa- smoking status of all adults in that house- preferences of adult and teenaged lem(10%). Among 9th-grade students, hold and 2)all smokers in the household smokers. Marlboro (75% (95% CI= ± 8.2%)), who were aged 25-64 years were inter- Newport(10%(95%CI= ±5.3%)),and viewed. The overall response rate for TAPS Camel(6%(95%CI= ±4.3%))were the the 10 communities was 75%;the first- For the TAPS survey,data on knowl- most commonly purchased brands. stage response rate was 82% (range: edge, attitudes, and practices regard- In all regions(the four regions were 760/&-86%)and the second-stage response ing tobacco use were collected from a Northeast[Connecticut,Maine,Massa- rate was 92% (range: 85%-94%). Cur- national household sample of adolescents chusetts,New Hampshire,New Jersey, rent brand use was measured by re- aged 12-18 years'by a computer-assisted New York,Pennsylvania,Rhode Island, sponse to the question, "What brand of telephone interviewing(CATI)system; and Vermont], Midwest[Illinois, Indi- cigarettes do you usually smoke now?" JAMA,April 8, 1992—Vol 267,No. 14 From the CDC 1893 �h lT Table 1.Percentage of self-reported cigarette brands usually bought by current smokers*aged 12-18 years who usually bought their own cigarettes,by Roarette brandt-Teenage Attitudes and Practices Survey,1989,and cigarette market shares,t 1989 Percentage(95%confidence interval) Benson Category No. Marlboro Newport Camel Winston Salem &Hedges Kool Merit Ventage Other Overall$ 865 68.7 8.2 8.1 3.2 1.5 1.4 1.0 0.5 0.1 7.3 (±3.4) (±1.8) (±2.1) (±1.2) (±0.8) (±1.2) (±0.6) (±0.5) (±0.2) (±1.9) Sex Male 477 68.9 7.3 10.9 3.6 0.5 0.2 1.9 0.7 0.2 6.0 (±4.5) (±2.4) (±3.4) (±1.8) (±0.6) (±0.4) (±1.1) (±0.7) (±0.4) (±2.3) Female 388 68.4 9.4 4.6 2.6 2.9 2.9 0 0.3 0 8.9 (±5.2) (±2.9) (±1.9) (±1.7) (±1.7) (±2.5). (±0.5) (±3.0) Race White 807 71.4 5.6 8.4 3.4 1.0 1.3 0.6 0.5 0.1 7.6 (±3.4) (±1.6) (±2.2) (±1.3) (±0.7) (±1.2) (±0.5) (±0.5) (±0.2) (±2.0) Black 41 8.7 61.3 3.1 0 9.7 3.3 10.9 0 0 2.9 (±9.7) (±15.7) (±6.2) (±7.2) (±6.4) (±9.1) (±5.8) Ethnicity** Hispanic 46 60.9 12.8 7.6 0 2.8 3.7 5.8 0 0 6.5 (±15.0) (±9.5) (±8.6) (±5.4) (±4.9) (±6.1) (±7.6) Non-Hispanic 817 69.1 8.0 8.1 3.3 1.5 1.3 0.8 0.5 0.1 7.3 (±3.5) (±1.9) (±2.1) (±1.3) (±0.8) (±1.2) (±0.6) (±0.5) (±0.2) (±1.9) Age(yrs) 12-15 195 74.8 6.1 8.7 2.5 0.9 0.4 1.1 0 0 5.4 (±6.3) (±3.7) (±3.9) (.t2.1) (±1.3) (±0.8) (±1.5) (±3.2) 16.18 670 67.0 8.8 7.9 3.3 1.7 1.6 1.0 0.6 0.1 7.8 (±3.9) (±2.0) (±2.4) (±1.5) (±0.9) (±1.5) (±0.7) (±0.6) (±0.3) (±2.2) Region Northeast 184 68.4 16.2 4.1 0 2.3 0 0 0.6 0.5 7.9 (±7.7) (±5.2) (±3.1) (±2.3) (±1.2) (±1.0) (±4.0) Midwest 247 70.2 10.0 7.3 3.4 2.2 0 1.1 0.5 0 5.3 (±6.2) (±3:9) (±4.8) (±2.5) (±2.0) (±1.3) (±1.0) (±3.1) South 281 67.2 5.0 6.1 6.2 1.1 2.9 2.1 0.4 0 9.1 (±5.8) (±2.3) (±2.8) (±2.9) (±0.8) (±2.9) (±1.5) (±0.7) (±3.6) West 153 69.6 2.0 18.1 0.7 0.6 2.3 0 0.6 0 6.2 (±8.1) (±2.2) (±6.3) (±1.3) (±1.1) (±2.2) (±1.1) (±4.0) Overall market share,t 1989 26.3 4.7 3.9 9.1 6.2 3.9 5.9 3.8 2.5 33.7 *Persons who reported smoking on 1 or more of the 30 days preceding the survey.Sample size=1396. tSource:Reference 9. JData were weighted to provide national estimates. ¶Excludes the racial category"other"(n=17). **Ethnicity for two persons was unknown. Table 2.Percentage of cigarette brand use self-reported by adult current smokers*,by cigarette brand-10 U.S.communities,19881 Percentage(95%confidence Interval) Benson Virginia All other Community No. Marlboro Winston Salem Kool Newport &Hedges Camel Merit Slims Moral brands Vallejo,Calif. 1,536 24.3 7.7 10.0 8.5 4.6 9.6 4.1 3.5 4.7 0.7 22.4 (±2.1) (±1.3) (±1.5) (±1.4) (±1.0) (±1.5) (±1.0) (±0.9) (±1.1) (±0.4) (±3.9) Cedar Rapids,Iowa 1,234 23.1 9.2 6.4 5.1 0.3 2.6 5.0 9.1 3.3 2.7 33.2 (±2.4) (±1.6) (±1.4) (±1.2) (±0.3) (±0.9) (±1.2) (±1.6) (±1.0) (±0.9) (±5.4) Fitchburg/ Leominster,Mass. 1,185 24.1 13.8 6.8 5.1 8.1 3.0 2.6 5.6 3.5 0.3 27.1 (±2.4) (±2.0) (±1.4) (±1.3) (±1.6) (±1.0) (±0.9) (±1.3) (±1.0) (±0.3) (±4.9) Paterson,N.J. 1,854 24.5 13.8 9.7 6.5 16.0 3.5 1.3 1.7 3.2 0.1 19.5 (±2.0) (±1.6) (±1.3) (±1.1) (±1.7) (±0.8) (±0.5) (±0.6) (±0.8) (±0.1) (±3.3) Santa Fe,N.M. 2,307 28.6 11.0 9.5 3.3 0.5 7.2 11.2 4.2 2.8 1.0 20.8 (±1.8) (±1.3) (±1.2) .(±0.7) (±0.3) (±1.1) (±1.3) (±0.8) (±0.7) (±0.4) (±3.3) Yonkers,N.Y. 1,494 24.0 6.2 9.6 6.4 10.4 4.1 1.4 4.1 3.3 0 30.5 (±2.2) (±1.2) (±1.5) (±1.2) (±1.5) (±1.0) (±0.6) (±1.0) (±0.9) (±4.6) Utica,N.Y. 1,347 21.1 11.7 9.9 4.6 6.8 3.4 3.6 5.2 1.6 2.3 29.7 (±2.2) (±1.7) (±1.6) (±1.1) (±1.3) (±1.0) (±1.0) (±1.2) (±.7) (±0.8) (±4.8) Raleigh,N.C. 1,546 13.1 12.8 13.8 4.4 8.0 4.3 2.5 6.9 5.2 1.4 27.6 (±1.7) (±1.7) (±1.7) (±1.0) (±1.4) (±1.0) (±0.8) (±1.3) (±1.1) (±0.6) (±4.6) Medford/ Ashland,Ore. 1,373 27.5 9.2 4.1 2.5 0.3 4.8 12.5 5.0 3.7 0.9 29.6 (±2.4) (±1.5) (±1.1) (±0.8) (±0.3) (±1.1) (±1.8) (±1.1) (±1.0) (±0.5) (±4.8) Bellingham,Wash. 1,539 23.3 10.5 6.6 3.1 0.2 4.6 14.6 6.6 2.7 0.8 26.9 (±2.1) (±1.5) (±1.2) (±0.9) (±0.2) (±1.0) (±1.8) (±1.2) (±0.8) (±0.4) (±4.5) Overall 15,415 23.6 10.6 8.8 4.9 5.6 4.9 6.1 5.0 3.4 1.0 26.1 (±0.7) (±0.5) (±0.4) (±0.3) (±0.4) (±0.3) (±0.4) (±0.3) (±0.3) (±0.2) (±0.7) *Persons aged 25-64 years who answered'yes'to the question"Have you smoked at least 100 cigarettes in your entire life?"and then answered yes"to the question"Do you smoke cigarettes now?" tUnweighted data. 1894 JAMA,April 8, 1992-Vol 267,No. 14 From the CDC 1 t r Table 3.Percentage of cigarette brand use self-reported by 9th-grade students who smoked and usually bought their own cigarettes*,by cigarette brand-10 U.S. comm♦INNities,19901 Percentage(95%confidence Interval) Benson Virginia All other Community No. Marlboro Winston Salem Koot Newport &Hedges Camel Slims brands Vallejo,Calif. 18 50.0 0 0 5.6 33.3 0 5.6 0 5.6 (±23.1) (±10.7) (±21.8) (±10.7) (±10.7) Cedar Rapids,Iowa 27 70.4 3.7 0 0 0 0 25.9 0 0 (±17.2) (±7.1) (±16.5) Fitchburg/ Leominster,Mass. 37 64.9 2.7 0 0 21.6 0 10.8 0 0 (±15.4) (±5.2) (±13.3) ('_10.0) Paterson,N.J. 30 36.7 3.3 0 0 60.0 0 0 0 0 (±17.3) (±6.4) (±17.5) Sante Fe,N.M. 71 25.4 0 1.4 0 0 0 69.0 0 4.2 (±10.1) (±2.7) (±10.8) (±4.7) Yonkers,N.Y. 47 40.4 2.1 0 0 44.7 0 0 0 12.7 (±14.0) (±4.1) (±14.2) (±9.5) Utica,N.Y. 56 37.5 3.6 1.8 1.8 53.6 0 1.8 0 (±12.7) (±4.9) (±3.5) (±3.5) (±13.1) (±3.5) Raleigh,N.C. 49 44.9 10.2 0 0 4.1 2.0 34.7 4.1 0 (±13.9) (±8.5) (±5.6) (±3.9) (±13.3) (±5.6) Medford/ Ashland,Ore. 33 42.4 0 0 0 0 0 57.6 0 0 (±16.9) (±16.9) Bellingham,Wash. 56 41.1 5.4 0 0 0 0 50.0 0 3.6 (±12.9) (±5.9) ('_13.1) (±4.9) Overall 424 42.5 3.3 0.5 0.5 20.0 0.2 29.7 0.5 2.8 (±4.7) (±1.7) (±0.7) (±0.7) (±3.8) (±0.4) (±4.3) (±0.7) (±1.5) *Students aged 13-16 years who reported they smoked one or more cigarettes during the 30 days preceding the survey. tUnweighted data. During October-December 1990,data highest among younger adults(i.e.,aged dents and third most popular among on preferences for cigarette brands 2534 years).Overall,the cigarette brand white 9th-grade students. among teenaged smokers aged 13-16 preferences of adult smokers were con- Reported by:KM Cummings,PhD,E Sciandra,MA, years were obtained from school-based sistent with known national market Roswell Park Cancer Institute;TF Pechacek, PhD, surveys of students from a random sam- share patterns (percentage of all ciga- State Univ of New York, Buffalo.JP Pierce, PhD, y Univ of California,San Diego;L Wallack,DrPH,Univ ple of 9th-grade classrooms in each of rettes sold in the United States, by of California,Berkeley.SL Mills,MD,Div of Cancer the 10 communities.The survey included brand-market share data are collected Prevention and Control;WR Lynn,DR Shopland,Na- both public and private schools and quarterly b a tobacco indust ana- tions]Cancer Institute,for the Community I Group; S- p p q y y industry tion Trial for Smoking Cessation Research Group;SE yielded representative samples of ap- lyst9). Marcus,PhD,National Institute of Dental Research, proximately 400 9th-grade students per Amon 9th-grade smokers across all National Institutes of Health.Epidemiology Br,Office p y gr p g � on Smoking and Health,National Center for Chronic community. Forty-six (96%) of the 48 10 communities, three cigarette Disease Prevention and Health Promotion; Div of eligible schools (i.e. schools with _-50 brands-Marlboro Camel and New- Health, Interview statistics, National Center for r > > Health Statistics,CDC. students in 9th grade)participated,and port-were consistently preferred(84%- 4129(86%,range:76%-91%)of the 4783 100%). Among the 424 teenaged smok- CDC Editorial Note:In both the TAPS eligible students completed the survey. ers who usually purchased their own and COMMIT surveys, at least 84%of Data in this report were limited to 9th- cigarettes, 180 (43%) purchased Marl- the adolescent current smokers who usu- grade students who reported they were boro, 126 (30%)purchased Camel, and ally bought their own cigarettes pur- current cigarette smokers(adolescents 85(20%)purchased Newport.In nine of chased one of three brands-Marlboro, who reported smoking cigarettes on 1 the 10 communities, one third or more Newport, or Camel. Brand preference or more of the 30 days preceding the of all 9th-grade smokers preferred Marl- is much more tightly concentrated survey) and usually bought their own boro cigarettes. among adolescent smokers than among cigarettes.Current brand use was mea- The preference for Camel and New- adult smokers in the 1988 COMMIT sured by responses to the question, port cigarette brands varied consider- baseline survey of adults and in the 1986 "What brand do you usually buy?" ably among communities. In five com- Adult Use of Tobacco Survey(AUTS)3 In all but one community, Marlboro munities (Santa Fe, Medford/Ashland, as well as in the overall market.9 was the preferred brand for at least 20% Bellingham, Raleigh, and Cedar Rap- Marlboro,Camel,and Newport were of adult smokers; in Raleigh, North ids)Marlboro and Camel were the most among the most heavily advertised cig- Carolina,the brand most popular among frequently mentioned cigarette brands. arette brands in the United States dur- adults was Salem. Winston was pre- In four other communities (Paterson, ing 199010; therefore, these data sug- ferred by more than 10%in six of the 10 Utica,Yonkers,and Vallejo),Newport gest that tobacco advertising may in- communities. Except for these three and Marlboro were the dominant ciga- fluence teenagers in their choice of preferences,cigarette brand use among rette brands. brands. adult smokers varied considerably Camel cigarettes were most popular In both surveys, Marlboro was the within and across communities; most among teenaged smokers in western and predominant brand used by adolescents. brands were mentioned by <10% of midwestern communities.Newport cig- Teenaged smokers may be attracted to smokers. In communities where the arettes were most popular among teen- the brand's image of strength and in- preference for Camels was high among aged smokers from communities in the dependence promoted in the long-run- adults (Santa Fe, Medford/Ashland, Northeast. Newport was the most pop- ning"Marlboro man"advertising cam- and Bellingham), use of Camels was ular brand among black 9th-grade stu- paign. JAMA,April 8, 1992-Vol 267, No. 14 From the CDC 1895 1 �G The regional preferences for Camel paign began, the brand ranked third tional items.12,13 In addition, schAl to- , and Newport brands among teenaged among teenagers surveyed in TAPS. bacco-prevention programs can play e smokers(regardless of race)were con- Other studies, conducted after TAPS, key role in reducing smoking initiation sistent in both surveys.A recent report report even higher rates of Camel and should include information about the from California showed a high rate of preference among adolescents,',' con- media's influence on smoking.11 Camel use among adolescent current sistent with the COMMIT survey re- smokers in that state.' These findings sults. References may reflect regional differences in ex- Cigarette brands that appeal to chil- 1.CDC.Reducing the health consequences of smoking: osure to cigarette brand advertising dren and teenagers also use promotions 25 years of progress—a report of the Surgeon General. P g g g P Rockville, Maryland:US Department of Health and and promotion. such as displays at sports and youth- Human Services,Public Health Service,1989;DHHS The preference of black adolescent and oriented events and distribution of pro- publication no.(CDC)89-8411. 