Others

The link between a journal's source of revenue and the drug recommendations it publishes (2011)

Description
Becker A, Dörter F, Eckhardt K, Viniol A, Baum E, Kochen MM, Lexchin J, Wegscheider K, Donner-Banzhoff N. (2011). The association between a journal's source of revenue and the drug recommendations made in the articles it publishes. CMAJ. 183(5):544-8. Epub 2011 Feb 28.
Categories
Published
of 5
6
Categories
Published
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Transcript
  Research CMAJ 544 CMAJ, March 22, 2011, 183(5) © 2011 Canadian Medical Association or its licensors C linicians rely heavily on the informationprovided by journals for their continu-ing medical education. 1 In Germany,the circulation of educational medical journalsseems to exceed that of medical journals thatemphasize research.There has been some debate over whethercommercial interests influence the content of sci-entific publications. Evidence from a systematicreview suggests that authors’ conflicts of interestare significantly associated with positive resultsin the studies they publish. 2,3 There is additionalevidence that pharmaceutical companies pres-sure leading journals or journals with highimpact to avoid publishing content that is con-trary to the marketing interests of the company. 4 Drug advertisements are used by doctors asa source of information, and they have beenshown to influence the prescribing behaviour of physicians. 5 However, the quality of drugadvertisements placed in major journals is notan adequate basis on which to rely when decid-ing whether to prescribe a specific medica-tion. 6,7 In 2004, Reuter and Zitzewitz 8 describedan advertising-related bias in the financialmedia. The authors found that publicationswere more likely to print articles recommend-ing funds from companies that had placed themost advertisements in the publication in theprevious 12 months.Little is known about such corporate influ-ence on educational medical journals. Educa-tional medical journals are often sent to physi-cians free of charge. These so-called “free journals” depend entirely on advertising to gen-erate their revenue. Only a few educational med-ical journals are financed entirely by fees paid bytheir subscribers. We investigated whether a The association between a journal’s source of revenue andthe drug recommendations made in the articles it publishes Annette Becker MD MPH, Fatma Dörter DMD, Kirsten Eckhardt DMD, Annika Viniol MD, Erika Baum MD,Michael M. Kochen MD MPH, Joel Lexchin MD, Karl Wegscheider PhD, Norbert Donner-Banzhoff MD MSc Competing interests: Annette Becker was aconsultant for GrünenthalGmbH, from whom she hasalso received a speaker’sfee. Erika Baum is thevice-president of theGerman College of GeneralPractitioners and FamilyPhysicians. MichaelKochen is a member of the Drug Commission of theGerman MedicalAssociation and is editor of the journals  Zeitschrift für  Allgemeinmedizin and  Arznei-telegramm . JoelLexchin has worked as aconsultant for Apotex Inc.and the Government of Canada, has provided experttestimony for the lawfirmof McGinnis, Lochridgeand Kilgore LLP and is aboard member for HealthySkepticism. KarlWegscheider is a memberof the Drug Commission of the German MedicalAssociation. NorbertDonner-Banzhoff is amember of the board for theGerman College of GeneralPractitioners and FamilyPhysicians and has a spouseemployed by Novartis. Noother competing interestswere declared.This article has been peerreviewed. Correspondence to: Dr. Annette Becker,annette.becker@staff .uni-marburg.de CMAJ  2011. DOI:10.1503 /cmaj.100951 Background: There is evidence to suggestthat pharmaceutical companies influence thepublication and content of research papers. M ost German physicians rely on journals fortheir continuing medical education. We stud-ied the influence of pharmaceutical advertis-ing on the drug recommendations made inarticles published in 11 German journals thatfocus on continuing medical education. Methods: We conducted a cross-sectional studyof all of the issues of 11 journals published in2007. Only journals frequently read by generalpractitioners were chosen. Issues were screenedfor pharmaceutical advertisements and recom-mendations made in the editorial content for aspecified selection of drugs. Each journal wasrated on a five-point scale according to thestrength with which it either recommended ordiscouraged the use of these