Financial ties between dsm-iv panel members and the pharmaceutical industry

Regular Article
Financial Ties between DSM-IV Panel
Members and the Pharmaceutical
Industry
Lisa Cosgrove a Sheldon Krimsky b Manisha Vijayaraghavan a a University of Massachusetts, Boston, Mass. , and b Tufts University, Medford, Mass. , USA Key Words
cies (22%) and speakers bureau (16%). Conclusions: Our
Confl icts of interest ؒ Ethics ؒ Financial interests ؒ inquiry into the relationships between DSM panel mem- bers and the pharmaceutical industry demonstrates that there are strong fi nancial ties between the industry and those who are responsible for developing and modifying Abstract
the diagnostic criteria for mental illness. The connec- Background: Increasing attention has been given to the
tions are especially strong in those diagnostic areas transparency of potential confl icts of interest in clinical where drugs are the fi rst line of treatment for mental dis- medicine and biomedical sciences, particularly in journal orders. Full disclosure by DSM panel members of their publishing and science advisory panels. The authors ex- fi nancial relationships with for-profi t entities that manu- amined the degree and type of fi nancial ties to the phar- facture drugs used in the treatment of mental illness is maceutical industry of panel members responsible for revisions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) . Methods: By using multimodal
screening techniques the authors investigated the fi nan-
cial ties to the pharmaceutical industry of 170 panel
Medical journals began introducing confl ict-of-inter- members who contributed to the diagnostic criteria pro- est (COI) disclosure requirements for authors two de- duced for the DSM-IV and the DSM-IV-TR. Results: Of the
cades ago beginning with articles of original research. A 170 DSM panel members 95 (56%) had one or more fi - growing awareness of the importance of author disclosure nancial associations with companies in the pharmaceuti- in biomedical publications is refl ected in the rising num- cal industry. One hundred percent of the members of the ber of medical journals that have adopted COI policies panels on ‘Mood Disorders’ and ‘Schizophrenia and Oth- over the past decade and the support for such policies er Psychotic Disorders’ had fi nancial ties to drug compa- among professional societies. If fi nancial COI among nies. The leading categories of fi nancial interest held by medical researchers can bias the outcome of a study (as panel members were research funding (42%), consultan- recent research shows) [1, 2] , there is as much reason to Department of Urban and Environmental Policy and Planning Tel. +1 617 627 3394, Fax +1 617 627 3377, E-Mail [email protected] believe it can also bias the recommendations made by States, most of whom have taken instruction on the DSM members of advisory panels [3–5] . The importance of [17] . protecting the integrity and public trust in scientifi c and In this study we investigated the fi nancial relationships medical advisory committees has been widely discussed that members of the advisory boards to the DSM-IV and [6, 7] . Yet, there remain areas that lack the transparency the DSM-IV-TR have had with the pharmaceutical in- of fi nancial COIs that have become standard procedures dustry, which manufactures drugs used by clinicians for the treatment of mental disorders. We assessed the extent To date there has been no study examining the rela- and types of fi nancial relationships for each of the diag- tionship between the pharmaceutical industry and the nostic panels. scientists comprising the advisory panels that recom-mended changes in the Diagnostic and Statistical Manu-al of Mental Disorders ( DSM ) , a leading medical manual used for the diagnosis of psychiatric disorders. Pharma- The Diagnostic and Statistical Manual for Mental Disorders ceutical companies provide substantial funding for con- ( DSM ) is organized around working groups or panels. Most of the ventions, journals, and research related to what is includ- panels address a specifi c category such as ‘Mood Disorders’. The ed in the DSM, because what is considered diagnosable members of each panel have signifi cant infl uence in determining directly impacts the sale of their drugs [10] . This ‘uneasy whether a new diagnosis should be added or an older diagnosis re-vised for the next edition of the manual. From the latest edition of alliance’ [11] was evidenced when a prominent journal the DSM (DSM-IV) and the edition with text revision (DSM-IV- reported that it was diffi cult to fi nd research psychiatrists TR), we identifi ed 228 individuals associated with the development to write an editorial about the treatment of depression of the volume. After deleting clerical staff, we were left with 170 who did not have fi nancial ties to the pharmaceutical expert members who comprised a total of 18 distinct panels. companies that manufacture antidepressant medications Each panel member was put through a series of ‘screens’ to de- termine whether he or she has had any fi nancial associations with [12] . Recently some members of the American Psychiat- one or more pharmaceutical companies whose business is poten- ric Association (APA) have expressed concern about the tially affected by decisions or recommendations made by the pan- potential for COI that arises with the increase in industry el. The ‘screens’ involved tracking publications of the panel mem- support [13] . For example, at the annual meeting of the ber for disclosures of potential COIs (in journals that have disclo- APA in 2004 there were 54 industry-supported symposia. sure policies), and submitting the panel member’s name into the news media database Lexis-Nexis and Internet search engines. We Also, pharmaceutical advertising revenue in APA jour- also submitted panel members’ names into databases of the US nals, totaling USD 7.5 million in 2003, increased 22% in Patent and Trademark Offi ce to screen for holdings of intellectual property in a drug whose sales could be affected by recommenda-tions of the DSM. For example, the FDA’s approval of Sarafem (fl uoxetine hydrochloride) for the treatment of premenstrual dys-phoric disorder was contingent upon expert testimony that con- Objectives of the Study