2. CDC.Differences in the age of smoking initiation adult smokers for Newport is also con- motional items(e.g., T-shirts,posters, between blacks and whites—United States. MMWR sistent across surveys and may reflect and caps)that may appeal more to chil- 1991;40:754-7• the increased occurrence of mentholated dren and teenagers than to adults.12 One 3. CDC. MMMWR 19W;39:665,671-3.among ere cigarette advertisements targeted to of the national health objectives for the 4. Pierce JP,Gilpin E,Burns DM,et al.Does tobacco blacks."Further research is needed to year 2000 is to eliminate or severely JMA advertising1;2266:315young s g people to stat, smoking? determine whether preference preceded restrict all forms of tobacco product ad- 5. DiFranza JR, Richards JW, Paulman PM, et al. or followed such targeted advertising. vertising and promotion to which per- RJR Nabisco's cartoon camel promotes camel ciga- rettes too children.JAMA 1991;266:3149-53. The COMMIT data for adolescents sons aged -18 years are likely to be 6. COMMIT Research Group.Community Interven- indicate a slightly different pattern of exposed(objective 3.15).13 tion Trial for Smoking Cessation(COMMIT):summary of design and intervention. J Natl Cancer Inst brand preference than do the TAPS The forces that influence smoking ini- 1991;83:1620-8. data. The higher preference for Camel tiation are complex and may include ad- 7. Allen x,Moss A,Botman S,Winn D,Giovino G, among the COMMIT respondents eom- vertising s TAPS"peer influence and habits of Pierce : Teenage attitudes and practices survey :methodology and response rates(Abstract). pared with the TAPS respondents may family members.'-'-' The exposure of In:Program and abstracts of the 119th annual meeting reflect the difference in age composition youth to tobacco advertising can be re- of the American Public Health Association.Washing- ton,DC:American Public Health Association,1991. (adolescents aged 13-16 years compared duced by 1) prohibiting the use of im- s. shah BV. Software for Survey Data Analysis with 12-18 years) and sample frames agery in advertisements by allowing only (SUDAAN) version 5.30 (software documentation), Research Triangle Park, North Carolina: Research (the 10 U.S.communities compared with words and a picture of the product itself Triangle Institute,1989. the overall U.S. population). (i.e., "tombstone"advertising);2)pro- 9• Maxwell JC. The Maxwell consumer report: 1991 year-end sales estimates for the cigarette industry.The difference may,however,reflect hibiting tobaccosponsorship of sporting Richmond,Virginia:Wheat First Securities,1992. a growing effect of the "Old Joe" ad- and other events that have a substantial 10. Endicott RC.The top 200 brands.Advertising Age vertising campaign. Recent evidence youth audience; 3) prohibiting tobacco 1991;Nov 11:41-2. 11. Cummings KM,Giovino G,Mendicino AJ. Ciga- suggests that the advertising campaign advertising in publications that have a rette advertising and black-white differences in brand for Camel that began in 1988 and fea- substantial teenaged readership;4)pro- preference.2. Richrds JW Health Fischer 19PM.Smokescreen:102: 98-701.how tures a dromedary cartoon character ap- hibiting tobacco billboards located near tobacco companies market to children.World Smoking peals more to children than to adults.' schools and other areas where youths and Health 1990;15:124. Public Health Service. Healthy people na- In 1986r Camel ranked seventh among congregate (e.g., Parks and shopping dional health promotion and disease prevention the youngest age group (17-24 years) malls);5)prohibiting paid tobacco place- objectives—full report, with commentary. Washing- of smokers responding to the AUTS3; ments in movies and videos;and 6)pro- PCi:U Health ser;of Heart nxxHuman s publication in 1989,1 year after the advertising cam- hibiting tobacco advertising on promo- no.(PHS)91-50212. Surgeon General's Report on Smoking and Health MMWR. 1992;11:183 526 000 smoking-attributable deaths oc- bacco consumption and production; THE 1992 REPORT of the Surgeon Gen- curred yearly in the Americas; 100 000 and taxation and legislation)is crucial to eral,Smoking and Health in the Amer- of these deaths occurred in Latin Amer- development of a systematic program icas, was released on March 12, 1992. ica and the Caribbean. for prevention and control of tobacco The report, developed in collaboration 3. In Latin America and the Carib- use. with the Pan American Health Organi- bean, the current structure of the to- An executive summary of the Sur- zation, examines epidemiologic, eco- bacco industry, which is dominated by geon General's report is available from nomic, historical, and legal aspects of transnational corporations, presents a the Public Information Branch, Office tobacco use in the Americas. formidable obstacle to smoking-control on Smoking and Health, National Cen- The major conclusions of the report efforts. ter for Chronic Disease Prevention and are: 4. The economic arguments for sup- Health Promotion,CDC;telephone(404) 1.The prevalence of smoking in Latin port of tobacco production are offset by 488-5705. Copies of the full report are America and the Caribbean varies but the long-term economic effects of smok- available from the Superintendent of is 50%or more among young persons in ing-related disease. Documents, P.O. Box 371954, Pitts some urban areas;substantial numbers 5. Commitment to surveillance of to- burgh, PA 15250-7954 (SIN 017-001- of women have begun smoking in recent bacco-related factors (e.g., prevalence 00478-2 for the English edition and S/N years. of smoking; morbidity and mortality; 017-001-00479-7 for the Spanish trans- 2.By 1985,an estimated minimum of knowledge,attitudes,and practices;to- lation)for$12 each. 1696 JAMA,April 8, 1992—Vol 267,No, 14 From the CDC Dane Starling ice President ANHEUSER—BUSCH COMPANIES R Corporate Representative March 27, 1995 The Honorable Gayle Bishop Chair - Contra Costa County Board of Supervisors 651 Pine Street Martinez, CA 94553 Dear Ms. Bishop: Anheuser-Busch Companies, Inc. wishes to file the attached four pages in opposition to the proposed county ordinance amendment prohibiting advertising tobacco products and alcohol beverages on billboards within 2,000 feet of schools. Sincerely, Dane Starling Vice President and Corporate Representative Attachments Anheuser-Busch Companies, Inc. 1485 Response Road Suite 108 Sacramento, CA 95815 (916)920-0886 Anheuser-Busch Companies, Inc, Contra Costa County Proposed Ordinance to Ban Alcohol & Tobacco Billboards Within 2,000 Feet of Schools On March 28, 1995, the Contra Costa County Board of Supervisors will conduct hearings to consider prohibiting the placement of any billboards for tobacco products and alcohol beverages within 2,000 feet of any school. The Board should reject the proposed ordinance because it is too restrictive and is founded on ill-conceived principles of harm to children when they see such billboards. The proposed ordinance is flawed because: • It is founded on the unsubstantiated belief that the placement of billboards "near" schools in some way induces students to drink, despite the fact that school-aged children are perfectly well aware that drinking alcohol beverages is illegal. • It is based on the incorrect assumption that advertising is a cause of underage drinking. Much research has been done on this topic and there is no support for the proposition that advertising causes abuse or underage drinking. Advertising is not even on the list of factors considered by most researchers as a cause of underage drinking, but rather, the behavior of peers and parents are what most typically are credited as the main influences. • Yet, we are a responsible corporation and intend to make positive contributions to communities, wherever we do business. That's why we sponsor a number of programs that address underage drinking, including our"Family Talk"program which guides parents through the difficult discussions that we believe they should have with their children, as well as programs for more effective checking of identifications to make sure that licensed retailers only sell our products to people of legal age. We are also sensitive to perception and community standards. That's why, regardless of the scientific evidence,that we do not allow any of our billboards to be placed within 500 feet of a school anywhere in the United States. • At the very least,the ordinance is overly restrictive and excessively blunt. Two thousand feet is a distance of nearly seven football fields. A circle with a 2,000 ft. radius around each school would literally encompass an area of at least 288 acres. Such a large area, in some cases, would surely include highways, as well as other commercial areas that presumably are not what the supervisors have in mind for the ban. • The evidence shows that billboards do not cause underage drinking, but rather, they deliver information to adult consumers and help to position brands favorably at the expense of competitors' brands. The ordinance potentially will hurt retailers located in or near banned areas, and by comparison, it will aid those located elsewhere in the county as well as all retailers in adjacent counties In effect, the ordinance, if adopted, would unjustly determine winners and losers among retailers and taken as a whole, the retailers in the county would lose business to neighboring counties. • If our ultimate goal is to prevent underage drinking and other forms of alcohol abuse, restricting advertising is a waste of valuable time and energy. • No one wants kids drinking illegally. But moving a billboard 2000 feet away from a school is not the answer. • Young people recognize that drinking is a part of adult society. Many of these kids see their parents drink—and the overwhelming majority of those parents drink responsibly. In a predominantly drinking society, then,we can't hope to prevent drinking problems among our youth by attempting to shield them from the facts that(a) drinking takes place, (b)there are many different brands of drink, and (c)manufacturers advertise their brands to adults who choose to drink. • In fact, research authorities, adults and young people alike agree: Parents and peers are the primary influences on drinking decisions and behaviors—not advertising (see attached). • With that in mind,we should be focusing our time and energy on those things that will make a real and lasting difference on this issue—focusing on those actions that will prepare children to make smart choices about drinking: which include not drinking illegally and drinking responsibly when they become of legal age. • How can we do this? Members of the alcohol beverage industry are already working hard in this area and are willing to do more in cooperation with those who share a vested interest in stopping the problem. Let me provide you a few examples (see attached). • Again, it is on these types of significant programs that we should be focusing our time and energy if we really want to make a difference—not on empty gestures that unnecessarily penalize businesses who are as committed to addressing the issues as the rest of us. After all, these businesses are made up of people who, like the rest of us, drive the same highways and are working hard to raise good children. With the same interests at heart, let's work together —where it really counts—not at odds. ADVERTISING EVIDENCE Research Authorities • In 1993, Dr. Joseph C. Fisher published one of the most comprehensive reviews of the compendium of research available on this subject, titled, Advertising,Alcohol Consumption and Abuse: A World Wide Survey. In that book, Fisher states the following: In the beginning of this review, the effect of advertising was stated in the null form: Advertising does not affect alcohol consumption or abuse. Based on the studies in the literature it does not appear that,for practical purposes, the null hypothesis can be rejected but instead should be accepted. • In 1994,Dr. Morris Chafetz, the founding director of the National Institute on Alcohol Abuse and Alcoholism, reviewed Fisher's book for the New England Journal of Medicine, one of the most prestigious medical journals in this county (Volume 330, March 10, 1994,Number 10). In that review, Chafetz wrote: Fisher's conclusions are clear and solid and do not go beyond the data or the methods. He succinctly puts the question of advertising into perspective: when it comes to fighting for brand shares, advertising money is well spent, but if used to try to make people drink or drink abusively, advertising money is wasted...people can not be made to do something the are unwilling to do. • These conclusions are consistent with the experiences of our society. The abuse and misuse of alcohol beverages rises or falls irrespective of advertising expenditures. Adjusting for inflation, beer advertising expenditures more than doubled between 1976 and 1988. During the same time, per capita consumption remained relatively flat, while major indicators of abusive drinking (including drinking among high school seniors) entered a period of steady decline. Adults • Among nine things that might cause young people to start drinking alcohol, 65 percent of adults (age 18 or older) cited schoolmates and friends as the most important factors. (Roper Starch Worldwide Survey, February 1994). Young People • Among six things that might influence their decisions about drinking, 61 percent of American Youth(ages 8-17) say that their parents have influenced them the most (Roper Youth Report, January 1994). PROGRAM EXAMPLES Family