Tobacco Documents Online

The New England Abstracis in the advertisicg sections Journal of Medicine Established in 1812 as The NEW ENGLAND JOURNAL OF MEDICINE AND SURGERY VOLUME 327 OCTOBER 15, 1992 NUMBER 16 Original Articles Withdrawal Syndrome after the Double-Blind Cessation of Caffeine Consumption ..... 1109 K. SILVERMAN, S.M. EVANS, E.C. STRAIN, AND R.R. GRiFFITHS- Treatment of'Advanced Squamous-Cell Carci- noma of the Head and Neck with Alternat- ing Chemotherapy and Radiotherapy ... 1115 M. MERLANO AND OTHERS A Prospective Study of Recurrent Febrile Seizures ............................... 1122 A.T. BERG AND OTFlERs Deep-Vein Thrombosis and the Incidence of Subsequent Symptomatic Cancer ........ 1128 P. PRA.DO\I AND OTHERS Images in Clinical Medicine Coiled Larva of Trichireella spiralis .......... 1134 D. LACOMIS Special Article The Publication of Sponsored Symposiums in Medical Journals ...................... 1135 L.A. BP.RO, A. GALBRA[TH, AND D. RENNIE Review Article The Pathogenesis and Treatment of Kidney Stones ................................ 1141 F.L. CoE, J.H. PARKS, AND J.R. ASPLIN Case Records of the Massachusetts General Hospital A Three-Year-Old Boy with Obstructive Uropathy and a Heart Murmur of Increasing Intensity ................... 1153 M.D. FREED AND.M.F.. KING Editorials Clinical Importance of Caffeine Withdrawal 1160 J.R. Ht:cHEs The Best Medicine for Febrile Seizures...... 1161 J.M. FREEMAN Cancer and Clotting - Trousseau's Warning 1163 R.L. SILVERSTEIN AND R.L. NAC)nlAV Sounding Board Beyond Advance Directives - Health Care Surrogate Laws ........................ 1165 J.A. MENIKOFF, G.A. SACHS, AND '11. SIEGLER Correspondence Tuberculous Meningitis in Patients Infected with the Human Immunodeficiency Virus ....... 1170 Multidrug-Resistant Tuberculosis .............. 1172 Asthma in Children .......................... 1174 Genetic Testigg for Hypertrophic Cardiomyopathy 1175 Endoscopic BTliary Drainage for Acute Cholangitis 1176 The Unfinished Business of Physician-Payment Reform ................................ 1176 Correction to Statements about Lead in Infant Formula ............................... 1178 Book Reviews ........................... 1179 Books Received .......................... 1182 Notices .................................. 1183 Corrections Tuberculous Meningitis in Patients Infected with the Human Immunodeficiency Virus ....... 1170 Correction to Statements about Lcad in Infant Formula ............................... 1178 Corticosteroids in the Treatment of Optic Neuritis 1184 The Prevention of Heart Failure - A Ne -• Agenda 1184 Oxned, Published, and €'Copyrighted, 1992, by the Massachusetts Medical Society #E;lt1BYJ Gx rsz x.rz:s. 5-G I G I T 4;: 10 3 #0 2 00 133 i0# 12 309y 1t11 5 922 h!-? S ? Utl It1 t'1ICF.0F ILt1S Fnp1C:;; Q F'FF'IQLICHL'S ?QIJ }1 zF_F-P F_; htat{ AF:EOR t'1I 4310_~ -I'nE \ttY E.cu.D,lurR\rt, ut \Ituu.t\e ll',• . •-xi:kt•17!11) i.puh• li+hrd t,rrl,k frorn rdituriel ullicrs ,u 1l1 tilt.uiu... •nii•t, Ikntun, \l.\ O,tl I S-ti(M>t, tiuh,cription pricc' S'.I:LIKI lxr )t•.u ~•. nud-i la.. lx~.t.t r luid at I4.tun .ind at edditi in.d m.uliup i.fli j IY)ti l\t.\ti'1'CFZ ~rnd uJ~lrru chen~i~ tu 1',U• 1tur lS113, \\'.Jih ., \1.\ l!!2'I1-l4IU'.i.

