A discussion of the evidence that the pharmaceutical industry influences how physicians evaluate drugs in ways that encourage sales of their products and that are not always in the best interests of the patient.
The standard biomedical paradigmfor drug treatment ofdisease follows a straightforwardalgorithm. A disease is identified anddiagnostic criteria are developed.Promising medications are tested usingthe scientific method, especiallyrandomized controlled trials (RCTs),to determine safety and efficacy. Theresults of these studies are presentedto the scientific community in peerreviewedjournals and scientific debateoccurs.
If a drug is approved by the FDA,the results of the trials and the conflictingscientific opinions are presented tophysicians. Practicing psychiatrists mustevaluate the results for their own patientsand decide whether the benefits of amedication outweigh its shortcomings.
In this article, we discuss evidencethat the pharmaceutical industry influencesthis paradigm in ways thatencourage sales of their products andthat are not always in the best interestsof the patient.
An investigative article in the BritishMedical Journal by Moynihan andassociates1 describes how drug companiesdeveloped an explicit campaign topromote shyness as an important andtreatable disorder, successfully shapingmedical and public opinion aboutsocial phobia. A senior pharmaceuticalindustry executive involved in thepromotion of social phobia in Australiaacknowledged that marketing effortsexaggerated the prevalence of disease,and recognized the inherent conflict ofinterest when a drug company sponsorspublicity about a disease for whichit markets a treatment, as his companydid with social phobia.2
An article in the PharmaceuticalExecutive (a trade journal) discussesgenerally the making of a new disease,and how new markets can be createdby defining a new disease (such as themetabolic syndrome).3 Another way toexpand indications for a drug (and thusthe market) is to lower the threshold ofnormal, rendering more people illand in need of medication. Some criticsthink that pharmaceutical companieshave succeeded in medicalizingrisk factors or normal experiences, thusmaking otherwise healthy people thinkthey need certain drugs.4
Pharmaceutical companies havesuccessfully expanded the FDAapproveduses of various selective serotoninreuptake inhibitors (SSRIs) toinclude a variety of psychiatricdiseases, such as generalized anxietydisorder, obsessive-compulsive disorder,premenstrual dysphoric disorder,posttraumatic stress disorder, andbulimia nervosa.5 Trials for many offlabel uses such as premature ejaculation,migraine headache prophylaxis,and neurocardiogenic syncope areconducted and offlabel uses are oftensuggested.6 Drug company efforts toget patients on treatment have beenmotivated by a variety of reasons.Although there is no doubt that somepatients have benefited from thepromotion of treatment for thesesyndromes, drug company sales havecertainly increased.
In Australia, where direct-toconsumeradvertising (DTCA) is notpermitted, the value of campaignspromoting diseases can be clearlyseen. At the same time that Merck'shair growth drug finasteride wasapproved in Australia, a public relationsfirm hired by the manufacturerlaunched a vigorous campaign aboutthe dangers and tragedy of baldness.Eventually, newspapers published articlesabout the emotional trauma associatedwith hair loss, including expertopinions that losing hair could lead toemotional problems, harm to mentalwell-being, and loss of job prospects.These articles did not reveal that thestudies they cited were funded byMerck, and that the experts quoted hadbeen supplied by the company's publicrelations firm.1
For many practicing psychiatrists, theresults of peerreviewed, publishedRCTs are an essential element inmaking treatment decisions about drugtherapy. A systematic review of all clinicaltrials published in 4 leading, peerreviewedpsychiatry journals from2001 to 2003 found that pharmaceuticalsponsorship was common andwas statistically associated withoutcomes that favored the sponsor.7 Ofthe 162 RCTs published during thistime, 70% were funded by pharmaceuticalcompanies and 46% had atleast 1 author who had a financialconflict of interest either throughemployment or by having receiveddirect financial support from a pharmaceuticalcompany. Studies with atleast 1 author with a conflict of interestwere about 5 times more likely toreport positive results than were nonindustrysponsored studies.7
This article can only point to astrong correlation between industrysponsoredstudies and the reporting ofpositive results but cannot establishcausation. A summary of the types ofdesign and reporting modifications thathave been used in industrysponsoredpsychopharmacology trials is presentedin the Table.8
Clinical trials research on atypicalor secondgeneration antipsychotics(SGAs) provides an example of thisassociation between a trial's sponsorshipand its outcome.9 Of 59 RCTsstudying SGAs, those sponsored by thevendors of SGAs were significantlymore likely to find them superior to firstgenerationantipsychotics than werenonindustry-funded RCTs (P = .02).Furthermore, studies with first authorsemployed by the industry showed astrong trend (P = .05) toward favoring SGAs over conventional antipsychotics,compared with articles whose firstauthors were independent of industry.