Talk About Drinking: Developed with education, counseling and alcohol research professionals, this award-winning program is listed in the Prevention Materials Database maintained by the U.S. Department of Health&Human Services. It offers a set of guidebooks (both English- and Spanish-language versions) and a videotape that promote effective parent-child communication on the issue. More than 1.5 million copies of Family Talk material have been distributed, and the program continues to be available, free-of-charge, through a national,toll-free number(1-800-359-TALK). Make the Right Call& Barbara Babb, a former critical-care flight nurse from St. Louis, delivers a graphic, factual and impressive presentation to high school and college students about the consequences of underage drinking and drunk driving. Anheuser-Busch sponsorship has helped Ms. Babb take her presentation to schools across the country, reaching hundreds of thousands of students—reinforcing positive peer pressure on the issue. Operation Identification: To help retail establishments better identify those customers they are legally able to serve,Anheuser-Busch makes available through its wholesalers,plastic bands that are snapped on the wrists of those who show a valid identification. Anheuser-Busch wholesalers also make available a pocket-sized checklist of tips to help retailers spot fake IDs(available in English, Spanish and Korean), and booklets that contain photographs of valid drivers' licenses from all 50 states. Additionally, the nationally recognized TIPS server-training program devotes a portion of its course to techniques for checking and identifying valid IDS. Check Into A Winning Life: Developed and presented by Bob Anastas (the founder and former executive director of Students Against Driving Drunk), Check Into a Winning Life encourages students from a wide range of interests—be they athletes, students council members or club leaders—to band together and establish a single, focused group, one that draws on their collective talents and helps them deal effectively with the various issues facing their school- using positive peer pressure,networking and conflict-resolution skills. Caring Connections: This step-by-step guide helps concerned adults—lay people, educators, school counselors, etc.—make a positive difference in the lives of"at-risk" children, particularly those from troubled homes (homes where kids may not have the chance for positive parental communication on the issue of drinking)., The guides were written by a team of authorities in education, student assistance programs and counseling. To date, Caring Connections has been embraced by a variety of organizations concerned about this issue, including the National School Boards Association,the American School Counselor Association, the National Association of Elementary School Principals and the National Association of School Nurses. MRR723-1995 16:48 FROM DISTRICT 1 SUPERVISOR -TO 61078 P.01 ��- • 1130AI ri OF SX)PERVISC RS- -� p CtJN TRA COSTA.COX NTY i. n: i JIM ROGM ti SUPERVISOR,FIRST DISTRICT March 23, 1995 Ronald Beals Legal Counsel Gannett Outdoor by FAX(916) 312-8601 Dear Mr. Beals: Thank you for your letter to Victor Westman, County Counsel, of March 21, 1995. In your second paragraph, you imply that there is adverse precedent concerning the proposed County law in the Metromedia v. San piQao case and "numerous cases that have followed". In fact, none of these billboard regulation cases are similar to the County law. The Posadas case-which is similar to the proposed law, is conveniently omitted from your letter. In Posadas v. i'uerto Rico 106 Ct. 2986 (1986) the US Supreme Court applied the Hudson test and upheld the ban on all ads promoting gambling in Puerto Rico. Rejecting the industry argument that a legislature must permit unrestrained ads for any legal product, the Court upheld the ban because of the legislature's belief that the ads "would Increase the demand for the product advertised:" The Court didn't require solid findings, much less sciontific proof of this belief. Supporting the common sense notion that ad bans affecting cigarettes and alcohol -which addict and harm thousands of minors every year-would be given greater deference, Justice Brennan (dissenting in the Posadas case) specifically reserved judgment about ad bans on "cigarettes, alcohol..." A comparison of the Posadas ad ban and the County ad ban reveals that the County can meet the test much more strongly than the Posadas case which was upheld.by the Superior Court: 1 st Part and 4th Part of so test: Is it a legal activity? is the ban limited to achieve its purpose? Although I assume gambling in Puerto Rico is illegal for minors, Posadas had no targeting. The County law identifies an area inn_'37th CTAPPT ROnM 77n ► RIC'HM()Nr] f:A Q4Rn_1;.913R . TF1 FP14nNF M101'474.3931 . FAX 1.x,101 37A-UPQ MAR-23-1995 16:48 FROM DISTRICT 1 SUPERVISOR TO 61078 P.02 (near schools) where minors go by every school day, and where they walk around while on breaks and while gotting a snack or a lunch. Because the billboards are still perfectly usable for all the ads, the effect of the ordinance on your ability to place cigarette and alcohol ads as part of a regional purchase is, to quote your letter, "not substantial". This is in sharp contrast to the 100% ban upheld in Posadas. 2nd Part of Hudson test: Does the ad ban seek to implement a substantial government interest? Although gambling may qualify as a substantial government interest, no one has even thought of accusing gambling of killing one thousand Americans per day(as cigarettes do), or of being involved in the overwhelming majority of serious violent crimes (as raloohol is). Your letter also fails to mention not only Posadas, but also the Dunaain case(716 F2d 775 cert denied)upholding a State ban on Mr alcohol ads and ween at (433 AE 2nd 138 cert denied 103 c.t. 31) upholding a State ban on liquor price ads. Your letter also reiterates the scare tactic that San Diego paid a$500,000 legal bill alter losing the Metromedia case, ignoring the distinction that Metromedia was extremely complicated litigation involving a broad package of billboard regulation and ignoring the distinction that San Diego's law was overturned. By the way, if ad bans near schools are so illegal, why did the billboard industry support the 500 foot ban passed by the Richmond City Council? Why did your allies In the cigarette industry include a 500 foot ban in their recent unsuccessful ballot initiative? Won't it be clear to a court that the billboard industry is on record supporting a similar law(in Richmond), and that you're concerned about the degree of the law(2000 feet versus 500 feet) not the mid, of ban (no cigarette and alcohol billboards near schools). The special nature of cigarette and alcohol ads, as they affect minors, is further evidenced by the absence of cigarette and hard liquor ads on N and radio. Billboards are the last best refuge left to reach minors e.g. the Joe Camel ads which moved Camel from a .05% share of the vital minors' market to a 33% share in a few years. Your letter invites us to get into a debate about how many cigarette and alcohol ads are within 2000 feet of schools. The question is not whether your industry has followed its usual practice of changing the types of ads when the political heat is on (by avoiding having cigarette and alcohol ads appearing on -.10 f,74 342-9 P.002 MAR--23-1995 16:49 FROM DISTRICT 1 SUPERVISOR TO 61078 P.03 your.billboards today). The question is whether kids in the future should be exposed to these ads. Your letter also invites us to pit one neighborhood against another by adding more billboards in areas that currently don't have them. The property owners near schools still can put up signs even if Gannett and Patrick agree not to. Also, they can challenge the law, since they are affected, as well as the billboard companies. Who knows what protection your Industry can get the State legislature to provide, so that we would be stuck with more new billboards (per your deal) and cigarette and alcohol billboards near schools ( per new statutory protection)? This is the same type of diversionary pit one neighborhood against another scheme that your industry (unsuccessfully)tried to foist off on the Richmond City Council. In regards to your threat of litigation, I note that the$500,000 scare tactic was the only argument that your polling indicated would defeat the 1991 Measure J in Richmond. Once your free speech challenge loses, the path will be cleared for Contra Costa cities to pass their own bans. That's great long term financial news for the County budget (due to avoiding cigarette and alcohol health and crime costs from minors who are encouraged to start illegal use early on). But is removal of the scare tactic fig leaf, and the opening of the floodgates of new laws really what you want? Why haven't you challenged a much broader Utah ad ban? Do you really want the ugly truth from scientific studies that cigarette and alcohol ads increase minors' use coming out in a highly publicized court battle? I would hope that Gannett would reconsider and join us in supporting protecting minors from ads for products that are illegal for minors, especially when scientific studies (and common sense) show that these ads contribute to addicting and killing children. Sinc rely, JIMROGERS First District Supervisor JR:hf cc: Board of Supervisors County Administrator(for distribution to relevant staff) Gounty Counsel TOTAL P.03 MAR-23-19 5 15:481 510 374 3429 P.003 MAR-27-1995 16:52 FROM PINOLE LIBRARY TO 6461059^� P.01 I 3241 Colusa Street Pinole, CA 94564 March ,27, 1995 Via Facsimile Transmission to: 64B-1059 Contra Costa County Board of Supervisors 651 Pine Street Martinez, CA Dear Members of the Board: I understand that a proposed ordinance to prohibit alcohol and tobacco advertising billboards near schools will be on your agenda tomorrow, March 28. I strongly urge you to _enact this ordinance. The tobacco and alcohol industries together spend billions of dollars each year to promote their products. A significant {portion of their advertising is designed to attract young users. This kind of advertising has no place near our schools. I hope you will stand up to the tobacco and alcohol industries on this issue and enact this measure. I hope that the cities in Contra Costa County then will follow suit by enacting similar ordinances . Thank you very much for your consideration of my comments. Very truly yours , Gerald H. Schlintz .10BACC O JQ � ° TOBACCO PREVENTION PROJECT a r CONTRA COSTA COUNTY HEALTH SERVICES DEPARTMENT O e `11.q CON°,M';=-1 A�o CRs A" Briefing Report: Youth and Tobacco Advertising With the Board of Supervisors considering an ordinance to ban tobacco and alcohol advertising on billboards within 2,000 feet of schools in the unincorporated parts of Contra Costa County, here is some information about how that advertising impacts on youth: i' ♦ Although the County has had remarkable success in reducing the prevalence of smoking among adults (from 26% to 19% since 1989), youth use of tobacco has not decreased. ♦ Experts agree that there are four major causes for teen use and addiction to tobacco, including the widespread and high level tobacco sales to minors, parental and family influences, peer pressure and tobacco advertising and promotion. ♦ Ninety percent of adult smokers began smoking before age 19; sixty percent were smoking by age 14. The most vulnerable age is even younger - between 12-16. ♦ Susceptibility to smokeless tobacco use by adolescent boys - the absence of a clear decision not to use - was higher for this county than the rest of the state, especially among non-Hispanic whites. ♦ Exposure to peer smoke among adolescent rose from 33 percent to 47.1%. Exposure to tobacco advertising also rose and now equal the rest of the state. ♦ More youth recognize Joe Camel than Mickey Mouse. ♦ In Contra Costa County, more 12-14 and 15-17 year olds recalled Camel, Virginia Slims, Winston, Salem and Newport brand advertising than in the rest of the state. ♦ Teens exposed to advertising are more likely to perceive benefits to smoking and are more likely to intend to smoke in the future. ♦ Adolescents with a high exposure to cigarette advertising are as much as twice as likely to be smokers as those reporting a low exposure. ♦ Although selling tobacco to minors is illegal in California, the tobacco industry nationwide sells more than 947 million packs of cigarettes and 26 million containers of smokeless tobacco to 3 million kids annually - worth $1.26 billion annually. (More) 597 Center Avenue,Suite 325 ♦ Martinez,CA 94553 ♦ (510) 313-6522 ♦ FAX: (510)313-6219 More about Youth and Advertising.... ♦ The tobacco industry is one of the leading advertisers on billboards - second in 1993 behind retail stores. ♦ The tobacco industry spends an estimated $13 million in Contra Costa County to market and promote their products. By contrast, the voter-mandated cigarette