Vol. 327 No. 16 PUBLICATION OF SPONSORED SYMPOSIUMS IN ~fEDICAL JOURti'ALS - BERO ET AL. 1135 SPECIAL ARTICLE THE PUBLICATION OF SPONSORED SYMPOSIUMS IN MEDICAL JOURNALS LISA A. BERO, PH.D., ALISO:Y GALBRALTH, B.A., AND DRUMMOND RENNIE, M.D. Abstract Background. An increasing proportion of sper.ding by the pharmaceutical industry has gone to funding symposiums that are published by peer-reviewed medical journafs. This study tests the hypothesis that such sponsorship, particularly by a single pharmaceu- tical company, is associated with a promotional orienta- tion of the symposium and a distortion of the peer-review process. Methods. We counted the symposiums published in 58 journals of c~itvcal medicine and surveyed the journal editors regard'ing their policies for symposium issues. We analyzed the symposium issues that appeared in the 11 journals that pttbished the most symposiums in order to determine tfle sponsor or sponsors, the topics, whether the tities were misieading, whether brand names were used, and whetiw the featured drugs were classified by the Food and Drug Administration as innovative or approved. Results. The number of symposiums published per AN increasing proportion of marketing expenses of the pharmaceutical industry has gone to funding symposiums, `+ith funding rising from $6 million in 1975 (adjusted for inflation to 1988 dollars) to $86 million in 1988-t Although the number of sponsored symposiums published in one cardiology journal has increased, the extent of symposium publication within the clinical literature as a whole has not been deter- mined? Given that advertising influences prescribing, we hypothesized that industry spending on sympo- siums had increased because the symposiums serve as a vehicle for marketing the industry's products. 3-6 Anecdotal reports suggest that there is a promotional orientation in some published syTrtposiums.',' This study analyzes the relation between the type of spon- sorship and promotional characteristics, such as the focus on a single drug rather than on a cross section of the latest advances in biomedicine, the featuring of noninnovative ("me-too") drugs that are difficult to sell in a highly competitive market, and the featuring of unapproved drugs that cannot be legally adver- tised. Commercial pressure from the pharmaceutical in- dustry could distort the peer-review process of publi- cation in medical journals. Sponsoring companies could bypass this process by selecting the topic of a symposium, selecting and paying the guest editor and speakers, selecting and editing the papers that are published, and paying the journal to obtain editorial From the imt'utue fuc }icahh Policy Studee. School of Wdicine (LA.S., A.G., D.R.). and the Divivon of Clinical Pturmacy, School of Phanrtacy (L.A. B. ), Universiry of Calefornia, San Ftaacixo: an1 the Jour»ul c J theMteri- can Medical Assaciatiow, S.xn Francisco (D.R.). Adhcss rrprint requests to Dr. Bctu at the Institutt for ciulth Policy Studia-Scfiool of :`Saiicinc, Univr,sity of Califomia, San Franciscro. 1 i48 Suacr St.. 11th R.- San Fruxisco- CA 94109. Supponc,l in part by d)e Asncrican Assacixioa of Rctimi ft:rxxu. Wast:ing- ton. D.C., and by a C:rce: ik.e)optnent A..ard (2)(T007:) from the California Tobucco-Relud Disetst Research Prcgram. year increased steadily from 1966 through 1989. Forty- two percent of those analyzed (262 of 625) had a single pharmaceutical company as the sponsor. These sympo- siums were more likely than those with other sponsors to have misleading titles (P<0.001) and to use brand names (P<0.001), and less likely to be peer-reviewed in the same manner as other articles in the parent journal (P<0.001). Of the 161 symposiums that focused on a single drug, 51 percent concerned unapproved therapies; 14 percent con- cemed drugs classified as bringing important therapeutic qains. Conclusions. Symposiums sponsored by drug com- panies often have promotional attributes and are not peer-reviewed. Financial relations among symposium par- ti-dpants, sponsors, and journals should be completely disclosed, symposiums should be clearly identified, and journal editors should maintain editorial control over con- tributions from symposiums. (N Engi J Med 1992;327: 1135-40.) control. We examine the published policies of editori- al review for symposiums and describe the results of a survey of journal editors and their policies with regard to such publications. METHODS A symposium was dcfincd as a collection of papers published as a separate issue or as a special section in a regular issue of a medical journal. Special editions, issu-s with single themes, and collections of abstracts from meetings were not classified as symposiums. Selection of Journals and Identification of Symposiums Fifty-eight journals were selected for the study on the basis of their citation rates and clinical readership. We identified the 100 medical journals that had the highest citation rates in 1987 accord- ing to SCI fournal Citation Reports, and we used the presence of a drug advertisement in the journal as a proxy for clinical readership. A computer search of the TiEDLINE data base for the words "symposia," "symposium," "proceeding," and "congress" in the journal title identified symposiums published by the 58 journals from January 1, 1966, through December 31, 1989. Data from the survey of journal editors were used to supplement the MEDLINE data. Eleven journals that had published 10 or more symposiums from 1980 through 1989 were selected for the content analysis: American ffeart,/ourrtal, Amarican journal of Cardiology, Amtrican fournal of .1ledicine, American journal of Obstetrics and Gytucolo{r, British,Journal of Anatsihesia, Cancer, Cirndation, fl)pattruinn, journal of Aller{v and Clinical Immunology, Ridfuy International, and T-antplantation F'rocnd- ings. At the medical library of the University of California, San Francisco, we found 625 symposiums, 314 of which (50 percent) were identified by the ASEDLINE search. (A complete list of the 625 symposiums analyzed is available through the National Auxil- - iary Publications Scr.•ice.') The rate of identification of sympo- siums through MEDLINE was poor, bccausc the key words used to identify symposiums were not systematically entrred by the codcrs at the National Library of Medicine bcfore 1991 (Kotzin S, Ntuion- al Library of Medicine: pcrsonal contmunic:uion). •See NAPS tlttcumcnt no. 04979 for 40 pages of supplementary matecial. Onkr from NAPS c/o Microfiche Publ;cations, P O. Box 3513, Grand Ccntral Station, New York. NY 10163-3513. Remit in advance (in U.S. funds only) $13.75 fur the first 20 pages plus 5.30 per additional page fur photocopies, or $4 for mictufiche, Oubide the U.S, and Canada add paaage uf $4.50 fur the titat 20 pages and S t for each 10 pages thercaRcr ($1.50 for micrufiche poetnge).