Could the industry-sponsored trialssimply have been bigger, better, andmore able to identify a true superiorityof the sponsored drug? Two lines ofevidence argue against this possibility.First, the researchers who reported thisdifference in outcome found a trend (P= .07) toward higher overall RCT qualityscore in the studies that were notindustry funded. The second line ofevidence is indirect. A review by Hoganand associates10 of studies comparingone cholinesterase inhibitor withanother for Alzheimer disease (AD)concluded that, All studies were fundedby pharmaceutical companies, coauthoredby their employees, and reportedresults that favoured the sponsor's product.Similarly, industry-funded studiesthat compared 2 long-acting inhaledcorticosteroids and found a differencebetween them, found evidence in favorof the sponsor's product in 37 of 38 ofthe peer-reviewed published RCTs.11
|Table Summary of examples of design and reporting modifications in industry-sponsored trials|
|Technique used||Examples from the literature|
|Using doses outside the usual range||Haloperidol at a dose of 20 mg was the comparator in studies of EPSwith second-generation antipsychotics|
|Altering the dosing schedule||Twice daily amitriptyline was the comparator in evaluation of daytimesleepiness (vs paroxetine)|
|Using self-serving measurement scales||“Worst EPS score” rather than total or last observed EPS score was used to show risperidone causes the same amount of EPS as placebo|
|Selecting major findings and end points post hoc (secondary analyses or "data dredging")||Although not the initial hypothesis of the trial, side-effect differences at 1 week were used to compare venlafaxine and fluoxetine|
|Masking unfavorable side effects||Specific questions about sexual side effects of SSRIs elicit a 73% positiveresponse; change to unstructured open-ended questions drops the rate to 2%|
|Repeatedly publishing the same or similar positive studies||Multiple publication of the same study (stand-alone or as pooled data) is discussed in the text and in an editorial|
|Selectively highlighting findings||Ignoring placebo as being as efficacious as the drugs (and having fewer side effects) in the comparisons of 2 antidepressants for anxiety symptoms|
|Editorializing for the sponsor in abstracts and conclusions||The study drug heralded as the “new first-line treatment” when data do not support this conclusion|
|Publishing the obvious to emphasize a point for marketing||Publications of the risks of nortriptyline in older adults with heart disease in comparative trials versus various SSRIs|
|Touting nonsignificant but favorable differences and negating dropout differences statistically||A conclusion that paroxetine was better tolerated than clomipramine was based on insignificant differences|
|Selection of participants and trial duration||Short trial duration of 8 weeks was used to compare olanzapineor risperidone versus chlorpromazine in treatment-resistant patients; chlorpromazine typically shows effects over a longer time frame|
|Withholding unfavorable results||Withholding unfavorable results|
|Masking sponsorship||Ghostwriting is discussed in the text; disclosure of conflict of interestis now more commonly required|
Pharmaceutical sponsorship is widespreadin the SSRI research literature;96% of comparative trials involvingSSRIs published from 1980 to early2005 received some form of industrysupport.12 In 2003, Melander and associates13found that when data submittedto the Swedish drug regulatory authoritywere compared with the publishedliterature on SSRIs, several negativeindustry-sponsored trials were identifiedthat had never been published orhad appeared only within publicationsreporting pooled data.
