Tobacco Documents Online

Commentary Conflict of Interest The New McCarthyism in Science M.a\1" -zc'ienzists and juut'nal eclitol•., eslmuse the 1>1•inciltle that a piece of scientific work or cl'lticislll is supposed to be jutl,CtI solely on its sc•ic•ntitic mel•i-. In rcc•ent }•eats• however, inuny biome<lic•al jotu•nills. Ja.Ua amon;- thenl,' huve at',opt- ed new polic•ies that t•eyuire <ilithot's to disclose potential cnnflic•ts of interest. These disclosures are intended to pl'o- vicle information that is supposed to aid edicors and readers in th-ir evaluations of the «•ork: in the words of'the former editor of The \'c•rr F,iigtcrud .lotu•licil cij .11c'clic'ilie, such dis- c•1o<ure "alerts re<uier., to information th,, v may Nri~h to have as the}• :lssess the published 1•epol•t."= Examples of scientific mi: condttct rans~ing from : loppine: s to fi<tud have been teide- lv publicized: disclosure polic~e, are supposed to reduce these Imoblenl., by allowing 1•eader. to make ~t more informed and therefore a better intelZn•etation of published work. Y,etter interpretation of published work is a lofty goal that evel;\•one would endorse. On the othel• hallcl, infol•mecl judg- nlent is not bettet• judpnent, as anyone who trnclel•- : tan<Is the reasons for blind assigllment and assessment in a clinical trial can attest. Unfortunatel' v, and perhaps surpris- ingly, these policies of mandatol;t• disclosure thtcal•t the pl•in- ciple that a wol•k should be judged solely on its merits. By emphaSizing ct•edentials, the: e policies foster an ad hominenl approach to evaluating science. Although intended to rait'-e scientific discoul•se to a highel• standard of ethics, these new disclosure policies themselves contlibute new ethical I>rob- lems, b}• abridging the right of honest scientists to an im- ,taltial heal•in* of their u•ol•k. Ultimately, as I explain below, these policies Ncill 1•echu•e, l-ather than improve, the overall objectivity of scientific discouhe. They have already led to it ltolic}• of censorship in one journal. What Constitutes a Conflict of Interest? The phrase coufl irt ofinterc .t hazz a pejorative connotation, but it actually denotes a nearly univelN~al circumstance. Con- flict of interest refers to any situation in which an individual with responsibility to others (which includes professional t•e- sIx,nsibilities) might be influenced, consciotlslti• or unconsciotls- IN•• by financial or personal factors that involve self-interest. A conflict of interest b,y it<elfdoes r.-ot indicate wrongdoing- it merel' v refers to asettint~ in which factol•s exist that might influence one's conduct. The conventional view is that where the potential fol• such influence exists, there is 1•eason for concern, and moreovel•% this potential implies an increased probabilit}• of misconduct. For a<c'ientist who is labeled as -, e D- o! E: o' c C a:, ,.0 DePa~: it - .,- : u -_ - =- ._ ,r:, ,( t, . y. .-d it c tncnt '~ -.~w . . ; . :.a 4~ ... _ , __ '-L9a_~ Cr R~:hm~n ~s thC Loc:er 2782 _-. . ,i. - 2 •. _ Zc3 N'~ 21 UMI Article Clearinghouse has reproduced this material with permission of the copyright owner. Further reproduction is prohibited. ® hnvinv a Crrnflict of' illtele. st, ihe concern is an lllcl'etl-ed possibility for a wot•lc to he biased. The label of "conflict of interest" thttS as a type of' "risk fhctor" of whether a 'cientist hil.: hultecl his or hel• work because of the influence of tlle conflicting factor. 1Iuch of the cliscus~ion about factors that cause <t conflict of interest has focused on financial influences, but the topic is by no means linliterl to money. According to Sc•ieuc•e; :' which has recently adopted a new conflict-of-interest policy,' the range of factors that nlight cl•eate a conflict of intel•e: t could inclurle the sexual orientation of the re~earchel• (: ttch as a homosexual ; tud}•ing homosesuality^') or a tenaciously held belief in specific scientific tieol•ies:' Along the: e: anle lines, u•e might expect conflict-of-interest concerns to be raised, for example, about Seventll-Day Adventists n•ho are studying the health effects of the comparatively ab.temiou,; life-style of theit• fellon• Adventists. Whereas policie.s at .1:1.11.-1 and The \'ex• E»glccncl .lomwrrl of .11<•clic'iue emphasize financial conflicts, Scieurt, asks authors to clirillre "any relations-hip: that they believe could be comtl•uecl as causinE~<t conflict of interest, whether or not the individual believes that is actu- ull}• so:" In othel• n•nrcls, to c:oml>1}• with disclosure policies at Science, attthol•~ miLht need to disclose to ellitol's theil• 1•e- lig,ion lulcl sexual orientation along lCitll tl1C•Il• financial Itol•t- folio. Conflict of Interest May Be Temptation, but Is Temptation Sin? If h<tvinl,'' a conflict of illterest k viewed a~ it risk filctol• for producing biased «•orlc, one must aLree th(tt clisclosure of conflicts would he of 'omv hellt to readel•:; in iclentifvin.v antthols who are "at t~l•eatvr risk" fitl• introducing bias into their R•ol•k. L1nfo1'tllilatel~, ~~hilc (lisclo>tlrc~ nla% • label some- one ils having a conflict of interest, it does not reveal whethel• there acttl-llly is ~l pl•ublem with the work or whether tlle implicit prediction is a'•fal<e positive." Not everyone sub- jected to