1136 THE NEW ENGLAivD JOURNAL OF bIEDICINE Oct. 15, 1992 Data were collected to examine the correlation between the finan- cial support the journals received from the drug industry and the publication of symposiums. The percentage of subscriptions that are paid for and the page ratio of advertisements to editorial copy were used as indicators of financial support 9 To test the hyp,)thesis that smaller journals were more likely to publish symposiums in order to increase revenues, we determined whether journals with low citation rates and small circulations published more sympo- siums than journals with higher citation rates and larger circula- tions. Data on the ratio of advertising space to copy, the percentage of paid subscriptions, circulation, and publisher were collected from Busintss Publication Rates and Data,10 Ulrirh's Intemational Periodicals Director); t the survey of editors, and the journals themselves. Characteristics of Symposiums For each symposium issue studied, data were collected on spon- sorship, misleading titles, the symposium topic, the use of brand names, and the editorial-review process. The symposium sponsor was determined to be one of the following: (1) a single pharmaceuti- cal company; (2) more than one such company; (3) one or more pharmaceutical companies plus a nonpharmaceutical organization; (4) another, nonpharmaceutical sponsor or sponsors, including uni- versities, medical societies, government, and foundations; and (5) no sponsor mentioned. The title of a symposium was considered misleading (1) if it oid not refer to drugs (instead mentioning "frontiers," "horizons," or "outcome measures," for example) but the topic of at least half the papers was drugs or (2) if the title referred to drugs generally (such as "new treatments" or "new therapies") but at least half the papers described treatment with a single drug. The authors reached a con- sensus about the ratings for the titles. A symposium topic was considered "nondrug" if at least half the artides did not describe studies involving pharmaceuticals (e.g., describing surgical interventions or quality-of-life measures). It was considered "single drug" if at least half the articles focused on a single drug (e.g., a clinical trial repeatedly comparing one drug to other drugs, or a pharmacokinetic study). The topic was considered "multiple drug" if at least half the articles described the results of studies involving more than one drug without featuring any particu- lar one (e.g., a comparison of a variety of drugs with a surgical intervention). A symposium was considered to use brand names if the brand name of a pharmaceutical product appeared at least once in any section other than the Methotis section. If the masthead of the symposium issue listed the same editorial staff members as the regular issues of the journal or if the symposium was part of a regular journal issue, the review process was classified as the "same" as that of the journal. By assuming that a symposium issue with a masthead was reviewed in the same way as other articles in the journal, we may have overestimated the number of peer- reviewed symposiums. The review was considered "special" if the process was described, but diffcred from that of the journal. When the topic of a symposium was a single drug, we determined the therapeutic class of the drug12't4 (with additional data from pharmaceutical companies and the symposium), the drug's date of approval by the Food and Drug Administration"-'s (with additional data from pharmaceutical companies), and the therapeutic rating assigned to the drug by the FDA. A therapeutic rating of A denoted important therapeutic gain, B modest therapeutic gain, and C little or no therapeutic gain.'2 Survey of Journal Editors A confidential questionnaire was mailed to the editors of the 58 journals selected for the study. A second mailing yielded a response rate of 81 percent (47 of 58). Thirty-eight respondents were editors of journals that had published symposiums between 1966 and 1989, and nine were editors of journals that had not. The response ratcs for journals that published symposiums (79 percent) and those that did not (90 percent) did nL.t differ significantly (P = 0.725). The questionnaire also asked the editors' opinions of sytnlwsiums, their editorial policies toward them, and the financial arrangements for their publication. Statistical Analysis The median values for continuous variables were compared by the Kruskal-Wallis test, because the distributions were not normal. exh charaLten.cic shown. Dichotomous variables and proportions were compared by chi- square analysis. RESULTS Extent of Symposium Publication The number of symposiums published per year by the 58 journals selected for the study increased from 71 for the period from 1966 to 1971 to 307 for the period from 1984 to 1989. Ten journals did not