drugs. We lookedfor differences in these ratings between free journals (i.e., those financed entirely by phar-maceutical advertising), journals with mixedsources of revenue and journals financed solelyby subscription fees. The journals were alsoscreened for the simultaneous appearance ofadvertisements and recommendations for thesame drug within a certain period, which wasadjusted for both journal and class of drug. Results : We identified 313 issues containing atleast one advertisement for the selected drugsand 412 articles in which drug recommenda-tions were made. Free journals were more likelyto recommend the specified drugs than journalswith sources of revenue that were mixed orbased solely on subscriptions. The simultaneousappearance of advertisements and recommen-dations for the same drug in the same issue of a journal showed an inconsistent association. Interpretation: Free journals almost exclusivelyrecommended the use of the specified drugs,whereas journals financed entirely with sub-scription fees tended to recommend againstthe use of the same drugs. Doctors should beaware of this bias in their use of material pub-lished in medical journals that focus on con-tinuing medical education.  Abstract See related commentary by Kesselheim, page 534  Research CMAJ, March 22, 2011, 183(5) 545  journal’s source of revenue was associated withrecommendations for or against the use of spe-cific drugs in the editorial content of the journal. Methods Using advice from four general practitioners,each of whom runs a practice and is involvedwith the training of medical students, we ident-ified 11 German educational medical journalsthat are widely read by general practitioners. The journals were divided into three classes accordingto their source of revenue: “free journals,” whichare completely financed by paid advertisements;“mixed-revenue journals,” which are financed byboth subscription fees and paid advertisements;and “subscription journals,” which are com-pletely financed by subscription fees.All of the issues published in 2007 in theselected periodicals were included in the searchfor eligible articles and advertisements (Table 1).We selected nine innovative drugs or classes of drug that are used to treat common illnesses andthat were heavily promoted at the start of the sam-pling period (Table 2). These drugs were chosenfor two reasons: they were still protected by patentsand were more expensive than other drugs used totreat the same conditions, and there was some con-troversy over either their effectiveness or the rangeof indications for which they should be prescribed.All advertisements and articles related to thesedrugs and published in the selected journals in 2007were recorded. Articles that referred to other drugsor articles that did not make any recommendationsfor the specified drugs were excluded.Articles were electronically scanned and pre-sented to raters in a neutral layout. Each articlereferring to one of the selected drugs was evalu-ated for the strength of its recommendation for oragainst the use of the drug. The evaluations weredone using an ordinal five-item scale ranging from–2 to +2. A positive integer indicated a strong ormoderate recommendation for use of the drug, ascore of zero indicated a neutral statement, and anegative integer indicated a moderate or strongrecommendation against use of the drug. The eval-uation was conducted by two dental students whohad been trained to use the rating system and whowere blind to the selected journals (interrater relia-bility, Cohen’s κ  = 0.73). Ratings followed a stan-dardized procedure and were based on the contentand titlesof the articles, as well as on the way inwhich the authors worded their recommendationsfor or against the use of the selected drugs.We compared the strength of recommenda-tions across classes of journal revenue sources Table 1: Characteristics of 11 journals commonly read by general practitioners in Germany that were included in the study*Journal title CirculationNo. of issues pervolume (frequency)Frequency ofadvertisements† Available by subscription Free journals  Ärztezeitung 63 900 226 (daily) ++ Usually free Medical Tribune 67 000 52 (weekly) ++ Usually free Der Hausarzt  55 000 20 (biweekly) ++ Usually free Der Allgemeinarzt  50 000 20 (biweekly) ++ Usually free Munchener MedizinischeWochenschrift — Fortschritte der Medizin 60 000 51 (weekly) ++ Usually free Mixed-revenue