tax (Prop 99) provides less than $150,000 annually for tobacco education and prevention efforts. ♦ Research shows striking increases in the illegal youth market for cigarettes following the introduction of high profile advertising campaigns for tobacco. ♦ Sharp increases by teen age girls in smoking in the late sixties and early 1970's are linked to soaring sales of widely advertised cigarettes for women. ♦ Smoking among California teenagers 16-18 years old declined from 12.5% in 1975 to 6.2% in 1988. When Joe Camel was introduced in 1988, the rate of teen smoking climbed by .7% a year through 1990. ♦ Since RJ Reynolds began the Joe Camel campaign, Camel's share of the illegal children's cigarette market increased from .5% to an incredible 32.8%, representing $476 million in annual sales. ♦ An exclusive study by Advertising Age magazine and Gallup shows 68% of Americans believe cigarette ads influence children and teens to smoke. And 66% of the total, including 60% of smokers, believe some cigarette ads are specially designed to appeal to young people. ♦ Two-thirds of Americans want the government to impose greater restrictions on cigarette advertising. ♦ Publications aimed at readers under 18 do not accept ads for tobacco or alcohol. The billboard industry has no restrictions and often places tobacco and alcohol billboards near schools, churches, parks, playgrounds, health centers and sports facilities. For more information, call Julie Freestone at 510-313-6522. Contra Amalia Gonzalez-Del Valle Costa CSAP Project Director County (510)313-6389 FAX(510)313-638-t t .' - 4 Substance Abuse Administration 597 Center Avenue,Suite-"V 3 ZD Martinez,California 94553 �r US iA 1 r Board of Supervisors Public Hearing on Billboard Ordinance - March 28, 1995 The Community Substance Abuse Services and the. Community Partnership Prevention Project supports the Board of Supervisor's proposed Billboard Ordinance Billboard Advertisement �9 Since the W70s the leading advertiser on billboards has been the tobacco industry. Alcoholic beverages are second. In 1985,nine of the 10 leading billboard advertisers were tobacco or alcoholic beverage companies. Nation-wide a third of all billboard revenues came from tobacco or alcoholic beverages [Advertising Age 9/4/86 in Marin Institute Fact Sheet]. Substance Abusers in Contra Costa County In Contra Costa County eighty thousand residents are dependent on or are serious abusers of alcohol and other drugs. These substances were involved in 64% of the suicides, 85% of the homicides, and 58% of all fatal car crashes [1/21/94 PEHAB Report]. Juvenile Arrests by Alcohol Charges Nearly 1,600 juveniles were arrested in Contra Costa County on alcohol charges during 1989-1993. Following the statewide trend, alcohol arrests have been on a steady decline in this time period. From 1990 through 1992, the CCC juvenile alcohol arrest rate remained slightly above the state rate [48.4 vs 47.0 in 1990; 38.7 vs 37.3 in 1991; 32.9 vs 30.1 in 1992] in 1993 CCC juvenile alcohol arrest rate was 23.7 as compared to the State's rate of 27.1. Public drunkenness accounts for about half of the juveniles arrested on alcohol charges in West County and East County.Liquor law violation arrests make up half of juvenile alcohol arrest in Central County, 42% in East County, 68% in South County, and 78% in Lamorinda. Driving under the influence accounts for 31% of juvenile alcohol arrests in West County. . [CSAS Substance Abuse Problem Indicator Data] A regional breakdown of the type of disciplinary actions received ixcc's off- sale alcohol outlets shows that in West County, of the 88 disciplinary actions received by off-sale outlets, only about 18% involved minors. In sharp contrast, in Central County, of a total of 149 violations received by off-sale licensees in 1994, 60% of the violations. involved minors, and in East County, of the 84 violations received in 1994, 44 or 53% involved minors. The proportion of youth-related violations is largely a function of increased enforcement through the use of police stings as well as an indication that off-sale outlets in there regions are not sufficiently careful about limiting youth access to alcohol. . [CSAS Substance Abuse Problem Indicator Data] Death Classification by Alcohol Test Examining death classification by the breakdown of all investigated deaths testing positive for alcohol reveals that of the 853 deaths where alcohol was found to be present in the victim,the largest percentage 27% were homicide victims, 24% were either motor vehicle or other accident victims, 22% were victims of suicide, and in 3% of the cases the death classification was undetermined. . [CSAS Substance Abuse Problem Indicator Data] Risk Behaviors and Social Norms Many factors play a role in drinking and smoking problems,but one clear contributor is the slick advertising campaigns that alcohol and tobacco companies aim at women, youth, and minority communities. The glamourous images in the ads,which are particularly enticing to young people, saturate inner-city neighborhoods across the country. Such advertising pervades our entire society, assaulting each of us dozens of times each day from billboards, bus cards, television, magazines, and a host of other sources. Drive through any inner-city, ethnic neighborhood and one of the first things you'll notice is a huge number of billboards carrying ads for alcohol or tobacco products specifically targeted to youth, women, and ethnic minorities. We need to recognize that tobacco and alcohol are gateway drugs encouraged through advertising and mass media promotional campaigns. We need to remember that alcohol remains the most popular drug, that there is no indication that drinking is declining among California youth, and that with alcohol and drug abuse comes violence [Miller San Francisco Examiner 3/22/95]. For example, a recent policy paper from the California Elected Women's Association for Education and Research shows that in 80 percent of crimes, alcohol has been used within the past 72 hours. In another report issued by the Robert Wood Johnson Foundation, as many as two thirds of homicides and serious assaults involved alcohol. Socio-Economic Factors You'll probably also notice that the concentration of bars and liquor stores is far higher in communities of color than in comparable White neighborhoods [Center for Science in the Public Interest]. The corner liquor store, a fixture in inner-cities communities across the United States is a community institution well known by local citizens long before researchers began to examine the proliferation and over-concentrateof alcohol outlets. Equally obvious to inner-city residents is the knowledge that alcohol outlets are magnets for a host of public health and safety problems - shootings, assaults, drug dealing, gambling, prostitution,harassment of passerby,public drinking and drunkenness, vandalism, graffiti, and loitering. We cannot dismiss as mere coincidence that African American and Latino communities are most affected. The over-concentration of alcohol outlets and billboard advertising appears to be one aspect of an overall social and economic policy that serves to oppress inner-city communities and to a large extent it is a form of institutional racism [Mosher and Works "Confronting Sacramento: State Preemption,Community Control, and Alcohol-Outlet Blight in Two Inner-City Communities"]. Community Strategies Rather Than Individual Solutions • Alcohol availability -where people buy and drink alcohol- and communities with alcohol problems are closely related. Because the community level is where most alcohol problems must be dealt with on a day-to-day basis communities which act vigorously to regulate alcohol availability are better able to manage alcohol problems than inactive ones. The key for change is collaborative efforts between elected officials, public agencies, and community groups working together. Cities and counties have substantial powers to control problems of alcohol availability through local planning and zoning ordinances which can be used to address problem of retail alcohol outlets and advertising, and in local city codes affecting the use of public places for drinking can be used as effective community based prevention strategies. The alcoholic beverage industry continues to expand its activities, it creates new products and campaigns aimed at expanding the times, places, and occasions where it is acceptable and desirable to have a drink. As this marketing activity expands, it is important that communities make clear their own standards, build effective grass-roots support, encourage local political leaders to become actively involved, and persuade city councils to enact creative legislation to protect its citizens and to promote healthy communities. Remember • Power comes in two forms: money and people - The alcohol industry spends a large amount of money because they represent a small number of people. The focus should be changing alcohol policy through community organizing: citizen groups, the recovering community, parents, government administrations, health providers, legislators, educators, and youth themselves. These constituencies represent an enormous number of committed people directly affected by the problems and willing to work for an environmental approach to the prevention of alcohol problems. • The consequences of alcohol use extend far beyond the individual with alcohol problems. Consequences include alcohol related fatalities and injuries, birth defects, violent crimes and suicide, the social costs resulting from reduced productivity, premature death and the expenses of treatment and support,and the incalculable cost of human suffering. • THE PARTNERSHIP APPROACH FOCUSES ON ENVIRONMENTAL FACTORS WHICH ENCOURAGE AND REINFORCE ALCOHOL RELATED PROBLEMS RATHER THAN ON INDIVIDUAL FACTORS.EMPHASIS IS ON GROUPS AND COMMUNITIES AT RISK, EMPOWERING COMMUNITIES, AND REDUCING HARMFUL COMMUNITY CONDITIONS! FACT SHEET: ALCOHOL AND TOBACCO ADVERTISING ON BILLBOARDS Since the 1970s, the leading advertiser on billboards IE B `r• a D has been the tobacco industry. Alcoholic beverages are second. In 1985, the latest year for which figures were OCT 2 S 19$9 available, nine of the 10 leading billboard advertisers " .k,kTE IN t!t'S,i.ITI.� were tobacco or alcoholic beverage companies. Nation- EZE��O>iJCL� CN7iRI wide a third of all billboard revenues came from tobacco �� •jfUl 1; °' ' or alcoholic beverages. (Advertising Age, 9/4/86). The combination of alcoholic beverages, tobacco and billboards has created a number of problems: 1. Billboards are the most intrusive and offensive form of advertising. It is the nature of billboard advertis- ing to force motorists and pedestrians to look at the ads. No one is forced to read ads in newspapers and maga- eines, but these ads are unavoidable when displayed on billboards. Billboards make the American public a "captive audience" to tobacco and alcoholic beverage advertising while at the same time blighting residential _r.. ... neighborhoods. 2. Unlike the print media, the billboard industry exer- cises no meaningful restraints on the placement of its ads. Publications aimed at readers under age 18 do not serve this function. A :onsumer may briefly glance at a accept ads for tobacco or alcoholic beverages. The bill- magazine ad once. but saturate a neighborhood with board industry, on the other hand, exercises no such billboards and the consumer will see the same ad 30 times restraint over the placement of its ads. They simply put a day, every day. It is no surprise then that in many low- billboards anywhere and everywhere they can. Today income areas more than half of all billboards carry ciga- billboards touting cigarettes and booze can be found rette and alcoholic beverage ads. Likewise, sports next to homes, schools, churches, parks, playgrounds, stadiums, neighborhood shopping areas and mass health centers, in sports stadiums, shopping centers, on transit systems are an increasing showcase for tobacco taxi stands and buses, alone rural roads and city streets and alcoholic beverage ads. —literally everywhere. 