publish any symposiums during the study period, 21 journals stopped publishing symposiums, and 3 journals began publishing symposiums in 1990. (The Appendix con- tains a list of the 58 journals surveyed and the number of symposiums published by each.) Characteristics of Journals That Published Symposiums There were no significant diETerences with respect to circulation, citation rate, publisher, percentage of paid subscriptions, or ratio of advertisement to edito- rial copy between journals that published symposiums and those that did not. Two hundred eighty-three of the symposiums (45 percent) were published in three journals (American Heart fournal, American ,Journal of Cardiology, and American Journal of Medicine) that are owned by two large publishing companies. Reasons for Publishing Symposiums Twenty of the 38 survey respondents who published symposiums answered a question about the percent- age of the journal's net revenue that was derived from this type of publication. Six stated that more than 6 percent of their revenue was derived in this manner, eight gave a figure of 1 to 5 percent, and six stated that their journal derived no revenue from publishing sym- posiums. Seven of the 30 editors (23 percent) who answered a question about why they published sym- posiums stated that financial gain for the journal was the main reason. Two editors reported that theirJour- nals charged $400 to $1,500 per page to publish sym- posiums, and another reported charging a flat fee of $100,000. The journals charged an average of $15 per reprint, and reprint requests for symposiums Table 1. Characteristits of 625 Symposiums, According to Cate- gory of Sponsor.* CHARACTERLSTIc CATEGOAY OP SPONS°R DRUG C°N• wULnnE ?ANY nus aI NoN. SINGLE DRUO DRUG NUNDIIW DRl'C NONE COMPANY COMPANIES SfUNSUR SPONSORS IISiLD (N s `.6:) (.. - 29) (N - 4°) (N - $9) (N - 155) ALL prn'eN of grouP (numlkr oJsim(wsiums) no.<f .rvm~'siunLs Topic Single tin g 49 (129) 21 (6) 21(19) 1(1) 4(6) 161 Multiple 31 (80) 28 (8) 12(11) 9;8) 14 (21) 128 drugs Nonlnis 20 (53) 52 (I5) 67 (60) 9,100) 83 (128) 336 Misleading 19(50) 3(l) 12(11) 2 (2) 2(3) 67 title Use of brand 21 (56) 10(3) 21 (19) 2(2) 8(12) 92 narT,es •P<O.t1Ut by the chi-squue Ieu f.x Ihe comrui~m u(,p,mu)r cetegkKica v tth re~ir.'t m

Vol. 327 No. 16 PUBLICATION OF SPONSORED SYMPOSIUMS IN MEDICAL JOURNALS - BERO ET AL. 1137 ranged from 100 to 200,000 (average, 25,000). Addi- tional reasons given for publishing symposiums were to disseminate information (63 percent of respond- ents), to stimulate reader interest (40 percent), and to allow discussion of a topic in depth (10 percent). Content Analysis of Symposiums The descriptive characteristics of the 625 sympo- siums analyzed are shown in Table 1. Forty-two per- cent had a single pharmaceutical sponsor, 26 percent had a single drug as the topic, 15 percent used brand names, and 11 percent had misleading titles, such as "Anxiety: Quest for Improved Therapy" for a sympo- sium describing only a brand-name form of bu- spirone.ls (A complete list of the symposiums with misleading titles is available through the National Auxiliary Publications Service.*) Sponsorship by a single pharmaceutical company was associated with single-drug topics, misleading titles, and the use of brand names. The types of drugs most frequently featured in the 161 single-topic symposiums were cardiovascular medications (86 symposiums), antiinfective agents (18), nonsteroidal antiinflammatory drugs (12), and immunosuppressant agents (11). The date of FDA approval and the drug's therapeutic rating could not be determined for three and five symposiums, respec- tively. Eighty-one of 158 symposiums that focused on a single drug (51 percent) featured products that had not received FDA approval. Twenty-nine percent of the single-topic symposiums featured drugs that had not received FDA ratings. Of the rated drugs, 14 per- cent were assigned a rating of A (presenting an impor- tant therapeutic gain), 47 percent a rating of B (a moderate therapeutic gain), and 40 percent a rating of C (little or no therapeutic gaiii). The proportion of all FDA-approved drugs that received an A rating before 1990 was 13 percent, whereas 37 percent were f?-rated and 50 percent were C-rated. Fifty-nine percent of the C-rated drugs were featured in symposiur.s beiore they were approved, as compared with 33 percent of the B-rated drugs and 27 percent of the A-rated drugs (P = 0.014). Editorial Review of Articfes in Symposium Issues For 44 percent of the symposiums analyzed (275 of 625), the review process followed for the symposium articles was not specified. Forty-six percent (286 of 625) were subjected to the same type of peer review as the regular articles published in the journal, and 10 percent (64 of 625) received some type of "special" review. Published policies for special review varied among the symposiums, ranging from review by a publication committee, to review by the guest editor only, to review provided by the sponsoring company. Regardless of the type of review, 94 percent of the 'Sce NAPS docu¢rnt no. 04979 for 40 pages of supplementary material. Order from NAPS uo \Sierafiche Pub!icztioess. P.O. Box 3513. Grand Central Station. New Yofi;. 