journals Deutsches Ärzteblatt  403 000 52 (weekly) + Yes, and free to members ofthe German MedicalAssociation Subscription journals  Arzneiverordnung in der Praxis 3 500 4 (quarterly) 0 Yes  Zeitschrift für Allgemeinmedizin 3 500 12 (monthly) 0 Yes, and free to members ofthe German College ofGeneral Practitioners andFamily Physicians  Arznei Telegramm Not available 12 (monthly) 0 Yes Der Arzneimittelbrief  14 000 12 (monthly) 0 Yes Tägliche Praxis Not available 4 (quarterly) 0 Yes *Journals were classified based on their sources of revenue. Free journals relied entirely on paid advertisements, mixed-revenue journals were financed by amixture of subscription fees and paid advertisements, and subscription journals relied solely on subscription fees from their readers.†++ = many paid advertisements, + = classified advertisements only, 0 = no advertisements.  Research 546 CMAJ, March 22, 2011, 183(5) using a linear mixed model with revenue class asthe factor and journals within revenue classes asthe random effect to account for any potentialcluster effect caused by multiple recommenda-tions appearing within the same journal.We did a logistic regression analysis to deter-mine whether a recommendation for a drug wasmore likely to appear if an advertisement for thatdrug was present in the same issue of the journal.This analysis referred only to issues of the free journals. The presence of a recommendation for aspecific drug was the dependent variable. Inde-pendent regressors included the drug (clopidrogelwas used as the reference), the journal and thepresence of an advertisement for the drug. Theinfluence of an advertisement varied according todrug and journal because we included the corre-sponding interaction terms. We used backwardselection so that only significant interaction termswere retained in the model (log-likelihood ratiotest,  p < 0.05). Three analyses were done: one forads in the same issue; one for ads in the previous,same or subsequent issue; and one for ads in upto five previous issues (depending on the numberof issues of the journal published per year). Results We screened 465 issues from 11 periodicals foreach of the 9 drugs or drug classes. The character-istics of each journal are given in Table 1. Weidentified 638 advertisementsfor the selecteddrugs (312 issues, ranging from 14 to 161 per jour-nal) and 297 articles that made recommendationsfor or against the use of the selected drugs (250issues, ranging from 2 to 105 per journal) (Table2). More than one advertisement of the same drugin the same issue was counted only once. We iden-tified 592 advertisements and 256 articles makingrecommendations concerning the selected drugs infree journals, 46 advertisements and 7 articles inthe mixed-revenue journal and 34 articles with rec- Table 2: The selected drug and drug classes and the frequency with which they were advertised orrecommended in articles published in the selected educational journalsDrug or class of drug (indication)No. of issues in whichadvertisements for theselected drug appears n = 638 No. of issues in which at leastone article recommended use ofthe selected drug n = 412 AChE inhibitors (dementia) 36 21Ezitimibe (hypercholesterolaemia) 75 18Incretin mimetics/enhancers (diabetes) 102 79Clopidogrel (prophylactic forcardiovascular disease)38 38Glitazones (diabetes) 56 74Varenicline (smoking cessation) 62 29Duloxetine (depression, urgeincontinence) 3 6ARBs and sartans (hypertension, heartfailure) 174 105Insulin analogues (diabetes) 92 42 Note: AChE = acetylcholinesterase, ARBs = angiotensin receptor blockers. Table 3: Strength of recommendations for or against use of the selected drugs, by journal’s source of revenueStrength of recommendation; no. (%) of articlesSource of revenue  n –2 –1 0 1 2 Mean score*Paid advertisements† 256 0 1 (0.4) 2 (0.8) 18 (7.0) 235 (91.8) 1.9Paid advertisements andsubscription fees7 1 (14.3) 1 (14.3) 0 1 (14.3) 4 (57.1) 0.86Subscription fees 34 10 (29.4) 18 (52.9) 3 (8.8) 3 (8.8) 0 –1.03 Note: 2 = strong recommendation for use, 1 = moderate recommendation for use, 0 = neutral recommendation, –1 = moderate recommendation against use, –2 =strong recommendation against use.*  p < 0.001, F  test.†Free journals.  