5. Although billboards are the only tobacco ads 3. The billboard industry is saturating low-income which the public is forced to look at, warning labels on neighborhoods with billboards advertising tobacco and billboards are either impossible to read or non-existent. alcoholic beverages. This has resulted in thousands of For example, print ads for smokeless tobacco include a new billboards being placed indiscriminately in ethnic new set of health warnings but billboards are exempt. neighborhoods. A much higher percentage of billboards Likewise, health warnings on cigarette ads are virtually in ethnic neighborhoods advertise alcohol and liquor invisible. Unless a motorist stops his car, gets out, and products compared to billboards in other locations. For walks up to a billboard he cannot read the health warn- example, during the first nine months of 1986, liquor ing. Ironically, Americans are being bombarded by bilI- marketers spent nearly 16 times as much on"junior bill- board ads for cigarettes but they almost never see the boards" directed at Blacks as they spent on advertising health warning which is virtually invisible during the six .; to the general public. A 1987 survey conducted by the to ten seconds a motorist has to observe a billboard from City of St. Louis found twice as many billboards in black a moving car. neighborhoods as white. Almost 60014 of the billboards 6. While a total ban on tobacco or alcoholic beverage in the black neighborhoods advertised cigarettes and alcoholic beverages. advertising may be constitutionally suspect, billboard C fcontrols including a total ban on all commercial bill- 4. As the most visible form of alcoholic beverage and boards have been repeatedly upheld by both federal and S cigarette advertising,outdoor advertising is unique in its state courts.In the case of Metromedia v.San Diego,453 t ability to reinforce drinking and smoking as social U.S. 490 (1981), the U.S. Supreme Court ruled that a norms. The best way to maintain the smoking popula- community may totally ban all"commercial billboards" tion and encourage drinking is to reinforce the cue to as a legitimate time, place, and manner restriction on smoke or drink in as many places as possible. Billboards speech. For additional information contact. SCENIC AMERICA 216 7th Street, SE Washington. D.C. 20003 (202)546-1100 oaaa 1212 New York Avenue, NW, Suite 1110 Washington, D.C. 20005 OUTDOOR ADVERTISING ASSOCIATION OF AMERICA CODEOFADVERTISING PRACTICES Outdoor Advertising delivers an advertiser' s message to the consumer. This role in the arena of public discourse requires both a vigilant defense of free speech and a sensitivity to contemporary community standards. We of the Outdoor Advertising Association of America recognize the need to balance these demands and recommend that each OAA-A member company formally adopt standards that reflect the following code of advertising practices: 1. Establish exclusionary zones which prohibit advertisements of all products illegal for sale to minors that are either intended to be read from, or within 500 feet of, established places of worship, primary and secondary schools, and hospitals. 2 . Establish voluntary limits on the number of billboards in a market that may carry messages about products that cannot be sold to minors. 3 . Continue to assert the right to reject creative content that - is in poor taste, misleading, sexually explicit or overly suggestive. 4 . - Intensify our traditional commitment at both the national and local levels to display public service messages from non- profit organizations. 5. Encourage greater diversity of advertised goods and services in all markets. The OAAA will establish a , national council composed of local leaders to advise us on community issues. Outdoor Advertising Association of America. Inc. Tel: (202) 371-5566 Fax: (202) 3, 1-1284 � y for 1Costa Hls� In 1989 and 1990 Contra Costa County convened several county-wide Alcohol and Drug Summits. As a result, a county-wide Action Plan was development for prevention, treatment and law enforce- ment related to substance abuse. In 1990 the county Board of Supervisors adopted the Contra Costa County Action Plan,which was then taken to the polls and ratified by the voters in 1990. The Action Plan called for each regional district to form a Partnership Coalition comprised of community residents and providers of treatment, prevention and law enforcement services. These coalitions were formed to develop and support community-wide strategic plans for prevention of alcohol, tobacco and other drug abuse (ATODA) in the four major regions of the county: East,West, Central and South. At the county level, the Partnership Forum, a federation of representatives appointed by the regional coalitions and the Substance Abuse Advisory Board, promotes the Community Partner- ship principles and provides a vehicle for coordination of county-wide prevention initiatives and activities. Fug In 1990 the Contra Costa Health Services Department applied for, and received, a grant from the federal Center for Substance Abuse Prevention (CSAP). CSAP funds 252 Community Partnerships across the country. The grant for the Partnership for a Drug-Free Contra Costa was for$500,000 per year for five years, for a total of$2.5 million divided among the four major regions of the county- Pu ose: The Partnership was formed to facilitate communities in acquiring and effectively using knowledge, skills and resources necessary to respond to alcohol, tobacco and other drug abuse, and its related problems. Approach: The Partnership effort is a collaborative process where community members are considered to be :'XP__;L-�, .L;bponsible for defining the nature and scope of alcohol, tobacco and other drug abuse problems. Agency and community advocates work together to plan, execute and evaluate prevention activities to meet the specific substance abuse problems in their region. Vision: The Partnership is working towards a drug-free Contra Costa County, with commitment among members to involve each and every member of the community (including all ethnic, cultural and socio-economic groups) to work towards a safer and healthier environment by reducing alcohol, tobacco and other drug abuse, and its related problems. C€}... .. OST VISION A Community Partnership working towards a drug free Contra Costa County. MISSION Each and every member of the Contra Costa County community will work towards a safer and healthier environment by the reduction of Alcohol, Tobacco, and Other Drugs [ATODA] use and abuse and their related problems. PRINCIPLES The Partnership is based on a bottoms up approach to prevention. This principle is practiced through the following actions: o Doing with the community rather than doing for them. o Create opportunities for capacity building. o Promote shared leadership. The Partnership values diversity and respects differences. These principles are practiced through the following actions: o Ensure inclusion of all ethnic, cultural, and socio-economic groups. o Create opportunities for dialogue, reflection, and action. o Implement a multi-sector community-wide prevention strategy. o Promote strategic alliances and collaborative efforts. o Develop partnerships that share risks,resources,responsibilities, and rewards. The Partnership is intentional in its commitment to community empowerment. This principle is practiced through the following actions: o Acknowledge differential power status between groups. o Develop democratic structures to ensure group representation. 0 Ensure equal access to resources, opportunities, and choices. o Create a communication system that is open and inclusive. o Implement policies that encourage conflict resolution and problem solving. o Promote relationships that are predicated on responsibility, experience, and skills. �c MANUAL FOR COMMUNITY PLANNING TO PREVENT PROBLEMS OF ALCOHOL AVAILABILITY September 1, 1988 Prepared by: Friedner D. Wittman, Ph.D. Patricia Shane, M.P.H. This Manual has been prepared as a report at the Prevention Research Center of the Pacific Institute for Research and Evaluation under contract from the California Department of Alcohol Programs (Contracts Nos. A-0154-5, A-0043-6, and A-0024-7). Additional support was provided by National Institute on Alcohol Abuse and Alcoholism Research Center grant AA06282 to the Prevention Research Center, Pacific Institute for Research and Evaluation. Dissemination will be made in conjunction with the U.C. Berkeley Institute for the Study of Social Change. ti C1 APTERI OVERVIEW: COMMUNITY PLAJ iNING FOR ALCOHOL AVAILABILITY Alcohol Availability in California Communities. Alcohol in everyday life. Alcoho is beverages are a widely-accepted and ever-present part of daily life in California communities. People drink to celebrate at weddings, births, graduations, promotions, holidays, and funerals; they drink to relax at the nd of the workday and on weekends; they have a drink over lunch with fellow workers and business associates; they drink to be seductive on dates an I to be sociable at parties; they drink at home as part of family traditions; t iey drink for religious reasons; they drink for personal reasons; they drink imply because others around them are drinking. About seventy percent of Califorria adults drink (Fay, l.ipow, Fay, 1984). On the average, Californians over 15 years of age drink about 3.2 gallons of absolute alcohol per capita each ear (U.S. Alcoholic Epidemiologic Data, 1985). About 44 percent of California adult drinkers have a drink at least once a week, and 9 percent drink ive or more drinks in one sitting at least once a week (Wallack, 1978). A,though consumption levels in California have been steady or have decline slightly in the last few years, in keeping with national trends, California's I ier capita consumption rate is still in the top fifteen percent among the 50 states (DISCUS, 1985). California high school students report that by the twelfth grade, they drink almost as much as adults. One study conducted in a suburban high school indicates that over 90 percent of high school seniors report having used alcohol, twenty-seven percent within the last week (Skager, 1986). As many as 30 percent report having ridden in a car with a driver who had been drinking (Wittman, Shane, and Grube, 1987). Alcohol availability at the community level. Alcohol is readily available throughout the variety of commun ty settings in which Californians live day to day. Alcohol is widely present in both private settings (such as homes, private clubs, workplaces) and pu blic settings -- public places and public events. Public places where alco olic beverages are permitted to be sold and/or consumed include bars, estaurants, grocery stores, convenience stores, specialty boutiques, colle a student unions, gas stations, hotels, beaches, parks, sidewalks, parking, lots, sports stadiums, airplanes. Public events are community occasions hat are, open to the public: alcohol is available at county fairs, block parties, Superbowl victory parties, the Olympics, various holiday celebrations throughout the year, fund-raisers for community charities. Altogethe , assumin& California follows national patterns, approximately 46 percent f all drinking in the state occurs in public places or at events attended by tJ e public (Clark, Midanik, and li{nupfer, 1981:201). 4 Three kinds of availability. California's abundant availability at the community level may be characterized in three ways. (1) Retail availability is provided by the distribution of retail alcohol outlets in the community. In addition to traditional bars and restaurants, an astonishing variety of stores sell alcoholic beverages, ranging from flower shops to gasoline stations. Alcoholic beverages are priced at the same levels as soft-drinks, milk,juice, and coffee, and are often packaged, promoted, and placed on shelves to compete with these other beverages. On the order of 25 percent of all commercial retail outlets in a community sell alcohol beverages (Wittman and Burhenne, 1985). About 10 to 15 percent of all taxable retail sales go toward purchase of alcoholic beverages (Benham, Wittman and Wallack, 1983). (2) Public availability is provided by cities and counties which permit drinking in public, and whichpermit the use of alcoholic beverages in public facilities such as parks, beaches, sports stadiums. Public availability also includes alcoholic availability at public events and special occasions. Many Californians do not realize that although the state does not permit public drunkenness, it is left to cities to regulate drinking in public. (3) Social availability is determined by generally-accepted, commonly- followed drinking patterns and practices in the community. Happy hours, drinks before dinner, drinking over business deals, teenage "kegger" parties, New Years' parties, religious confirmations, and graduation celebrations are examples of occasions and circumstances at which drinking is customarily and routinely included in community activities. Beverage industry and retailer contributions to availability. Retailers and beverage producers are constantly on the look-out for new times, places, and occasions at which to sell alcoholic beverages. The last few years have seen rapid growth in the sale of alcoholic beverages in gasoline station "minimarts." Applications have been made for alcohol sales licenses in fast- food restaurants, flower shops, donut shops, and mortuaries. The beverage industry has produced highly-portable beverages, such as pre-mixed cocktails with tops that may be used to drink the beverage, and new products, such as wine coolers, alcohol/soft drink combinations, and light beer, which are intended to compete with soft-drinks and other non-alcoholic beverages. In a word, alcoholic beverage retailers are quick to open new outlets as market opportunities warrant. The California of Alcoholic Beverage Control Department functions primarily to ensure orderly, legal activity in the sale of alcoholic beverages, rather than to limit or restrict such sales (Morgan, 1979). In 1987, approximately 70,000 retail alcohol outlets are available for 26 million Californians; this works out to about one outlet for every 370 citizens, or about one outlet for every 200 drinkers. Over the past few years, the number of places selling beer and wine have expanded at a greater rate than the growth of the state's population (Alcoholic Beverage Control, annual reports on retail licenses). 