1'Y 10163-3513. Remit in advatxe (in U.S. funds only) t 13.75 for tfx first 20 pages plus 5.30 per a iditional page for photocop"tes, or S-'S for microfiche. Oazsidc the U.S. and C-u add postage of 54.50 for the fi:st 20 pages and SI for ex3 10 pages tfxrcafter (SI.50 for m'teroftche postage). editors (34 of 36) replied that they retained the right to reject a manuscript. Eight editors reported that their special review procedures were distorted by a variety of pressures from symposium organizers. Table 2 shows that the rigor of the review process varied with symposium sponsorship. The type of re- view also depended on the journal that published the symposium (P<0.001). Ninety percent or more of the symposiums published in the American Journal of Ob- stetrics and Gynecology, the British Journal of Anaesthe- sia, Cancer, Circulation, and Kidney International were subjected to peer review of the same type as regu- lar journal articles. In contrast, for more than 70 per- cent of the symposiums published in the American Heart Journal, the American Journal of Cardiology, and the American Journal of Medicine, no review policy was stated. Seventy-five percent of the symposiums that did not have a stated review policy were sponsored by single pharmaceutical companies (P<0.001). Table 3 shows that a focus on a single drug, the use of misleading titles, and the use of brand names were also associated with the lack of a review process. The Gold Standard for Symposium Publication We defined an ideal or "gold standard" for the pub- lication of a symposium to mean that the symposium did not have a misleading title, did not mention brand names, and was peer-reviewed in the same manner as the parent journal. There was an association between the type of sponsorship and whether or not a sympo- sium met this gold standard (Fig. 1). Thirteen percent of the symposiums sponsortia by single pharmaceuti- cal companies (34 of 262) met the standard, as com- pared with 74 percent of those with nonpharmaceuti- cal sponsors (66 of 89). DISCL`SSION Our data suggest that industry-sponsored syrt-tpo- siums are promotional in nature and that journals often abandon the peer-review process when they pub- lish symposiums. Published criteria for the acceptance of symposiums support our findings that the stand- ards for review vary among journals."'22 Financial pressures on journals appear to contribute to the in- creasing publication of symposiums. Table 2. Types of Editorial Review in 625 Published Symposiums, According to Category of Sponsor.* TYPE OF REYIEw CATEGORY OFSIONSOR ANGLE DRUG COMPANY (vs2t.:) MULTIPLE DRUO COMPANY DRUG ['LUSNONDRUG COMPANIES 1lONSOR (N- 29) (N- 90) ~1 NON- DRUG SPONSOLS (N- 69) NONE LISTED (N- 155) {xrcent of group (number o,r nmpoait:mal Not stated 78 (205) 41 (12) 42 (38) 12(11) 6(9) "Special" 4(11) 34(10) 23 (21) 1100) 8(12) Peer review as in parent journal IS (46) 24(7) 34 (31) 76(68) 86 (134) -P<0,001 by the chf-sqnare test foc ttx comparison of types of rcvkw. See the Meehods seetion for definitions of Lhe dittercnt types. Pen-enuges do not t.xal e:utly 100 becsuse of tounding.

1138 THE NEW ENGLAND JOUR`AL OF JiEDICINE Oct. 15, 1992 Table 3. Characteristics of Symposiums According to Type of Review.* cH1IACTERISTIC TYPE OF REVIEW NOT STATED "SPECIAL- (N = 275) (N = 64) PEER REYIEW AS 14 PAtUNi /Ol'LV AL (N = --4) ptrrent of group (no. of s) mposiumr) Topic Single drug 48(131) 16(10) 7 (20/ Multiple drugs 27(75) 8(5) 1706) Nondrug 25(69) 77(49) 76(218) Misleading title 22(60) 2 (1) 2(6) Use of brand 20(56) 14(9) 9(27) r1arnes 'P<0.001 by the chi-square test for the comparison of typa of rcvicw xith respa.t to sympositun topic uw) use of miskad'iay tisks. and P- 0.0013 for the comparison with respect to tbe use of braad names- Some perttatsges do tat total exxtty 100 because of m»ndng. Symposiums may be used to bring noninnovative and unapproved drugs to the att,-ntion of prescribing physicians. It has been argued that in so doing, they educate readers about important advances in medi- cine. However, the majority of symposiums sponsored by single pharmaceutical companies focused on single drug products, and the percentage of A-rated drugs featured in symposiums was no higher than the per- centage of all FDA-approved drugs with such a rat- ing. The featuring of unapproved and noninnovative drugs suggests that manufacturers of these drugs are seeking preapproval promotion to allow them to get a jump on the market, increase product recognition, or