cluded that premenstrual dysphoric disorder was a distinct clinical entity that should be included as a mental disorder in the DSM Continuously produced since 1952 by the APA, the [18] . DSM, currently in its sixth revision [15] , is the offi cial Panel members were screened for any fi nancial affi liations they manual for psychiatric diagnosis in the United States.1 had with the drug industry between the years 1989 (the DSM-IV was published in 1994) through 2004. By using multiple screening Its classifi cation system is used by government agencies techniques to gather published or Internet data on fi nancial affi lia- and for all mental health professionals who seek third tions, we were able to avoid a methodology that relied solely on party reimbursements. The manual provides the stan- self-reporting (e.g., surveying panel members). dard psychiatric taxonomy found in psychiatry and psy- Financial associations of interest for this study include: hono- chology textbooks [16] . Nearly 400,000 mental health raria, equity holdings in a drug company; principal in a startup company, member of a scientifi c advisory board or speakers bureau professionals, including psychiatric nurses, social work- of a drug company; expert witness for a company in litigation; pat- ers, psychologists and psychiatrists practice in the United ent or copyright holder; consultancy; gifts from drug companies including travel, grants, contracts, and research materials. We use the term ‘fi nancial interest’ in describing the relationship between panel members and the pharmaceutical industry rather than the 1 Outside of the United States the offi cial coding system for mental illness term ‘confl ict of interest’ (COI) because the latter term implies an is the International Statistical Classifi cation of Diseases and Related Health interpretation of the interest. Thus, we choose not to defi ne COI. Problems, 10th Revision (ICD-10). According to the DSM IV-TR, the prepara-tion of DSM-IV has been closely coordinated with the preparation of Chapter Rather, we identify categories of fi nancial interest and reserve judg- V ‘Mental and Behavioural Disorders’ of ICD-10.
ment on whether they represent a real, perceived, or potential COI. Our approach is congruent with other publications that make a distinction between the fi nding of ‘fi nancial interests’ and the judg- We chose 20% of the panel members (every fi fth name) for an ment of ‘confl icts of interest’ [19–21] . Specifi cally, there is less dis- interrater reliability test. One investigator, who was not involved agreement about what constitutes a ‘fi nancial interest’ than there in coding, reviewed the data for 19 names representing 44 coding is about what makes a ‘confl ict of interest’. decisions. This investigator missed 2 coding decisions; all other The specifi c screening methods we applied included Lexis-Nex- coding decisions matched the results of the coding team. Our test is, Internet search engines (such as Google and Yahoo), the US Pat- demonstrated that the most likely error of using another coder was ent and Trademark Offi ce Internet Site on patents pending or missing a fi nancial interest. This error was minimized by having awarded, and Medline. Each fi nding of a panel member’s fi nancial three independent coders who compared fi ndings and reached con- connection to a drug company was coded: H = honorarium; RF = sensus. Our methodology tends to err in understating rather than research funding; RM = research materials (equipment, drugs, cell overstating the fi nancial interests of panel members. cultures, etc); EQ = equity in a company; CB = member of a cor-porate board (advisory board or board of directors); SB = speakers bureau; CON = consultant; ET = expert testimony; DINP = drug industry nonprofi t (a nonprofi t organization funded primarily by the pharmaceutical industry such as the Novartis Foundation);P = holds a patent, patent application or royalties on a product rel-evant to the treatment of mental disorders, and CIFS = collaborator Of the 170 DSM panel members 95 (56%) had one or in industry-funded study. Individuals classifi ed under CIFS are more of the eleven fi nancial links to a company in the DSM panel members who participated in a study as a coinvestiga- pharmaceutical industry. Figure 1 shows the percentages tor or collaborator, where the principal investigator (rather than the of the panel members listed in the DSM-IV and DSM- panel member) was described as funded by the pharmaceutical company. IV-TR with fi nancial linkages to drug companies. Unless A DSM panel member who received a fee from a drug company otherwise noted, the percentages given for each DSM cat- for speaking at a session of a symposium sponsored by a pharma- egory include the members from both the 1994 and 2000 ceutical company was coded as an H (received an honorarium). editions. In 6 out of 18 panels more than 80% of panel However, when a university or medical school receives unrestrict- members