In contrast, positive data were oftenpublished more than once, both as standalonepublications and in publicationsthat pooled data. Of the 42 studies submittedto the agency, only half showed theSSRI to be better than placebo and 19of these 21 were published in 22 standalonepublications. Of the 21 studies thatshowed nonsignificant differences, only6 were published in stand-alone publications.The International Committee ofMedical Journal Editors now requiresregistration of all clinical trials, whichwill make it more difficult to concealnegative data.14
In at least 1 context, the suppressionof negative studies on SSRIs has provedto be deliberate and explicit. An internaldocument released by Glaxo-SmithKline written in 1998 discussedhow to manage the results of 2 negativeclinical trials of paroxetine in the pediatricpopulation. The document statedthat the clinical trials results were insufficientlyrobust" to support an applicationfor a regulatory approval for use inpediatric depression and recommendedthat the company "effectively managethe dissemination of these data in orderto minimize any potential negativecommercial impact. It went on to state,It would be commercially unacceptableto include a statement that efficacy hadnot been demonstrated, as this wouldundermine the profile of paroxetine.15Only 1 of 3 studies of paroxetine in childrenwas eventually published.16
Selective publication may affect clinicians'impressions of safety as well asefficacy. Whittington and associates17reviewed published data on SSRIs in childrenand adolescents and compared themwith unpublished data obtained by theUK's Committee on Safety of Medicines(CSM). The CSM obtained the unpublisheddata during an investigation ofSSRI safety in children. The meta-analysisof the published data suggests thatSSRIs have a favorable risk-benefit ratio;however, the addition of the unpublisheddata would suggest an unfavorable riskbenefitprofile.17 A review of the unpublisheddata led Britain's Medicine andHealthcare Regulatory Authority to banthe pediatric use of all SSRIs other thanfluoxetine. In contrast, the research rhetoricof the sponsored trials tends tospin the meaning of the findings inpublished, positive studies.14 Onepublished trial, for example, concludedthat paroxetine is generally well toleratedand effective for major depressionin adolescents.18
The use of ghostwriting and publicrelations/marketing firms has becomecommonplace. During the course of alawsuit, it was disclosed that 85 internaldocuments about sertraline werecoordinated or prepared for Pfizer by amedical information company. Articledrafts had been prepared, then awaitedan author with good academic credentials,often to be determined, to beadded before submission to a highimpactjournal. Articles coordinated bythe medical information company were published in bigger name journals(with higher impact factors) and werelikely to be cited more often than articlesnot prepared by the company.19
Internal memos outline howSmithKline Beecham used a publicrelations firm to ghostwrite letters tothe editor to counterattack claims by arival drug maker about discontinuingparoxetine because of side effects.20 Thewillingness of academics to participatein this strategy is a sad chapter indeedin the history of science.
Industry-sponsored rhetoric in peerreviewedpapers often differs from theconclusions of independent authors.An industry-supported paper ondonepezil states that use of the drug isassociated with delayed nursing homeplacement in patients with AD and concludes with the statement: . . .doctors and caregivers need to beeducated that, in the same way as theactual benefits of treating hypertensionor hyperlipidemia are seen only afteryears of treatment, treatment of ADwith donepezil needs to be maintainedto see important long-term benefits.21In comparison, an independent, systematic,nonindustry-sponsored reviewconcludes: Because of flawed methodsand small clinical benefits, thescientific basis for recommendationsof cholinesterase inhibitors for the treatmentof AD is questionable.22
How the findings of clinical trials aretranslated into clinical practice is a keyfeature in the evolution of medical care.The pharmaceutical industry is heavilyinvolved in this aspect of drug use. Ananalysis of the exhibition hall at the 2002American Psychiatric Association (APA)Annual Meeting showed that 54% of allcompanies with booths were in violationof APA or FDA rules about marketingto physicians.23 One drug companybrought in more than 30 psychiatristsfrom Mexico to the APA meeting in 2002,all expenses paid, a practice that is againstthe PhRMA Code of Interactions withHealth Care Professionals24 and the AMACode of Ethics.25
The character and extent of promotionalactivities can be extravagant. In2004, the vendors of gabapentin agreedto plead guilty and pay more than $430million to resolve criminal chargesand civil liabilities in connection with their illegal and fraudulent promotionof the drug.26
DTCA is an acknowledged marketingtechnique. Advertisements for paroxetine,marketed for social anxietydisorder, and atomoxetine, marketed foradult attentiondeficit disorder, takecommon symptoms experienced at leastoccasionally by many healthy people (eg,fear of public speaking, being distractedand disorganized) and suggest that thesemay be symptoms of a "treatabledisease," which should be discussedwith physicians.26
The effectiveness of DTCA, from anindustry point of view, is suggested bya comparative study of patients in theUnited States with those in Canada(where DTCA of prescription drugs isillegal). More of the US patients requested drugs that were advertised ontelevision, and US patients were morelikely to leave the physician's office witha new prescription. About half the time,however, the prescribing physicians feltthat these new prescriptions were"unlikely" or only "possible" choicesfor patients with those complaints.28
Industry influence in clinical trials ishere to stay. The pharmaceutical industryfunds the majority of clinical trials,spending about $12 billion in 2002.29This spending has led to advances inscience and the discovery of new drugsthat have benefited patients enormously.The industry also spends large sums topromote their products, many of whichprovide little actual benefit at greatexpense to patients. By itself, industrysponsorship does not mean that a clinicaltrial is flawed; however, the corporations'fiduciary duty to make profitsfor their stockholders virtually ensuresthat promotional considerations willaffect the way diseases are identified,drugs are studied, studies are written,decisions are made about publication,and information from the sponsoredtrials is disseminated.