the temptation of placing his or her self-interest above responsibility to others succtnnb• a that temht~ttion. What is worse, tlle scientist who is labeled a~ having a conflict of interest hit~ no possibility of pleading innocent to havin,q produced biased work, because the bias m,ly be sub- conscious, ancl no one call claim to be inuntme to subconscious influences. Indeed, Sc•ic-uc•e's editor, Daniel E. Kosliland, Jr, states explicitly that an author's view of how much influence the potentially conflictin,g f<lctor had is il•1•elevant.' Because of the hos~ibility of a subconscious influence, alleging that someone has a conflict of intet•ezt has an insidiot.sanlbigrltitN•. If an investigator deliberately alters research findings, the work issimlll}• frutcltllent. In contrx~t, ~uheonsrio,t~ tilting of -stucl}• design, data collection, anal}•sis, or interpretation be- cal:~:,, of one's finarcial ;:uppol•t or othcl.aoti;rs .. incom- petence, not fraud. The c•11<u ge of conflict of interest thus Commentary

carries the implication that the work is; possibly fraudulent, but the charge impugns .m investigator <is possibly being deceitful without requiring the accuser to make an open al- legttion of fraud. Nor does the charge put any burden on the accuser to pirove<iny «'rongcioing or incompetence. I use the terms charge and allegatiox here deliberately. Theoretically, labeling someone as having a conflict of inter- es t does not constitute an accu4ition. Nerertheless. the label ofcon:lict of interest is so commonly used with the intent to discredit a person ora work that it is disingenuous foranyone to claim that no accu,~ation is intended when de_ cribing con- flicts of interest. Part of the problem with the current em- phasis on conflict-of-interes t disclosures is that those who are innocent of fiaud or of any slanting of their n•ork are tar- nished along with the guilty, without any real knowledge of who has in fact been influenced by a financial lure or some other factor, Can Scientists Be Objective? Most people think of objectivity a.-, a state of mind, a Holy Grail of mental outlooks. Other: doubt that a mind can have an~•thi.ng in it and still be objective, since all of life 5 expe- riences are shaping to some degree. If one presumes that ties to business preclude objectivity, it seems as if we must look elsewhere, presumably toward untainted government or ac- ademic scientists, to find objective minds. But can we cor- rectly presume that government scientists have the high ground with respect to being disinterested in research re- s ults? Are academic scientists immune to pressures toward specific research outcomes? Obviously not: both government and academic scientists may tirell have an interest in obtain- ing provocative results, since publicity and prominent pub- lication may bring the rewards of promotion and further research funding. Indeed, for today 5 scientists, the rewards of publicity in terms of future research funding and career advancement may mean considetably more than personal financial gains. The previously published findings ofcilleagues, friends, or researchers in collaborating groups may also con- stitute a set of pressures favoring some study outcomes over others. No one works in a>,•acuum, and everyone's career is influenced in some way by what one publ:s hes. The reality i_7 that despite our best hopes, no one is truly free of pressures that might distort intellectual endearors. The reason to ask for disclo`ure of conflicts of interest is to in- form readers about a possible lack of objectivity. But if objec- tivity is an ideal state of mind, and everyone issubject to the iu- fluence of some forces that reduce objectivity, a more direct %vxy of assessing someone s objectivity would be to ask specifically about it rather than asking about stock portfolios. Journals could simply ask authors to disclose how objective they think they are. Cynics would argue that authors cannot be objective about assessing their own objectivity, so their responses would not be usefuL It is also easier to lie about one's objectivity than one's financial ties, for which recotYLs usually e.dst. For these rea- sons, asking directly about objectivity might not have a high enough sensitivity in detecting biased work to satisfj• zealots. On the otherhand,although relyingon the rrnk factorof financial ties may have more sensitivity, it has a lower speciScity, fingering the innocent along with the guilty. It remains unclear wh°ch ap- proach is worse. At least asking about objectivity itself haz!, the virtue of being direct, and allows the author the opportunity for self-evaluation where it really matters. Science, in its newpolicy, does ask authms to assess their own objectivity, but only in the negative: it it,5ks whether there is any factor that they believe has compromised their objectiv- ity. Koshland,a appiyinga double standard, addressed the ques- tion of the objectivity of the eclitois at Science, whose salaries are paid in part by adve•tiseia of products that might relate to the content of some articles. Koshland atsstu•ed readers that ad- vertisements will not influence editorial policy-that is,11e as- s ured the readers that the editors will be objective despite the financial influence of advertisers. Perhaps