5 i Alcohol Problems in California. Although most Californians handle their drinking well, the use of alcohol brings with it enormous health, social, and economic problems. Below is provided a summary of alcohol problems facing Californians. See Appendix A for further discussion of causes and extent of these problems. Medical and health problems include: Approximately 3,400 Californians die each year from alcohol cirrhosis and related diseases. Alcohol is a major factor in deaths from heart disease, cancer, and respiratory diseases (U.S. Public Health Service, 1987). Approximately 1.62 million Californians have serious drinking problems, including alcoholism, if one accepts the widely-accepted figure that about 10 percent of adult drinkers fall into this group. Alcohol is estimated to be a factor in 50 percent of traffic fatalities and injuries, which includes over 70,000 Californians annually (California Highway Patrol, 1985). Alcohol-related traffic accidents are the leading cause of death for men aged 15-24 years of age (FARS, 1985). Approximately 342,000 Californians are arrested each year for driving while intoxicated (Hepperle and Klein, 1985). Alcohol is involved in approximately 66 percent of all fatal accidents, 45 percent of drownings, 50 percent of skiing accidents, 53 percent. of fire deaths, 37 percent of suicides, and 22 percent of home accidents (Klatsky, et al, 1981; Haberman and Baden, 1978). Alcohol-related birth defects are a leading cause of infant mortality and morbidity (Petrakis, 1985; U.S. Dept. of Health and Human Services, 1987). Social and economic costs of alcohol are also enormous: Disturbances from public inebriety and illegal behavior in public on or about the premises of alcohol outlets are of concern to local officials in 40 percent of California cities (Wittman and Hilton, 1987). Family disruptions and family violence are estimated to be involved with alcohol in more than 50 percent of cases. This includes 56 percent of fights and assaults in homes and 60 percent of child abuse. FBI figures indicate that 70 percent of murders, 41 percent of assaults, 50 percent.of rapes and 55 percent of all arrests involve alcohol. Federal Bureau of Justice Statistics Bulletins report that nearly 50 percent of convicted jail offenders had been drinking just before the current offense (Kalish, 1983). Informal estimates from local police departments in California communities suggest that fifty to seventy percent of misdemeanors are alcohol-involved. The cost of treatment programs to serve people with alcohol problems runs into billions of dollars nationally, and into tens of millions in California. National costs for hospital-based treatment of alcohol-related disorders have been estimated at $9.487 billion annually (Harwood, 1984:34). The California Department of Alcohol and Drug Programs (DADP) in 1986-87 budgeted $55.1 million per year for prevention, treatment, and rehabilitation 6 services to people with alcohol problems ($40.9 million state, $13.3 million federal, $0.9 million reimbursements). DADP Funding of $43.3 million to the counties is considered to have generated $112 million in alcohol services ($40 million for residential treatment, $36 million for non-residential treatment, $36 million indirect services) (Department of Alcohol and Drug Programs, 1987:54). Relation of alcohol problems to alcohol availability. From alcoholism to alcohol problems. Since Repeal of Prohibition, alcohol- related problems have been attributed primarily to alcoholic individuals, and the retail availability of alcoholic beverages has been thought not to be an important factor in the incidence or prevalence of alcoholism and other alcohol-related problems. Public health research into alcohol problems at the community level over the past twenty years has challenged this belief. As Figure I.1 shows, this research suggests that most alcohol problems are associated with the general drinking population, rather than with alcoholics per se (Moore and Gerstein, 1981). Alcohol availability and alcohol problems. The research also indicates that increases in levels of economic and physical alcohol availability increase the likelihood of alcohol problems in a community (Smart, 1982; Grossman and Coate, 1985; Room 1984; Watts and Rabow, 1983; MacDonald and Whitehead, 1983). Even the drinking of those considered to be alcoholic (e.g., public inebriates) has been shown to be influenced by changes in alcohol availability such as liquor strikes (Room, 1984) and wartime restrictions (Smart, 1974). Appendix A illustrates further the relationships between types of alcohol problems and kinds of alcohol outlets. Local controls on alcohol outlets to prevent community-level alcohol problems. In light of this information, California communities are faced with re-evaluating the community's many alcohol outlets and drinking occasions which have long been taken for granted to be non-problematic. The community's alcohol problems have been considered the responsibility of health professionals and self-help groups. Problems with alcohol outlets have been seen as the responsibility of the California Department of Alcoholic Beverage Control (ABC) rather than as a local responsibility. In general, the wishes of retailers to open new alcohol outlets have not been seriously questioned, and alcohol outlets have increased steadily over the past decade as fixtures in California communities. Recent public health research suggests that local agencies and community groups can do much more than previously thought to take effective responsibility for the community's own patterns of alcohol availability. Now that problems with availability appear to be a function both of people's uses of alcohol and the settings associated with those uses, effectiveprevention approaches require attention to both. Retail availability of alcoholic beverages must be monitored more closely for potential alcohol-related health and social problems. 7 FIGURE 1.1. ALCOHOL—RELATED PROBLEMS AMONG LIGHT, MODERATE, AND HEAVY DRINKERS ALCOHOL-RELATED PROBLEMS WITH F,RCFNdSTS,4 FGA4 BFCUGFR VCF P0USF X06 UGHT Rc MOOEgA,E -- OpINKERS HEAVY RS ;,pPINKE CONTRIBUTION OF DRINKERS TO REPORTED PROBLEMS FIGURE 2.1 National surveys reveal that the majority of certain alcohol- related problems affect moderate drinkers--those having up to two drinks per day. Heavier drinkers are more likely to suffer some sort of alcohol problem than are moderate drinkers, but there are so many more moderate drinkers that their contribution to the total problem adds up. Source: Uni- versity of California,Social Research Group(Berkeley. CAlifornia). . 8 Regulatory activity by the California Alcoholic Beverage Control Department. The California Alcoholic Beverage Control Department has lost much of its former strength for dealing with alcohol outlets and related problems of availability. The ABC in 1986 had less than 150 officers statewide to oversee the operations of more than 70,000 alcohol outlets. This amounted to one-tenth the effective strength in sworn ABC field officers compared to the ABC's 1955 field-officer staffing level. From 1955 to 1986, District Office investigator strength declined from about one officer per 43 outlets to one for each 466. Recently, steps have been taken to reverse the decline in ABC's personnel. The Department's 1987-88 budget augmentation created funding for an increase of 64 positions statewide (50 investigators and 14 support staff). While the ABC has experienced a decline over the past thirty years, California state law toward alcoholic beverages has become increasingly liberal, and state taxes on alcoholic beverages have not kept pace with inflation (Morgan, 1979; Mosher, 1979). Alcohol has thus become increasingly available, and consumption has generally increased over the past twenty years, while state-agency control staff has declined in relation to numbers of outlets. Emerging interest in uses of local resources to regulate alcohol availability. City and county powers to regulate alcohol availability. Cities and counties have the power, through their local planning and zoning ordinances, to regulate the availability of alcoholic beverages to protect the "health, safety, welfare, and morals" of the community's citizens. These local powers are used in conjunction with the California Alcoholic Beverage Control Department's own powers to license the sale of alcoholic beverages. ABC District Offices routinely work with city and coun planning and police departments in the licensing and enforcement oalcohol outlets (see Chapters H and III of this volume for details). The ABC officially encourages close working relationships between its District Offices and local governments (Stroh, 1984). Although they have significant powers to regulate alcohol availability, most California communities do not take full advantage of them. Wittman and Hilton (1987) have shown that 41 percent of cities do not have any local planning and zoning ordinances available for the regulation of alcohol outlet availability. Twenty-nine percent have conditional use permit requirements for on-sale alcohol outlets only. About 30 percent of cities require conditional use permits for both on- and off-sales outlets. Fewer than 26 percent of cities have text in their ordinances with specific restrictions on alcohol outlets, and about 18 percent have ordinances that include both conditional use permit requirements and text restrictions. Local regulation of alcohol availability to prevent alcohol problems. As problems associated with alcohol availability have come to light, communities have become more interested in the preventive regulation of their alcohol outlets. Generally speaking, communities become active in using their local ordinances to regulate alcohol outlets when local community 9 groups become concerned about alcohol problems associated with outlets. These groups work closely with local agencies and with the ABC to establish appropriate restrictions in the local planning and zoning ordinance. The community groups then become active in seeing that their local ordinances are actually used for preventive purposes. Local officials are also interested in making greater use of local powers and resources. Asked whether local restrictions on alcohol outlets should be more or less restrictive, 47percent of California city planning officials responded "more restrictive"; while 2 percent responded "less restrictive" (Wittman and Hilton, 1987). Certain California communities have made extensive use of their local powers report and are encouraged that their efforts are reducing or have reduced problems related to alcohol outlets. Oakland, Los Angeles, Tulare and Oceanside are four highly instructive examples that are described in detail in Appendix B. Other communities that have taken their local outlets seriously are preparing preventive ordinances, or have recently instituted ordinances which are expected to show results in the near future. Low-income minority communities are among the more active communities in which local groups have taken initiative to develop local ordinances to supplement ABC controls. This.is not surprising ;since higher densities of alcohol outlets and higher alcohol-related crime rates are typically evident in low-income minority communities (Farrell, 1986; Watts and Rabow, 1983). Two of the Case-Studies in Appendix B involve minority communities in Oakland and in Los Angeles; other minority communities are working to follow their example. An alcohol recovery program in San Diego, for example, has taken the initiative to have its primarily Black and Hispanic district declared a trial planning area for local permit requirements. 