were found to have fi nancial ties to the phar- ed educational grants from a drug company, the company usually does not have a role in selecting the speakers, setting the honoraria, maceutical industry. These include 100% of the panels or signing the speaker’s honoraria check. For our study, a panel for the ‘Mood Disorders Work Group’ (n = 8) and the member who spoke at a symposium but did not receive direct pay- ‘Schizophrenia and Other Psychotic Disorders Work ment from a drug company (e.g., he/she spoke at a professional Group’ (n = 7), 81% for ‘Anxiety Disorders’ (n = 16), 83% meeting or Grand Rounds of a medical school sponsored wholly or for ‘Eating Disorders’ (n = 6), 88% for ‘Medication-In- in part by a pharmaceutical company) was not coded as H. In oth-er words, lecture honoraria paid by a university were considered as duced Movement Disorders’ (n = 8) and 83% for ‘Pre- nondisclosable fi nancial interests for the purpose of this study. Al- menstrual Dysphoric Disorders’ (n = 6) (see fi g. 1 ). The though not included in our analysis, we found 23 panel members mental illness categories denoted by ‘Mood Disorders’ who gave talks sponsored, in part, by drug companies but who were and ‘Schizophrenia and Other Psychotic Disorders’ are compensated by universities under unrestricted grants. the two main categories for which psychopharmacologi- While our methodology allowed us to ascertain varied fi nancial relationships that existed between DSM panel members and the cal treatment is standard practice, whereas it is far less pharmaceutical industry during the period of analysis, it did not likely for individuals diagnosed with ‘Substance-Related allow us to make causal inferences about those relationships. We Disorders’ (17%; n = 6) to receive such treatment (unless could not determine whether, or to what extent, an individual’s as- there is a coexisting mental disorder such as a mood dis- sociation with the pharmaceutical industry infl uenced his/her be- havior on a DSM panel or, conversely, whether participation on a DSM panel infl uenced his/her subsequent involvement with the The most frequent fi nancial link we found between the pharmaceutical industry. For the most part, the data on panel mem- DSM expert panels and the drug industry is ‘research ber associations with the pharmaceutical industry are atemporal. funding’. Among the 170 panel members, 42% received Some of the fi nancial relationships might have occurred before, research funding from pharmaceutical companies; 22% during, or after publication of the DSM volumes. Ethical consider- were consultants and 16% served as members of a drug ations are relevant whether the panel member’s involvement in the drug industry occurred prior to or after DSM publication. company’s speakers bureau (see fi g. 2 ). Three investigators independently conducted screens on the Of those panel members who had fi nancial links with panel members. Any questions about coding were resolved by a the pharmaceutical industry (n = 95) 76% had research fourth investigator. No panel member was coded as having a fi nan- funding, 40% had consulting income, 29% served on a cial connection unless there was unambiguous information con- speakers bureau, and 25% received honoraria other than from serving on a speakers bureau (see fi g. 3 ). More than half of the panel members with fi nancial ties were found Fig. 1. Percentage of panel members of the DSM with fi nancial ties.
Delirium, Dementia, Amnestic and Other Cognitive Schizophrenia and Other Psychotic Disorders Disorders Usually First Diagnosed during Infancy, Sexual and Gender Identity Disorders (TR) Medication-Induced Movement Disorders (TR) Committee on Psychiatric Diagnosis and Assessment Joint Committee of the Board of Trustees and Assembly of District Branches on Issues to DSM-IV to have more than one category of fi nancial interests with Limitations
the pharmaceutical industry. Eleven panel members had fi ve different fi nancial ties (see fi g. 4 ). When there were The results of this study need to be interpreted in light competing coding categories, only the most precise de- of several limitations. First, it is reasonable to expect that scriptor was coded. For example, a panel member listed some types of fi nancial relationships were not detected by on a speakers bureau, where he/she received honoraria, our screening methods. For example, expert witnesses was coded only as ‘SB’. Hence, the percentages in fi gure 4 serving in litigation are diffi cult to detect with standard represent the most conservative estimate of DSM panel screening tools. Second, our screening methods fell short members with multiple fi nancial associations to the phar- of allowing us to quantify or to set a temporal sequence for the association. In most instances information about the amount of money received from pharmaceutical com- Fig. 2. Categories of fi nancial interest held
by DSM Panel Members: total members (n = 170). RF = Research funding; CON = member of a corporate board; H = hono-rarium; CIFS = collaborator in industry- funded study; DINP = drug industry non-profi t affi liation; RM = research materials; Fig. 3. Categories of fi nancial interest held
nancial interests (n = 95). RF = Research funding; CON = consultant; SB = speakers bureau; CB = member of a corporate board; H = honorarium; CIFS = collaborator in in- dustry-funded study; DINP = drug industry nonprofi t affi liation; RM = research materi- als; EQ = equity in a company; P = patent; ET = expert testimony. Panel members with multiple financial interest Fig. 4. DSM panel members with multiple
fi nancial interests.