so, but this <tssur- ance is a self-assess ment. The editors of Science do not permit their contributing authors the same privilege. Objectivity May Be Free Interchange, Not a State of Mind Diogenes never found an honest man. The search for an objective person may be equally futile.On the other 11and, the search for objectivity itself may hold more promise, because not everyone agrees that objectivity is a state of mind. In The Opeu Societ y aud Its Exe»ries, Popper' «z•ote that objectiv- ity lies in the openness of free and critical interchange, and is therefore a process, rather than the state of anyone's mind. Popper wrote that scientists, like anyone else, "take many things as self-evident ... [and] accept them uncritically and even with the naive and cocksure belief that criticism is quite unnecessat•y:" d"' Acco.•cling to this view of objectivity, open- ness in scientific communication compensates for the fi<tilty of human opinion; communication n•ill achieve an objective result if criticism, whatever its motivation, is applied vigor- ously to all work and all ideas, refuting those theories that are Avz•ong and letting only the good ones survive. The objectivity of the process depends on each contribution receiving its due regard, whatever the motivations for bringing it. It depends on judging aAsork on its merits, rather than on the inferred state of mind of the author. It is precisely this social process of open dialogue and criticism that is threatened by editorial policies that focus attention on the circun-L-tances of the writer rather than on the substance of the writing. The ironic result of pointing attention away from the work and toward the author may be thus to stifle objectivity, by taking our attention away from substance and shifting it toward credentials. Does Censorship Help Readers? As an example of where this emphasis on mental purity can lead, consider the recent policy of Tlre .\'etu Engla nd Jou »ial of 3lediciue. Not satisfied with financial disclosure, the editor announced in 19J0 that The A'eo Errglaud Journal of 31ed- iciue will simply refuse to publish revien• articles oreditorials written by someone employed by a "company whose product figures prominently in the article or with a company making a competitive product.... «'hen authors have a financial as well as a scientific interest in their subjects, questions inev- itably arise that cast doubt on [their] objectivity:" This policy would have barred Thomas Edison from writing an editorial on the future uses of electricity, to spare readeis from being swayed by his possibly tainted views. The spirit of this policy n-ould also bar the director of the National Institutes of Health from writing eclitorial~ on the future of the organi- zation that pays his or her salary, because he or she has a financial tie to the subject. With this poliey of Tke .l'ere Euglmnl Joru•)ral of 1ledic•i xe, the concern about publishing work that might be tainted by JAMA. June 2. 1993-Vol 269, No. 21 Commentary 2783

an investigator's financial interests has mushroomed into a policy of censorship. While this policy will no doubt insulate readers ofThe Afem EuglandJourual from some biased views, not all readers appreciate being so coddled ° The New En- gland Journal policy dispatches any pretense that a work should be judged on its merits, since the work is deemed unfit for publication solely on the basis of the identity of the author. Exactly What Does the Reader Do With Disclosed Information? N'Vllen disclosure policies stop short of blocking publication, it remains unclear exactly how conflicts of interest are to be taken into account by editors, reviewers, or readers. Edito- rial policies usually explain that the disclosed information is just something that readers ought to have, and that such information can ottly help to understand the work. But this is most certainly not just knowledge for knowledge's sake. Exactly how is this information to be put to ase? If it is not to be used at all, why make disclosure mandatory? Some say that we should use the information by directing more intense scrutiny toward papers written by suspect au- thors who might be subject to "conflicts." This advice is misguided, because it implies that we should not take as critical a look at the work of supposedly unbiased authors. To the contrary, we should bring searching criticism t,, bear on the substance of any work, not just those papers that arouse our suspicions because of possible biases. It seems to me that the only possibility for using the in- formation is to adopt a skeptical attitude toward work for which a potential conflict of interest exists, unrelated to the substance of the work. Or. to put it more directly, to make an ad hominem judgment. But if the intent of these new policies on the disclosed information is to foster ad hominen judg- ments, which it seerrus, to be, then the intent is to push readers and editors alike toward irrationalism, steering us away from judging a work on its merits. Lofty motives notwithstanding, when readers are urged to judge a work by the source of funding, the sexual orientation of the investigator, the in- vestigator's religious beliefs, and other such considerations, irrationalism is the end resuit. Mislabeling People Is Unethical