enhance investor interest.23 The FDA has warned some pharmaceutical companies about their use of symposiums to promote unapproved uses of their products.3•21 One possible consequence of the publication of symposiums that appear in the form of peer-reviewed scientific literature is that they can be used to in- ~ ~ y E ~ 8. a cn 100 90 so 70 60 so 40 30 20 10 0 Single Drug Company (N = 262) Muftiple Drug Com- Drug pany plus Corn- Nondrug panies Sponsor (N = 29) (N = 89) Nondrug No Sponsor Sponsor Usted (N = 90) (N = 755) Category of Sponsor Figure 1. Symposiums That Met (Solid Bars) and That Did Not Meet (Hatched Bars) the Gold Standard for Publication, According to Category of Sponsor. To meet our gold standard, a symposium was required not to have a misleading title, not to use brand names, and to be peer- reviewed in the same manner as articles in the parent joumal. The percentage of symposiums meeting this standard was calculated separately for each category oi sponsor, and the percentage dif- fered significantly between categories (P<0.001 by the chi-square test). fluence the prescribing practices of physicians. The inaccuracies of drug advertisements have been well publicized, and prescribers usually approach an ad- vertisement with skepticism .1e,21'21 Our findings sug- gest that symposiums should also be approached with skepticism. A lack of peer review such as we found with symposiums is associated with publica- tions of poorer quality.30 Symposiums may be given more credence than they deserve, because they often resemble regular journal issues and may be presented as educational material. Courses of continuing medi- cal education, which are sometimes published as sym- posiums, have been found to have content that is bi- ased in relation to the funding source.sl The acceptance of symposium publications could distort the medical literature and ultimately alter phy- sicians' prescribing practices and patient care. Physi-• cians might be persuaded by promotional symposiums to prescribe drugs that are inappropriate, expensive, or both. The increasing amount of money that the pharmaceutical industry devotes to symposiums could contribute to rising drug prices. We offer the following recommendations to reduce the promotional nature of symposium publications. Recommendations For the FDA We encourage the FDA to continue to develop and enforce guidelines for industry-sponsored sympo- siums.32,'3 The definitions of promotional and educa- tional material need to be clarified. The FDA's defini- tion of "independent" material should include an absence of financial ties between each participant and the sponsoring company, a broad topic for the sympo- sium, and an avoidance of discussions of ut)approved drugs. A list of all sponsors and the amount contribut- ed by each should be made public. For Pharmaceutical Companies Pharmaceutical sponsors should continue to sup- port and publish symposiums that do not have a promotional orientation. No suggestions of speak- ers, topics, or data selection should accompany the funding. Pharmaceutical companies should sponsor symposiums jointly because our data suggest that joint sponsorship reduces promotional characteristics. When pharmaceutical companies sponsor a sympo- sium, the sponsorship should be acknowledged in all brochures describing the symposium, in announce- ments at the symposium, and in all subsequent publi- cations. For Journats and Journal Editors If journal editors publish symposiums, they should insist that the papers be subjected to review of the same quality as articles in the regular issues of the journal. Journals should publish their review practices in each symposium issue and should maintain strict control over the selection of the topics and of the guest editor. Journal editors should reserve the right to rc- ject any paper. If the participants include speakers associated with the sponsoring company, the journal

Vol. 327 No. 16 PUBLICATION OF SPONSORED SYMPOSIUMS IN iti4EDICAL JOURNALS - BERO ET AL. edit• : should nominate an extra speaker, not associat- ed with the sponsor, to give an overview of the confer- ence. Journal editors should refuse to publish sympo- siums that are sponsored by a single pharmaceutical company or that have misleading titles, feature single products, or use brand names. If journals do publish industry-sponsored symposiums, each article should contain a complete statement disclosing all financial ties, past and present, among the authors, the guest editors, and the sponsoring company. Each paper included in a symposium issue should be clearly identified as such. Each page of a sympo- sium issue and each reprint should name the sym- posium sponsor, and the symposium title should be followed by a subtitle saying "symposium proceed- ing." In addition, symposium issues should be distin- guished from the regular journal