panies was not disclosed. Also, disclosures were reported ceutical industry, because of the greater likelihood that strictly in terms of whether a person was a current or past the drug industry may be exerting an undue infl uence on recipient of industry support. Even among journals that the DSM. As previously noted, the DSM working groups have set high standards on author disclosure, no specifi c that had the highest percentage of fi nancial ties to the dates are given. For example, the following is a typical pharmaceutical industry were those groups working in disclosure statement: ‘Dr. X has received honoraria and diagnostic areas (e.g., mood disorders and psychotic dis- research support from Company Y and Company Z and orders) where pharmacological interventions are stan- he also serves on their speakers bureau.’ Because it was dard treatment. In light of the extreme profi tability of the not possible to quantify these relationships or locate the psychotropic drug market, the connections found in this precise time they occurred, only qualitative conclusions study between the DSM and the pharmaceutical industry can be drawn about the extent of the fi nancial relation- are cause for concern. For example, antidepressants and ships of DSM panel members with the pharmaceutical antipsychotics were the fourth and fi fth leading therapy industry. Also, no conclusions can be drawn about the classes of drugs in 2004, with annual global sales totaling impact these relationships may have had on a panel mem- USD 20.3 and USD 14.1 billion dollars, respectively [24] . ber’s behavior. However, there are ethical concerns in One antidepressant alone, venlafaxine (Effexor, Wyeth) these relationships regardless of the amount of money giv- achieved USD 3.7 billion in sales in 2004 [25] . The anti- en or the timing of the fi nancial association. A fi nancial psychotic drug market has been identifi ed as one of main relationship held during (or just before) participation as therapeutic areas for global market growth with sales of an expert panel member might infl uence or be perceived USD 8.5 billion in 2002 and projected sales of USD 18.2 to infl uence the outcome of a DSM work group. On the billion by 2007 [26] . Therefore, we recommend that the other hand, if their association with the pharmaceutical APA institute a disclosure policy for panel members of industry occurred after work on the panel was completed, the DSM who have fi nancial ties to the drug industry. panel members might be perceived as using their prestige This is consistent with the trend for greater transparency to leverage lucrative consulting arrangements with the in the membership of federal advisory panels. Raising pharmaceutical industry. In the latter case public trust in awareness about the real or perceived COI of panel mem- the independence of medical science is eroded when for- bers is an important public health issue [27–32] . mer panel members, who are valued as ‘thought leaders’ Transparency should also apply to journal publication [22] , exert their infl uence on prescription practices through [33, 34] . We undertook a search of Ulrich’s Periodicals consulting, public speaking, and participation in industry- Directory using the terms ‘psychiatry and drugs’, ‘psycho- pharmacology’, ‘drugs and mental illness’ and ‘psychiatry Finally, we relied on self-reported disclosure data that and medication’. When restricted to the descriptors ‘ac- was available in the open literature. Author noncompli- tive’, ‘academic/scholarly’, ‘English language’ and ‘refer- ance with journal COI policies has been cited as a prob- eed journals’, the search identifi ed 45 journals of psychi- lem [23] . Therefore, the percentages of DSM panel mem- atry. Of those, 19 had COI disclosure policies (42%). bers identifi ed as having fi nancial ties and the variety of Many of the fi nancial ties that were found in this study their ties to the pharmaceutical industry should be con- were discovered because of these disclosure policies. Thus, we also recommend that all psychiatry journals adopt COI policies following the recommendations of the International Committee of Medical Journal Editors. Comments
Receiving fi nancial support from a pharmaceutical Acknowledgments