Judging someone's work by the funding source, or by any other characteristic other than the content, raises an ethical problem.10 The ethical problem is similar in principle to the dis- credited practice of judging college applicants by their photo- graphs. If you are willing to skip over the content of a work and weigh it by externalities, you infringe on the rights ofthe writer to a fair hearing of his or her ideas and findings. After doing so you could hardly expect the privilege of having your own ideas judged by their merits, rather than by who you are. Since there are no official boundaries on what could be the reason for a con- flict of interest, whenever we stray from using anything but the substance of a work itself as the basis for judgment, we begin to substi;ute prejudice for reason; we abridge the rights of oth- ers and convert the free interchange of critical views into a shouting match about pedigrees. Let's End This New McCarthyism in Science Wbere will this juggernaut end? Should readers demand that the financial portfolios of editors and reviewers be pub- lished as footnotes to the masthead? Beyond journals, we have the problem of textbooks, for which we do not have editorial police to help us. We also have the problem that any conflicted soul is still allowed to speak at scientific meetings- although some meetings are now requiring disclosure state- ments for speakers. Should we soon expect to see color-coded name badges at meetings to identify the objective and not- so-objective speakers? Like McCarthyism in the 1950s, this frenzy of credential checking is apt to get way out of hand. Theoretically, even identifying the author of a work is problematic, since readers may judge a work by the repu- tation of the author or institution where the work was done. Thus, one could argue that publication should be anonymous. Realistically, however, authors and institutions need to be identified on published work, and sponsorship often needs to be identified as well. I would not argue in favor of suppressing any of this information, or anything else that authors or editors wish to reveal. On the other hand, it seems unnec- essary to have disclosures become a mandatory part of the review process. Just as it makes sense to mask the identity and affiliation of an author from reviewers (a step that only a few journals take) to keep the reviewers focused on the substance of the work, rather than on personal issues," it makes sense to keep the revelations about potential conflicts out of the review process. Let the disclosures, if they are to be made, come after the publication decision. I suppose that there will always be some people who prefer to make ad hominem judgments over substantive ones. The trend in journal policies, however, is to sanction such judg- ments by elevating the importance of the disclosures that fa- cilitate them. When journals start imposing censorship to pre- serve "objectivity," the pendulum has swung too far. Journals should assure their readers that they will keep their pages open to any views and any work of importance that is well conceived and well described. To protect uncritical readers, the option is always open to solicit opposing views. Requesting information about financial sources of support orother financial links is rea- sonable as long as the information is not used as a pretext for judging the validity of the work. Editorial policies should make it clear that the request for such information is not an attempt to classify the author as either objective or tainted. By preserv- ing everyone's right to a fair hearing, journals can keep the di- alogue open and keep the process of communication objective, in spite of individual biases. Then we can halt this new \1cCar- thyism in science and get back to focusing on the work of a sci- entist rather than on his or her life story. Kenneth J. Rothman, DrPH I am grateful to DimitriosTrichopoulos,llD; Jlatthetc Gillman, \ID; Stephan Lanes, PhD; Alexander Walker, biD; Nancy Dreyer, PhD; Cristina Cann; Harris Pastides, PhD; and Cuttis I/llison,,NID, for helpful advice. 1. Rennie D, Flanagin A, Glass R.M. Conflict of interea. in the publication of>cience. /.1SlA. 1991...'G(i2fiG-2G7. 2. RelmanAS. Deal3ng tiith contlict.e of interest \' Eng! J S1ed 1~1,i1G:l Ii L-I1~ 3. 3. Marshall E. when does intellectual passion become conflict of intere: t? Science. 1992;257: 62~01. 4. Koshland DE Jr. Conflict of interest policy. Scicnca 1992257:595. 5. Marshall E. Sex on the brain. Science. 1992257:C20-621. 6. Marshall E. The perils of a deeply held point of view. Science. 1992:157:G21-C"'~?. 7. Popper KR. The fligG Tide ojProphecy: Neget,.Sfarr, and the AJkrmath. In: The Open Society and Its Enentiex, vol 2. Sth ed. Princeton, NJ: Princeton University Press; 1966. & Relman AS. New 'information for authors'-and readers. N Eng1 J.ifed. 1990;;323:50. 9. Waltuch AL. New editorial policy on conflicts of interest. N Engl J.l1ed. 19`.10;;;L4:15G8-15(i9. 10. Rothman KJ. The ethics of research sponsorship. J Clix Epidemiol. 1991:44 (suppl)23S•9_RS. 11. NcNutt RA, Evans AT. Fletcher RH, Fletcher SW. The effect.s of blinding on the quality of peer revieK: a randomized trial. J~1.111. 1990;2u"3:1:i71-1'37M 2784 JAMA, June 2, 1993-Vol 269, No. 21 Commentary I I