issues with a different cover. If articles from symposiums are referenced in jour- nal articles, they should be listed separately or preced- ed by a prefix that clearly indicates their source. A similar style should be used in MEDLINE and other reference data bases. For Unfversfties We advise universities cosponsoring symposiums with drug companies to maintain complete control over the selection of the topic, the speakers, and any articles to be published. Universities should encour- age the funding of symposiums by multiple sponsors to avoid a promotional bias. For Speakers at Symposiutrns We suggest that before participating in a sympo- sium, potential speakers investigate the source of funding and the degree of editorial control, and that they avoid participating in symposiums with a pro- motional orientation. Speakers should also avoid sym- posiums that are solely sponsored by a company that funds their research.'I-hose who do choose to speak at industry-sponsored symposiums should not feel obli- gated to gratify their hosts by promoting their prod- ucts. Such speakers should disclose their connections with the sponsor during their talks. Professional organizations should discuss industry- sponsored symposium publications in their ethics guidelines. The American College of Physicians and the Pharmaceutical Manufacturers Association have recently issued guidelines for the acceptance of gifts and other marketing practices,' but the guidelines fail to address the problem of industry-sponsored publica- tions. The American IM edical Association has recently established a policy with regard to the publication of collections of articles that can serve as a model for symposium publications.' For Readers We urge readers to approach symposium issues that are sponsored by a single pharmaceutical company with skepticism and to evaluate each article for any promotional content. Readers should determine what peer-review process has been followed for an article or 1139 reprint before they consider the conclusions. Prescrib- ing physicians should ask representatives of the phar- maceutical company about the sponsorship and re- view the criteria used to prepare any reprints they distribute. In conclusion, pharmaceutical companies, journals, and the FDA should work together to eliminate pro- motional aspects of symposiums and ensure that the papers included are peer-reviewed. We arc indebted to Andrew Herxheimcr, Michael Cunningham, and Elliot Rapaport for their assistance with this project; to Ken- neth Feather, Robin Fox, Stanton Glantz, Jere Goyan, Philip Lee, Stuart Nightingale, William Parmley, Jane Smith, Richard Smith, and Ann Witt for their careful review of thc manuscript; and to the participants in the writing seminar of the Institute for Health Policy Studies. APPENDIX Clinical Medical Joumals Studied and Number of Symposiums Published by Each.* JOURNAL No. of $YNtOAVY.t 1966-1979 1980-1989 American Heart Journal 0 25 American Journal of Cardiology 32 79 American Journal of Medicine 18 67 American Journal of Obstetrics and Gynecology 0 10 American Journal of Ophthalmology 1 0 American Jcurnal of Psychiatry 15 0 American Journal of Surgery 6 4 American Review of Respiratory Disease 14 3 Anesthesiology ' 10 0 Annals of Internal Medicine 1 2 Annals of Neurology 0 4 Annals of Surgery 3 1 Antimicrobial Agents and Chemotherapy 2 0 Archives of Dermatology 43 3 Archives of Disease in Childhood 0 0 Archives of General Psychiatry 1 0 Archives of Internal Mrdicine 7 2 Archives of Neurology 0 2 Archives of Ophthalmology 0 0 Archives of Surgery I 1 0 Arthritis and Rheumatism 7 4 Blood 1 0 British Heart Journal 1 0 British Journal of Anaesthesia 5 11 British Jour-lal of Haematology 0 0 British Journal of Psychiatry 0 7 British Jcurnal of Surgery 7 0 British D:Pdical Journal 0 0 Cancer 6 13 Chest I 4 Circulation 3 11 Circulation Research 1 3 Diabetes 2 9 Diabetologia 1 2 Fertility and Sterility 2 0 Gastroentcrolo,Qy 3 0 Gut 6 4 Hypertensio•l 0 17 (Continued on next pegt) 'These 58 journals were sckcted foc study because they were ftrqucndy cited and coMamcJ at kact one dcug advertisement. Jouunals shown in boldface wcra setected for eontcnt analysis of their sy mposiunu beca,lu each joumal published 10 or motr symposiu ms from 1980 thtvu;h 1989. Data from the Mt;DUNE scarch and the survcy of journal editurs wcrz combincd to dctermine the numbcr of symposiums Iwblished e.ch yeu. If thete wcte discretwicies tVtw=n MEDUNE and the suncy dat3, with the survcy indicating that the journal did twt publish symposiums and MEDUNE irMl(cating tnat it did, and if t!x MEDUNE scfcrrnce ttxtfarrmi ro ourdcfinitiort of a sympmium, we countcd it1f SSF.DLINE did not list a sympesium and the survey did, wr venficd that the symt+osiumexistaf and eounNed it.7his a{+{xnach was necrsury because the MEDLJNE xanfi did rrot deta-t every syml+osium.