company should not automatically disqualify an individ-ual from serving on a DSM panel. However, the public We are most grateful for the assistance of two UMass graduate students, Michelle Levinson and Greg Noga and one Tufts Univer- and mental health professionals have a right to know sity undergraduate student Naomi Mower, all of whom conducted about these fi nancial ties, because pharmaceutical com- preliminary searches. Ms. Mower participated through a summer panies have a vested interest in what mental disorders are scholars grant from Tufts University provided through the univer- included in the DSM. Transparency is especially impor- sity’s endowment. Dr. Krimsky was her mentor under the program. tant when there are multiple and continuous fi nancial He received a small stipend for supplies. No other participants of the study were supported by grants. relationships between panel members and the pharma- References
1 Bekelman JE, Li Y, Gross CP: Scope and im- 13 Goldstein B: APA-pharmaceutical relations. 21 Boyd EA, Bero LA: Assessing faculty fi nancial pact of fi nancial confl icts of interest in biomed- relationships with industry: a case study. ical research: a systematic review. JAMA 2003; 14 Scully JH: Advertising revenue helps APA meet its objectives. Psychiatr News 2004; 22 Angell M: The Truth about Drug Companies. 2 Antonuccio DO, Danton WG, DeNelsky GY, Greenberg RP, Gordon JS: Raising questions 15 American Psychiatric Association: Diagnostic 23 Krimsky S, Rothenberg LS: Confl ict of interest about antidepressants. Psychother Psychosom and Statistical Manual of Mental Disorders: policies in science and medical journals: edito- DSM-IV, ed 4, text revision. Washington, rial practices and author disclosures. Sci Eng 3 Little M: Research ethics and confl icts of inter- est. J Med Ethics 1999;25:259–262. 16 Hare-Mustin RT, Marecek J: Abnormal psy- 24 Leading Therapy Classes by Global Pharma- 4 Gutfi eld R: Panel urges FDA to act on adviser chology and clinical psychology: the politics of ceutical Sales, 2004. Available at www.ims- bias. Wall Street Journal, Dec 9, 1992, B6. madness; in Fox D, Prilleltensky I (eds): Criti- h e a l t h . c o m / i m s / p o r t a l / f r o n t / a r t i c l e C / 5 Monbiot G: Guard dogs of perception: corpo- cal Psychology: an Introduction. Sage, Thou- rate takeover of science. Sci Eng Ethics 2003; 17 Mental health practitioners and trainees: Men- 25 Leading Products by Global Pharmaceutical 6 Krimsky S: The redemption of federal advi- tal Health, United States, 2002. Rockville MD: Sales, 2004. Available at www.imshealth.com/ sory committees; in Science in the Private In- USDHHS Substance Abuse and Mental Health terest . Lanham, Rowman & Littlefi eld, 2003. Services Administration, Center for Mental 71234024_71234109,00.html (accessed March 7 Schroter S, Morris J, Chaudhry S, et al: Does Health Services. Available at http://www.men- the type of competing interest statement affect talhealth.samhsa.gov/publications/allpubs/ 26 Global Market Forecast. Available at www.
readers’ perceptions of the credibility of re- SMA04-3938/Chapter21.asp (accessed Febru- 18 Cosgrove L, Riddle B: Constructions of femi- 27 Editorial: Experts and the drug industry. New 8 Stolberg SJ: Study says clinical guides often ninity and experiences of menstrual distress. hide ties of doctors. New York Times, Feb 6, 28 Cho MK, Shohara R, Schissel A, Rennie D: 19 Krimsky S, Rothenberg LS, Stott P, Kyle G: Policies on faculty confl ict of interest at US 9 Fava GA: Confl ict of interest in psychophar- Financial interests of authors in scientifi c jour- universities. JAMA 2000;284:2203–2208. macology: can Dr. Jekyll still control Mr. nals: a pilot study of 14 publications. Sci Eng 29 Rennie D: Fair conduct and fair reporting of Hyde? Psychother Psychosom 2004;73:1–4. clinical trials. JAMA 1999;282:1766–1768. 10 Eriksen K, Kress VE: Beyond the DSM Story: 20 Warner TD, Gluck JP: What do we really know 30 Korn D: Confl icts of interest in biomedical re- Ethical Quandaries, Challenges, and Best Prac- about confl icts of interest in biomedical re- search? Psychopharmacology 2003;171:36– 31 Starcevic V: Opportunistic ‘rediscovery’ of 11 Bodenheimer T: Uneasy alliance: clinical in- mental disorders by the pharmaceutical indus- vestigators and the pharmaceutical industry. N try. Psychother Psychosom 2002;71:305–310. 32 Fava GA: Confl ict of interest and special inter- 12 Angell M: Is academic medicine for sale? N est groups. The making of a counter culture.

Source: http://www.webduke.ch/literatur/Cosgrove2006_DSM-IVPanelMembersandPharmaceutical.pdf