Research CMAJ, March 22, 2011, 183(5) 547 ommendations but no advertisements in the subscription-based journals. Table 2 shows the fre-quency with which advertisements and articlesconcerning each of the selected drugs appeared.The results of our main analysis, the compari-son of the strength of recommendations to supportthe use of a drug across journal classes, are givenin Table 3. Readers of the free journals receivedalmost exclusively positive messages (ratings of +1 and +2) about the selected drugs, and thesedrugs were mentioned more often in the free jour-nals than in the journals that depended on sub-scription fees or a mixture of subscription fees andpaid advertisements for their revenue. Meanstrengths of recommendations were 1.90 for free journals, 0.86 for mixed-revenue journals and –1.03 for subscription-based journals (Table 3).Logistic regression modelling showed that,among the free journals, the frequency of recom-mendations depended on both the journal and thedrug (Table 4). Glitazones were three times morelikely (odds ratio [OR] 2.61, 95% confidenceinterval [CI] 1.59–4.27) to be recommended thanclopidogrel, the reference drug, as were incretinmimetics (OR 3.15, 95% CI 1.93–5.14), andangiotensin-receptor blockers were about fivetimes more likely to be recommended than thereference drug (OR 5.18, 95% CI 3.20–8.40). Intwo of the free journals studied,  Munchener Medi- zinische Wochenschrift  and  Der Allgemeinarzt  ,the presence of an advertisement for a specificdrug more than doubled the probability of a posi-tive recommendation for that drug appearing inthe same issue of the journal. The observed asso-ciations disappeared when looking at the pervi-ous, same or subsequent issues or at the previousfive issues. Interpretation Our study showed that recommendations for theuse of a drug depended on the source of the jour-nal’s funding. Free journals almost exclusivelyendorsed the use of the selected drugs, whereas journals that relied exclusively on subscriptionfees for their revenue were more likely to recom-mend against the use of the same drugs.A survey of Canadian general practitionersfound that they were unlikely to consider jour-nals that did not undergo a peer review process(typically journals sent to them free of charge) ascredible sources of information. However, morethan half of the doctors surveyed said that theyhad used free journals as a source of informationduring the previous month. 9 In 2006, Lexchin and Light 10 stated that rev-enue from journal advertising was one of the mostimportant factors associated with commercialbias. 11,12 Using print material as the sole source of information for continuing medical educationseems ineffective, but the integration of suchmaterial into multimedia interventions has beenshown to change the behaviour of physicans inpractice. 13,14 Print media are still among the mostimportant sources of information for doctors. 1 Rennie and Bero summarized the characteris-tics of free journals as follows: they publishmore advertisements than editorial content, theyare not owned by societies, they do not publishsrcinal work, the articles they publish are notcited, they are subjected to no standard level of peer review, and they are deficient in critical edi-torials and correspondence. 15 It is also worth not-ing that the conflicts of interests of either theauthors or the editors are not declared in these journals.  Munchener Medizinische Wochen- Table 4: Association between recommendations for the use of selecteddrugs and their advertisement in the same issue of free journals, adjustedfor drug class and journalVariable Adjusted OR (95% CI) Drug or class of drug AChE inhibitors 0.73 (0.39–1.33)Ezitimibe 0.64 (0.34–1.20)Duloxetine 0.18 (0.07–0.47)Glitazones 2.61 (1.59–4.27)Incretin mimetics/enhancers 3.15 (1.93–5.14)Insulin analogues 1.64 (0.97–2.78)Varenicicline 0.91 (0.51–1.63)ARBs and sartans 5.18 (3.20–8.40)Clopidogrel* 1.00 Journal  Ärztezeitung 0.36 (0.23–0.56) Medical Tribune 0.93 (0.56–1.56) Munchener MedizinischeWochenschrift — Fortschritte der Medizin 0.67 (0.39–1.14) Der Hausarzt  1.76 (0.96–3.22) Der Allgemeinarzt  2.43 (1.46–4.05) Presence of specificadvertisement (any drug) by journal (interaction terms)  Ärztezeitung 0.62 (0.38–1.03) Medical Tribune 1.62 (0.95–2.76) Munchener MedizinischeWochenschrift — Fortschritte der Medizin 2.19 (1.25–3.82) Der Hausarzt  0.67 (0.30–1.47) Der Allgemeinarzt  2.77 (1.04–7.35) Note: AChE = acetylcholinesterase, ARBs = angiotensin receptor blockers, CI = confidenceinterval, OR = odds ratio.*Reference drug.
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x