10 THE MARIN INSTITUTE for the Prevention of Alcohol and Other Drug Problems Spring 1993 Writing Commentary- Clinton's Liquor license ii drug policy glut zones out Page 9 development Page 12 i Locals to Alcohol Industry: 'We're in Charge" hey looked up and saw • Dallas officials are thinking of introduc- cloud-to-cloud bill- ing an ordinance to control the location of boards exalting the use any business that derives more than 75 per- of alcohol and tobacco. cent of its revenue from alcohol. They looked across The mayor of Indianapolis is working the street and counted with legal specialists to identify ways to con- corner-to-corner stores trol local liquor outlet proliferation through selling food for too zoning code or nuisance litigation in the much and high octane beer and wine for too wake of persistent citizen pressure and failed little,attracting a sidewalk culture of brown- state legislation. bag"short dog"drinkers,drug dealers,pros- "I think this is a trend,"says Vic Coleman, ' �"t"~ ' ' titutes, and gamblers. vice chairman of California's statewide They looked at their kids and decided Council on Alcohol Policy. `"Ten years ago, that waiting for someone else to rescue them the idea of local communities using zoning _ v=, from an environment oversaturated with al- and such to control alcohol availability was a cohol was riskier than trying to do the job theory. Now, the ideas are being applied." {'"' 1~ themselves. Within the past five years,inner- Friedner Wittman, a research scientist city communities across the country have who in 1978 co-authored probably the first gone beyond the rhetoric of local control: how-to manual for communities on the sub- h • activists in Chicago,Baltimore,and Mil- ject, puts it this wav: "I wouldn't call this a waukee succeeded in removing more than forest fire type of a trend. It's analogous to 1,800 alcohol and tobacco billboards from starting a campfire,where little embers waft low-income neighborhoods that suffer dis- around,only to start little fires elsewhere. It proportionately from the effects of these is spreading." products. Inner-city alcohol outlet over-concen- 4 • Oakland,California,officials are to vote tration "makes the use of the term `ghetto' on whether to levv an annual fee on liquor more pervasive," says Los Angeles City stores that would pay for inspectors to en- Councilman Mark Ridley-Thomas. "But the force community standards for selling alco- pendulum is swinging. We are using the hol. weight of the law more creatively to assure • A California state legislator is pushing a that the interests of the community are measure handing South Central Los Angeles served." tools to prevent the rebuilding of liquor And the alcohol industry is fighting as stores incinerated after the first Rodney King hard locally as it does at the state level to verdict. protect what it sees as its constitutional right - - 4 S not to be targeted out of business. In the often place conditions reflecting community inner city,say industry representatives,their concerns on newer outlets. stores are employers in the community, too, and are not responsible for what happens in "This issue the streets around them. represents The federal government retreated from exploitation of the business of being in charge of liquor laws after the end of Prohibition in 1933.Today, a community, Of 14 states allow distilled spirits only in state- Many of them have been working on the G C/OSS,"says owned liquor stores. The rest reflect a issue for a decade or more — since varied Oakland City hodgepodge of alcohol rules. Many are like businesses and supermarkets moved out and Councilman California, where the state issues liquor li- the liquor industry moved in. Few know the Nate Miley. censes and theoretically enforces alcohol issue better than activists in South Central laws, but politicians have decided that Los Angeles,whose statistics have been seared "YOU don't see shrinking tax dollars belong in places other on the news since the first Rodney King this . . . in rich than the alcohol law enforcement budget. verdict. areas." Today, even a state enforcer like Carl South Central's liquor store density — 682 Falletta, assistant director of Alcoholic Bev- liquor licenses within 40 square miles—is 10times erage Control for southern California, opts that of the rest of Los Angeles county. for communities determining how many In the wake of the last year's events that stores selling alcohol can exist per block and left 45 people dead and more than 1,000 whether outdoor customer conduct meets buildings damaged or destroyed,many South community standards. Central residents got what they had tried to "Someone has got to have control over achieve by working through the system for a alcohol availability," he says. "The way it decade: More than 200 of 728 alcohol outlets stands now in California,there is avoid. Even were historv. if the state had the personnel to get out there, "It's too simplistic to blame all of South it is difficult to show the connection between Central's problems on alcohol outlets,"says the sale of alcohol and the activities that Sylvia Castillo,who is in charge of the Com- attract crime around the stores. Examine munity Coalition for Substance Abuse Pre- our history and you'll see that various forms vention andTreatment's campaign to rebuild of the alcohol industry have put a kibosh on South Central without liquor stores. "But we such enforcement." want the alcohol and other drug issue fac- Falletta and Dr.Richard Scribner of the tored into the rebuilding — that there are University of Southern Californiasay that the certain dangerous land uses that foster more available alcohol becomes, the bigger criminal activities and bad health." the number of people who drink and commit South Central took its first formal look at crimes grows. Scribner is studying the issue alcohol saturation in 1983,when a coalition on a federal grant. of 22 churches examined crime associated "In 1990,"Scribner says, "for each addi- with alcohol outlets. Early efforts to deal with tionalliquorstoreincities of50,000residents, problems through increased lawenforcement there were 2.5 additional major crimes such failed. The people did their surveys and as murder,robbery or assault over and above made their demands. The City of Angels the other alcohol-related crimes." responded in 1984, with a use permit spe- Activists in these neighborhoods say their cifically for that area. biggest problem is the alcohol outlet that has In California,thev call it the conditional existed for decades,exempted from laws that use permit and when applied to the subject of 2 The Marin Institute Spring 1993 .:� . €ci L IL r ;11 alcohol outlet density, it can mandate mini- mum distances between liquor stores and schools, residences, and churches; arkin P g J=, lot lighting; security guards; earlier closing r ;f S'• hours than required by the state;and limita- tions on advertising and display of alcoholic " K beverages. The city can force liquor store owners to stop selling individual cups and remove graffiti and outdoor couches that encourage all-day, all-night loitering. In 1985,the same conditional use permit process extended to the rest of the city.Three years later,Los Angeles added an amendment '"" allowing it over time to revoke existing liquort outlet licenses if citizen complaints or evi- dence could prove a pattern of violations. Due to lack of inspectors,in the past five years only 12 offending outlets unable to comply in a reasonable time have been put out of ' r l !t business, according to the city zoning ad- , i- .� P ministrator. Although the city's planning commission has given permission to rebuild to 28 demol- ished liquor stores,Castillo says the hearings T r have given the communiryan unprecedented say over day-to-dav store operations. Ridley- Thomas says he hopes plans to use the envi- ronmental impact report process succeed as ra an additional way to keep the city from green- •- - - ri lighting these liquor stores. t "I'm on a private crusade," says Darryl ».~ Fisher, deputy to the city planning director, - o z who attends planning workshops on the ;: r K, subject around the state. "We have avery - LL serious problem and unless we deal with it,it o will deal with us." - s 9 A beer delivery arrives in Oakland,California—one of the many American cities taking steps to control the proliferation of alcohol outlets. Spring 1993 The Marin Institute 3 The city of Oakland hopes to handle the city for its first conditional use permit and overconcentration of existing alcohol outlets 1,000-foot rule for five problem streets that by improving on its southern neighbor's or- went into effect in 1978. It was members of dinance. Oakland's proposal guarantees neighborhood groups who monitored city enough enforcers to complete annual in- permit reviews and objected to requests from spections by levying a fee(yet to be decided) store owners for waivers. on liquor store owners- Today,Miley represents the poorest and "in the past Oakland has 29 percent ofits county's popu- richest of Oakland. �� lation, but 40 percent of the county's liquor stores "This issue represents exploitation of a year, says and 41 percent of the non-bar beer and urine sales. community, of a people, of a class,"he says. Dorothy Burse of Two-thirds of Oakland's 5762 liquor stores "It is a problem in our inner city,where low- Indianapolis, lack the use permit that gives neighborhoods a say income minorities live. You don't see this this alcohol over conditions that affect safety and health. sort of alcohol outlet density issue in rich Like the Los Angeles ordinance,this one areas. We're not abolitionists. We're trying ISSUe has would slap existing liquor stores with baseline to establish standards of behavior in a com- brought to- community standards and if in the maximum munity. People were allowed to sell drugs in gether whites time allowed for compliance the store owner the '80s. Now people want to be clear on . . . and blacks refused to make changes,the city could close what's acceptable. What they want to say is, . . . this has the business. 'You want to sell alcohol? You want to drink? given us new "We have a comprehensive plan whose Fine. This is how.'" motive is to prevent what happened in South Workingwith the mayor's office and Miley Strength. Central Los Angeles,"says City Councilman on the issue is county Supervisor Don Perata, Nate Miley. who held a public hearing two years ago on The plan complements last year's efforts, illegal drugs. when the council required new alcohol out- "It became real obvious that the bigger lets to be at least 1,000 feet away from schools, problem to the kids and families was alcohol churches, public gatherings, or other out- and that to create a healthy climate,we had to lets,and that any liquor store closed for more attack the alcohol and tobacco industry's than a year be banned from re-opening. marketing particularly to underage kids in If the ordinance passes,Miley says he will these neighborhoods,"says Perata. move for a moratorium on new alcohol out- The supervisor helped activists apply for lets. "I think it will take four years for the one of the 15 Robert Wood Johnson Founda- impact of the whole package to work," he tion "Fighting Back" community grants. says. "We're giving citizens the tools and will "We're going to educate these inner-city li- rely on them to monitor the stores. We feel quor store owners," says Perata. "We'll ex- we have a double whammy here—if a com- plain to them the laws, health issues. We'll plaint deals with the license,we can still take ask them to place signs in their windows them to the state Alcoholic Beverage Control, announcing that they don't sell to minors. and if it deals with a nuisance issue, we can The person selling the product isn't the en- look at their permit with the city." emy. We'll work with them. As for the Miley was a community organizer in alcohol industry, it thrives on being able to Oakland during the 1970s,when,as in South hold captive a select audience in the halls of Central, neighborhood activists pushed the the state capitol. It hates this local fight. We city to recognize crime and health problems plan to fight the alcohol industry on a thou- exacerbated by liquor store over-concentra- sand fronts." tion. It was Oakland Community Organiza- In California,this local strategy includes tions, for which Miley worked, that lobbied state legislators. Oakland Mayor Elihu Harris 4 The Marin Institute Spring 1993 is promoting state legislation to prohibit the vealed that Fair Park residents purchased only a state's alcohol enforcement board from re- fraction of the total amount of alcohol sold in their newing an existing license that fails to com- community. According to the report,in 1986 there ply with local zoning. were about 340 alcohol-related businesses inthe l3- For the second consecutive year,Assem- square-mile area, but only$4 million of$48 mil- blywoman Marguerite Archie-Hudson of Los lion in liquor sales were to Fair Park residents. Angeles [see Commentary, p. 121 has intro- In 1987, the city passed a permit re- duced a bill changing the state's formula for quirement that placed local conditions on distributing liquor licenses to prevent liquor already existing alcohol establishments in outlet clustering and requiring public hear- certain areas near schools, residences and ings for the transfer of licenses. The latter churches. The Dallas Merchants and Con- provision could stop the South Central liquor cessionaires Association filed suit against the store rebuilding,because current lawpermits city in June, 1990, to block enforcement. A financially troubled store