1140 THE NMV ENGLAND JOURNAL OF :1fEI)ICINE Oct. 15, 1992 APPENDIx, GONTINUED 10. Business publication rates and data. Part I. Vol. 71. Ho. 1. WAmcttc, Ill.: Standard Rate and Data Scrve:c, 1989. 11. Ulrich's intemational pcriodicals directory 1990-1991 29th cd Vol 1 Immunology 0 I 2 11 , . . . . New York: R.R. Bowkcr, 1990. Journal of Allergy aLd Clinical Immunology 12. Food and Drug Administration. Offices of Drug Evaluation statistical report. Washington, D.C.: Center for D.::1~ Evaluation and Research, Journal of the American Medical 0 0 1990. Association 13. Billups NF, Billups SM. Amcri :an r' ug index 1987. 31st cd. Philadelphia: Journal of Bone ar.d Joint Surgcry I 0 J.B. Lippincott, 1987. (American) 14. Physician's desk reference. 45th cd. Otadcll. N.J.: Medical Economics Journal of Infectious Diseases 4 0 Company, !991. Journat of Neurosurgcry I 0 15. Food and Drug Administration. Approved drug products with thcrapcutic i f P di i 0 5 equivalence. Ilth cd. Waihington, D.C.: Covemment Printing Officc, ,Journa o e atr cs 1991. Journal of Urology 2 0 16. Anxicty: quest for improved therapy. Am J Mcd 1987;82:Suppl 5A:1- Kidney International 5 23 3/. Izncet 0 0 17. Bennett JC. Drug company-sponsored symposia: pros and cons. Am J Mcd Metabolism 2 6 1987;83:812. New England Journal of Medicine 0 0 18. /dem. Plans for The American Journal ojMedicine for 1989. Am J Med Ncurolagy I 2 1989;86:I. Obstetrics and Gynecology 0 3 19. Finch KT. To our readers: publication guidelines for sy mposia. Am J Mcd 1989;86:A25-A26 Pediatrics 7 3 . Stroke 0 0 20. Freeman HL. Supplement policy of the journal. Br J Psychiatry 1989; 155:Suppi 6:6. Surgery, Gynecology and Obstetrics 0 0 21. Phillips 1. Supplements. J Antimicrob Chernother 1983;12:199-200. Thorax 0 0 22. Speller DCE. The relationship between the Joumal and the pharmaceutical Transplantation Pt•oceediagr. 6 56 industry. 1 Antimicrob Chcnwihcr 1987;19:711-2. Tratuplantltion 0 2 23. Kessler DA, Pines WL. The federal regulation of prescription drug advcrtis- All journals 240 400 ing and promotion. JAMA 1990;264:2409-15. 24. Cohn JN. Science and advertising. Circulation 1982;65:83940. 25. Hemminki E. Review of literature on the factors affoLting drug prescribing. Soc Sci Med 1975;9:111-6. R Fcra srarcEs 26. Meulds RFW, Bochner F, Wing LMH. Drug advertising. Med J Aust 1987; 146 660 1. United States Senate Commiuoe o. Labor and Human Resources. Advertis- 27. : . Dyer C. The penalties of issuing misleading advcrtiscments. BMJ 1987; ing, marketing and promoticeai practices of the pharmaceutical industry. 294:426-7. W,tzhiagton, D.C.: Gover»rxst Priati_-~-Offrce. 1991. (Senatc hearing 101- 23. Wickham W. Misleading advertising. Med J Aust 1987;147:160. 1d17.) 29. Wade VA, Mansfield PR, McDonald P1. Drug companies' evidence to 2. R.•SertsWC.Sotnetcaliticsofpresca-daycardioiogicjoumalpublishingin the USA. Cardiolosy l987;74:25462- 30. justify advertising. Lancet 1989;2:1261-i. Gs:rding the guardians: rrsea:ch on cditorial pecr review: selected prncccd- 3. Hemminki E. Review o( literatrre oa the fa<.ors affecting drug prescribing. Soc Sci Med 1975,9:111-6. ings frotn the First Intematioral Congress on Pecr Review in Biomedical Publication. JAMA t990 ^63:1311-441. 4. Mackowiak 11, Gagnat 1P. Fliecis d ptornotion on phamtaccutical de- 31. Bowman MA. The impact of drug company funding on the content of mand. Soc Sci Med 1985;20:1191-7. 5. Lezchia J. Pharmaceutical promotir» in Canada: convince them or confuse 32. continuing medical education. Mobi u 1986;6(1):66-9. Kessler DA. Drug promotion and scientific exchange: the role of thc clinical them. Int 1 Health Serv 1987;17:77-59. 6. Avorn J, Chen M, Hartky R. Sciatibe vasus commercial sources of influ- 33. investigator. N Engl J Med 1991325:201-3. Drug company supported activities ir, s;'entific or educational contcxts: ence on the '' be3sarior of p6ysician:xs. Am J Med 1982;73:4-8. 7. Fumcane T. Drug mmpzny-spoasorcd syaposia: pros and cons. Am J Med draft concept paper. Washington, D.C.: Food and Drug Administration, 1991. 1987;83:811-2. 34. PMA adopts advice against MD acceptance of drug industry gifts. FhC Rep 8. Musher DM, Frishmsrt W. Phar>mx~Rial company-sportwrcd symposia 1990;52(23):T+G3. and medical ethics. Am J Med 19E8,8.5:.596-7. 35. Standards for special collections of articles republished from JAMA. In: 9. Rennie D, Bero LA. Throw 't z,ray, Sun: the controlled circulation jour- nals. CBE Views I990;13:31-5. AMA scientificjoumals. Chicago: American Medical Association, October 15. 1991:1-9 (internal document).