The increased transparency that willcome from the registration of clinicaltrials13 and the more stringent rulesconcerning disclosure of conflict of interestare important steps in trying to limitpromotional influences on putativelyscientific activities. Nonetheless, thepsychiatric literature will likely continueto be contaminated by rhetoric thatpromotes the use of drugs or touts the superiority of one drug over another onthe basis of questionable data. The challengefor physicians is obvious.
Dr Shah is a clinician-educator fellow in geriatricsat the Johns Hopkins School ofMedicine in Baltimore. He is board-certified ininternal medicine.
Dr Finucane is professor of medicine, chair ofthe ethics committee, and a clinician at theJohns Hopkins Bayview Medical Center inBaltimore.
The authors report no conflicts of interestconcerning the subject matter of this article.
1. Moynihan R, Heath I, Henry D. Selling sickness:the pharmaceutical industry and disease mongering.BMJ. 2002;324:886-891.
2. Moynihan R. Drug firms hype disease as salesploy, industry chief claims. BMJ. 2002;324:867.
3. Breitstein J. The making of a new disease.Available at: http://pharmexec.com/pharmexec/article/ articleDetail.jsp?id=80917. Accessed April4, 2006.
4. Payer L. Disease-Mongers: How Doctors, DrugCompanies, and Insurers Are Making You Feel Sick.New York: J. Wiley; 1992.
5. Ables AZ, Baughman OL 3rd. Antidepressants:update on new agents and indications. Am FamPhysician. 2003;67:547-554.
6. Stone KJ, Viera AJ, Parman CL. Off-label applicationsfor SSRIs. Am Fam Physician. 2003;68:498-504.
7. Perlis RH, Perlis CS, Wu Y, et al. Industry sponsorshipand financial conflict of interest in the reportingof clinical trials in psychiatry. Am J Psychiatry.2005;162:1957-1960.
8. Safer DJ. Design and reporting modifications inindustry-sponsored comparative psychopharmacologytrials. J Nerv Ment Dis. 2002;190:583-592.
9. Montgomery JH, Byerly M, Carmody T, et al. Ananalysis of the effect of funding source in randomizedclinical trials of second generation antipsychoticsfor the treatment of schizophrenia. Control ClinTrials. 2004;25:598-612.
10. Hogan DB, Goldlist B, Naglie G, Patterson C.Comparison studies of cholinesterase inhibitors forAlzheimer’s disease. Lancet Neurol. 2004;3:622-626.
11. Thomas PS, Tan KS, Yates DH. Sponsorship,authorship, and accountability. Lancet. 2002;359:351.
12. Hansen RA, Gartlehner G, Lohr KN, et al. Efficacyand safety of second-generation antidepressants inthe treatment of major depressive disorder. Ann InternMed. 2005;143:415-426.
13. Melander H, Ahlqvist-Rastad J, Meijer G, BurmannB. Evidence b(i)ased medicine-selective reportingfrom studies sponsored by pharmaceutical industry:review of studies in new drug applications. BMJ.2003;326:1171-1173.
14. De Angelis C, Drazen JM, Frizelle FA, et al. Clinicaltrial registration: a statement from the InternationalCommittee of Medical Journal Editors. N Engl J Med.2004;351:1250-1251.
15. Kondro W, Sibbald B. Drug company expertsadvised staff to withhold data about SSRI use in children.CMAJ. 2004;170:783.
16. Marshall E. Antidepressants and children: burieddata can be hazardous to a company’s health. Science.2004;304:1576-1577.
17. Whittington CJ, Kendall T, Fonagy P, et al.Selective serotonin reuptake inhibitors in childhooddepression: systematic review of published versusunpublished data. Lancet. 2004;363:1341-1345.
18. Keller MB, Ryan ND, Strober M, et al. Efficacyof paroxetine in the treatment of adolescent majordepression: a randomized, controlled trial. J Am AcadChild Adolesc Psychiatry. 2001;40:762-772.