owners wishing to state district judge upheld the ordinance that sell their liquor licenses to do so without year, but reversed her decision in March, public input on the impact of the store on a 1991, saying the ordinance conflicted with neighborhood. state regulation of the liquor industry. In December, 1991, a Texas appeals court up- held Dallas' right to control liquor stores through zoning. Merchants appealed and this past April 7, the Texas Supreme Court reversed that decision on a six to three vote. "We are going to file a motion for a rehearing,"says Dallas Assistant City Attorney Enough is Too Much John Rogers. "I'm sure we'll probably go to the legislature and try and get it fixed. Based More than a thousand miles away,Texas on what the decision said, we'll also most also looked to zoning as the key to control- likely try and pass an ordinance to control the ling alcohol outlet density in impoverished location of any business getting more than 75 areas in the '80s. While South Central orga- percent of its revenue from alcohol—bars nizers were documenting crime around their and liquor stores. This is about the liquor liquor stores, residents of South Dallas/Fair lobby versus the cities. It's not anew fightand Park were building a parallel movement that this isn't the last round." former Dallas Councilwoman Diane Ragsdale Mike Coker, Dallas planning director, calls"massive." For years,says Ragsdale,resi- says he has mailed about 60 packets of infor- dents complained about the lack of super- mation on the city's use permit process to markets,drugstores and banks,and the pro- communities around the country. liferation of liquor stores "on every street "In anticipation of the new ordinance, corner that attracted illegal drug activity and we plan to bring before the City Council other crime." immediately a moratorium on the issuance Ragsdale describes Fair Park as a 52-acre of certificates of occupancy for businesses "historic park with museums in the heart of that make 75 percent or more of their gross the black community." Fair Park is "wet"— revenue from the sale of alcoholic bever- designated by the legislature in the'60s to sell ages,"says Coker. "This should stop 22 liquor alcohol—and surrounded by"dry,"mostly stores that went out of business under our white communities. current ordinance from reopening." A study compelled by activist demands re- Spring 1993 The Marin Institute 5 vation and blacks for neighborhood preser- vation.This has given us new strength. With the mayor's help,I think we're finally headed in the right direction." "Inboards" Compared to South Central Los Ange- les,Oakland and Dallas,it appears that India- To advocates around the country, the napolis citizens fighting inner-city alcohol saturation in economically impoverished ar- outlets since 1985 have been asking for basics. eas of alcohol outlets that sell to minors and Twice — most recently this past March — encourage sidewalk drinking and crime state bills have failed that would have re- amounts to "corporate violence." In their quired county liquor boards to notify nearby world, alcohol and tobacco billboards are schools,residents and churches of new license known as "killboards" and "billboard vio- hearings. But city activists and the city's new lence." mayor,who was district attorney for 12 years, sound buoyant. "It's just a setback,"says Mayor Stephen Goldsmith. "I've undertaken two efforts— beginning a legal working group for poten- tial nuisance litigation against the highly of- fensive outlets,and looking into zoning and One of the oldest billboard battlefields is code enforcement. The liquor industry is an Chicago, where the mayor asked all 50 city effective lobbying group,but we aren't giving wards last February to declare a moratorium up." on new liquor licenses to combat crime. That Indeed,when Goldsmith took office last same week,569 ofan estimated 1,000 alcohol year, an industry representative sent him a and tobacco billboards lacking city permits warning via the local newspaper. "If Mr. came down and the City-Wide Coalition Goldsmith tries to become a legislator as Against Tobacco and Alcohol Billboards on mayor,then we'll come up with a law to limit the West Side and Standing Up,Taking Back the powers of local liquor boards," said on the South Side claimed a three-to-four- Darrell Felling, lobbyist for the Indiana Li- year victory. censed Beverage Association. The Citywide Coalition had its origin in In Indianapolis,the battle isprimarily against an anonymous, self-described "guerilla for one liquor store chain with 19 stores open 19 hours black health,Mandrake the Magician,"named a day, six days a week, in poor neighborhoods that after a Robin Hood-like cartoon character. suffer disproportionately from alcohol-related In August, 1989, Mandrake began calling problems. radio talk shows about the excessive propor- Dorothy Burse has been fighting against tion of alcohol and tobacco billboards in the the local outlets with the Citizens Neighbor- African-American community. Mandrake hood Coalition during the past five years. whitewashed his first billboards in Decem- "In the past year,"says Burse, "inner-city ber, 1989. On Good Friday, 1990, the coali- neighborhoods have finally come together to tion he inspired coordinated a rally attended protest and compile facts.on these most of- by hundreds of residents whose message— fensive outlets. Because part of the affected immediate removal of all alcohol and to- area is being gentrified,this alcohol issue has bacco billboards— was repeated in similar brought together whites for historic preser- events in African-Americancommunitiessuch 6 The Marin Institute Spring 1993 as Grand Rapids, Los Angeles, Baltimore, and if they cooperated,we'd help them get and Harlem. permits for the ones that were within pre- Subsequently, coalition members ob- scribed distances of residences. They did it twined a city zoning ordinance book and list voluntarily,ifyou call havingyourwrist pinned of all billboards and discovered that hun- to the matby a 500-pound gorilla'voluntary.'" dreds of them did not have permits and that Chicago has made"tens of thousands of the city had failed to enforce building and dollars" in billboard revenue since the suc- zoning laws. cessful activist pressure,according to Grady, ", , , local In 1990, coalition members — including and"the city has several draft ordinances up communities young people whose health and leadership develop- its sleeve for leverage if the billboard compa- will be taking mentwas thefocusoftheeffort—conducted astudy nies don't move on removing the rest of the that found that almost half of Chicago's billboards illegal boards." matters into Promoted alcohol and tobacco use. 77he Chicago their own Reporter wrote that there were nearly five times as hands . . . . many of these signs per person in the eight low- If the licensed income communities of color as there were in the five beverage predominantly white areas. While the coalition did its work, a white industry is priest with an African-American parish on Baltimore's billboard activists hit the streets in headed in the the South Side kept the issue in the newspa- 1989, when Boisclair Advertising bought 1,300 wrong pers with his billboard whitewashing,arrests, 'junior" boards that mostly touted alcohol and direction, we trial,and exoneration on charges of defacing tobacco and were reported to bring in an estimated have only private property. Father Michael Pfleger also annual revenue of.$1 million to .$2.5 million. convinced 120 liquor stores to stop selling These 5-by-11 foot ads were attached to the walls of ourselves to alcohol to minors and to remove beer and liquor or convenience stores selling alcohol blame. cigarette advertising from the outside of their As with Chicago,the tactic that brought —Beverage buildings. down more than 1,200 billboards last year in Bulletin Both the coalition and Pfleger worked Baltimore was exposing the lack of permits. with Graham Grady, city zoning administra- "You can tell where the billboards were tor. because of old paint around them,"says Sylvia "That first coalition demonstration was Fulwood of the Coalition for Beautiful enough pressure for the mayor's office to Neighborhoods. "It's made a big difference. assign me to the issue," recalls Grady. "Our Our group went block by block, counting legal department said from the beginning billboards and organizing churches and that we couldn't go after these billboards on schools. We even took note of illegal drugs the basis of their content,because the consti- hidden under these boards on the side of tution protects free speech. That they were liquor stores. The mayor named a task force erected illegally was a different story." on the issue. But I'd say things really started Grady says he met separately with bill- happening after[the anti-billboard advocacy board companies and activists. "What ulti- group] Scenic America did a report on mately happened was that I demanded an Baltimore's billboards. Eventually, the city inventory,discovered that four big companies took Boisclair to court and the court ordered didn't even know in many cases how many the illegal boards removed on a timetable." billboards they had and where. They feared There are more than 1,000 alcohol and a class action lawsuit by the city. We let them tobacco billboards left in Baltimore that have know that we'd seek a court order to get the permits,according to Zoning Administrator illegal ones—within 75 feet of residences— David Tanner. But the public health activists Spring 1.993 The Marin Institute 7 in that city have added another success to the positive messages or community service messages. billboard movement.Working with Tanner's "You might see a billboard that reads, office, the Citywide Liquor Coalition has re- `Kids Get the Message: Don't Drink, Don't lieved the landscape of a situation that threat- Smoke,' near a billboard advertising those ened to create a permanent sidewalk lounge products,"says Yvette Wunderlich,who chairs culture. the coalition. "We think a mixed message is Bev Thomas of the four-year-old coalition better than no positive message at all,which says that last year the group's undercover is the alternative we had. The community investigative efforts culminated in a state law campaign we had, the rallies and surveys, that will require about 200 of 600 taverns opened people's eyes. It made them think. operating in residential areas seven days a It's a good beginning." week,6 a.m.to 2 a.m.,either to act as bars or According to the city's arrangementwith to apply for new licenses as package goods billboard companies, no more than 50 per- stores. cent of the city's billboards can carry alcohol "Our group went inside these places and and tobacco advertising, 25 percent of the realized that there were barstools, but they remaining boards must go to the coalition were not really operating as bars,"says Tho- for positive messages, and new billboards mas. `"They were selling alcohol to go and it cannot promote alcohol and tobacco. went about as far as the sidewalk. They didn't As with Baltimore,activists successf ilwith have proper facilities or license to do that. It regulating outdoor advertising went on to was creating an open drug market.As a result look at retail outlet behavior, working for an of our work and lobbying, the local liquor ordinance prohibiting the sale of malt liquor board backed us. This is a true victory and after 9 p.m. that went into effect last Febru- gives us some control because as package ary. goods stores, they will have to do business All of this community activity has led the between 9 a.m. and midnight and close on alcohol industry to call for counterattack. Sundays." Referring to"the demise"of California's State Alcoholic Beverage Control Department,the Beverage Bulletin,an industry publication,ran an editorial in January, 1992, advising its subscribers to put internal competition aside and pull together in the face of a "looming peril." Unlike their counterparts in Chicago or "The handwriting is on the wall that local Baltimore,Milwaukeeactivists compromised communities will be taking matters into their with billboard companies. Early efforts in own hands when it comes to liquor licenses," 1987 to rid the city of the billboards failed. By it wrote. "...What we envision is a gradual 1991,the Milwaukee Coalition Against Drug curtailment of liquor outlets to the point and Alcohol Abuse decided that replacing where people in business will end up with an the offensive boards with positive messages exclusive territorial franchise...At the same against drinking and smoking was preferable time,other areas will dry up where no alcohol to living with the advertisements. licenses will be permitted...If the licensed Today, one-third of the 284 billboards in a beverage industry is headed in the wrong defined area of Milwaukee have been replaced with direction,we have only ourselves to blame." ■ 8 The Marin Institute Spring 1993