19. Healy D, Cattell D. Interface between authorship,industry and science in the domain of therapeutics.Br J Psychiatry. 2003;183:22-27.
20. McHenry L. On the origin of great ideas: sciencein the age of big pharma. Hastings Cent Rep.2005;35:17-19.
21. Geldmacher DS, Provenzano G, McRae T, et al.Donepezil is associated with delayed nursing homeplacement in patients with Alzheimer’s disease. JAm Geriatr Soc. 2003;51:937-944.
22. Kaduszkiewicz H, Zimmermann T, Beck-BornholdtHP, van den Bussche H. Cholinesterase inhibitors forpatients with Alzheimer’s disease: systematic reviewof randomised clinical trials. BMJ. 2005;331:321-327.
23. Lurie P, Tran T, Wolfe SM, Goodman R. Violationsof exhibiting and FDA rules at an American PsychiatricAssociation annual meeting. J Public Health Policy.2005;26:389-399.
24. PhRMA Code of Interactions with Health CareProfessionals. Available at: http://www.phrma.org/files/PhRMA%20Code.pdf. Accessed March 31,2006.
25. AMA Code of Ethics. Available athttp://www.ama-assn.org/ama/pub/category/4001.html.Accessed March 31, 2006.
26. FDA Office of Public Affairs. Drug maker to pay430 million in fines, civil damages. FDA ConsumerMagazine. July-August 2004;38. Available at: http.//www.fda.gov/gdac/features/2004/404_wl.html.Accessed March 29, 2006.
27. Wolinsky H. Disease mongering and drug marketing:does the pharmaceutical industry manufacturediseases as well as drugs? EMBO Rep. 2005;6:612-614.
28. Mintzes B, Barer ML, Kravitz RL, et al. How doesdirect-to-consumer advertising (DTCA) affectprescribing? A survey in primary care environmentswith and without legal DTCA. CMAJ. 2003;169:405-412.
29. Brown K. The clinical trials business. Availableat: http://www.bccresearch.com/ editors/RB-171.html. Accessed January 16, 2006.
30. Marder SR, Meibach RC. Risperidone in the treatmentof schizophrenia. Am J Psychiatry. 1994;151:825-835.
31. Christiansen PE, Behnke K, Black CH, et al.Paroxetine and amitriptyline in the treatment ofdepression in general practice. Acta Psychiatr Scand.1996;93:158-163.
32. Costa e Silva J. Randomized, double-blindcomparison of venlafaxine and fluoxetine in outpatientswith major depression. J Clin Psychiatry.1998;59:352-357.
33. Modell JG, Katholi CR, Modell JD, DePalma RL.Comparative sexual side effects of bupropion, fluoxetine,paroxetine, and sertraline. Clin Pharmacol Ther.1997;61:476-487.
34. Zajecka J, Amsterdam JD, Quitkin FM, et al.Changes in adverse events reported by patientsduring 6 months of fluoxetine therapy. J ClinPsychiatry. 1999;60:389-394.
35. Rennie, D. Fair conduct and fair reporting of clinicaltrials. JAMA. 1999;282:1766-1768.
36. Trivedi MH, Rush AJ, Carmody TJ, et al. Do bupropionSR and sertraline differ in their effects on anxietyin depressed patients? J Clin Psychiatry.2001;62:776-781.
37. Bruggeman R, van der Linden C, Buitelaar JK,et al. Risperidone versus pimozide in Tourette’sdisorder: a comparative double-blind parallel-groupstudy. J Clin Psychiatry. 2001;62:50-56.
38. Nelson JC, Kennedy JS, Pollock BG, et al.Treatment of major depression with nortriptyline andparoxetine in patients with ischemic heart disease.Am J Psychiatry. 1999;156:1024-1028.
39. Roose SP. Modern cardiovascular standards forpsychotropic drugs. Psychopharmacol Bull.1992;28:35-43.
40. Lecrubier Y, Judge R for the CollaborativeParoxetine Panic Study Investigators. Long-term evaluationof paroxetine, clomipramine and placebo inpanic disorder. Acta Psychiatr Scand. 1997;95:153-160.
41. Conley RR, Tamminga CA, Bartko JJ, et al.Olanzapine compared with chlorpromazine in treatment-resistant schizophrenia. Am J Psychiatry.1998;155:914-920.