© 2003 American Thoracic Society
Conflicts of Interest and AJRCCMRestating Policy and a New Form to UploadAuthors submitting a manuscript to AJRCCM are instructed to disclose any financial relationship with any biotechnology manufacturer or pharmaceutical manufacturer that has an interest in the subject matter or materials discussed in the manuscript. The number of disclosures received is far fewer than estimates of commercial relationships. The discrepancy suggests that many authors do not fully understand what constitutes a conflict of interest and why AJRCCM requires disclosure of this interest. In this editorial, I discuss the definition of conflict of interest, why journals require disclosure, those interests that need to be disclosed, and a new form that authors need to upload at the time of submitting a manuscript online to AJRCCM. Definition of a Conflict of Interest A conflict of interest is a set of conditions in which professional judgment concerning a primary interest, such as the validity of research, might be influenced by a secondary interest, such as financial gain (1). The secondary interest is usually not illegitimate in itself, but it becomes a problem when it eclipses the primary interest. With this definition, a conflict of interest is a condition, not a behavior—being determined by circumstances, not by outcome. A conflict exists not only when judgment has been clearly influenced. It also exists when judgment might or might be perceived to be influenced (2). That is, a conflict exists before any actual breach of trust, and irrespective of whether a breach of trust actually occurs. The terms "competing interest" or "dual commitment" are used by some journals to minimize negative connotations of the word "conflict." Because "conflict of interest" is more widely used, AJRCCM will continue to use it. Objectivity, Belief, and Trust The goal of biomedical research is to improve the life and longevity of patients through the production of objective new knowledge. As with any other form of knowledge, science is subject to error (3, 4). Widely accepted principles have in time been proven wrong. Because knowledge is not absolute and objectivity is never a pure state, we range between doubt and belief when we read (5). A scientist scrutinizes a new report with a sprinkling of healthy skepticism, one of the norms of science (6). But an attitude of total skepticism is corrosive and inimical to learning. Somewhere in the middle between the extremes of total skepticism and uncritical belief lies trust. The overwhelming proportion of the information we come to accept as knowledge derives from trusting others. To be able to inform and guide us, a scientific report must be believable. But belief in any knowledge claim is compromised if we suspect that the claim serves the personal interests of the claimant. Disinterestedness The converse of interest is disinterest, and disinterestedness is another traditional norm of science (7). (Norms are the values or moral abstractions by which a community decides between what is right and wrong [6].) A disinterested researcher is motivated by curiosity, pursuing knowledge for its own sake, and is more interested in the rewards of intellectual discovery than in personal gain. Disinterestedness fosters trust in the objectivity of a scientist. The norm of disinterestedness comes from a time when research was supported by patronage rather than commercial sources, a time when investigators did not imagine financial gain. When asked who owned the patent on the polio vaccine, Jonas Salk replied, "Well, the people, I would say. There is no patent. Could you patent the sun?" (8). In contrast with a tradition of science for the sake of science, the primary obligation of a manufacturer is to shareholders: to deliver a sound return on investment. The large clinical trials sponsored by industry involve considerable investment of company money and investigator time. As links between researchers and industry continue to grow, the traditional norms of science are pitted against financial incentives and commercial agendas of the business world, and the public image of universities as independent and disinterested creators of knowledge is increasingly challenged (9). Governmental Encouragement of Partnerships with Industry Governments have played a substantial part in enticing universities to embrace economic interests and entrepreneurial values (10). In 1980, Congress passed the Bayh–Dole Act, which allowed universities for the first time to patent the results of federally funded research without requiring governmental permission. By removing previous barriers to partnerships between academic institutions and industry, the amendment created new incentives for lucrative relationships. In many countries, granting agencies have become increasingly interested in the commercial exploitability of the research they support, and potential links with commerce have become a major part of grant applications (11). The change in relationships between investigators and industry has been sharpest in patient-based research, because governments provide less support for this type of research than for basic research. In 1999, the National Institutes of Health spent $17.8 billion on research, mostly on basic research, and the top 10 pharmaceutical companies spent $22.7 billion, mainly on clinical research (12). In many countries, including the United States, investigators who do mechanistic research at a patient's bedside can continue this line of research only by also doing less intellectually stimulating studies for industry. Not withstanding instances of egregious behavior by manufacturers and investigators, industry also supports many well-designed and well-conducted trials, which are objectively reported even to the detriment of product development or sales. Over the last decade, the location for conducting clinical trials has changed. Between 1991 and 1998, the share received by academic medical centers decreased from 80 to 40% (13). More and more trials are conducted in community hospitals and coordinated by contract research organizations that pay private physicians a gratuity for recruiting patients into a study (14). Prevalence of Conflicts, Their Influence, and Infrequency of Disclosure Several studies have revealed that financial interests have an important influence on research. Stelfox and coworkers (15) classified 70 articles on the safety of calcium channel antagonists as supportive (30 articles), critical (23 articles), or neutral (17 articles). They contacted the authors and inquired about financial relationships. Almost all of the supportive authors (96%) had financial relationships with manufacturers of calcium channel antagonists, as compared with 60% of the neutral authors and 37% of the critical authors. Combined data from a total of 1,140 original studies (16–23) reveal that industry-sponsored studies are 3.6 times more likely to reach conclusions favorable to a sponsor than are studies sponsored by nonindustry sources (24). Chaudhry and coworkers (25) asked two groups of readers to read an article, telling one group that the authors had a conflict of interest and telling the other group that the authors were not conflicted. Although the articles were identical, readers rated believability 18% lower when they understood the authors to be conflicted. In the study by Stelfox and coworkers (15), 63% of the authors had a financial relationship with a manufacturer, yet a relationship was disclosed in only 2 of the 70 articles. Krimsky and coworkers (26) studied 789 articles published in 14 leading journals in 1992. Although 267 (34%) of the lead authors were later found to have a significant financial interest, no voluntary disclosure was published. In a study of 61,134 articles published in 181 journals in 1997, Krimsky and Rothenberg (27) found a positive disclosure of conflicting interests in only 0.5% of the articles. The editors of the journals were contacted, and 74% said that they always or almost always published statements of positive disclosure. As such, the low rate of published financial disclosures signals poor compliance by authors. Money and Other Conflicting Interests Conflicting interests are disclosed to alert readers to the possibility of bias. But every author and every individual in society has involvements that might lead to bias. Self-interest motivates everyone at times. Nonfinancial conflicts include the desire to achieve research findings worthy of publication in an elite journal, pressure to get promoted, and the hope of winning a prestigious prize. Pressure to get a grant from the National Institutes of Health can induce as much bias as money from a manufacturer. Nonfinancial conflicts are inherent to research. In contrast, financial conflicts are not inherent to research—they are optional. Researchers are not forced to have financial conflicts, and restrictions based on them do not violate academic freedom (28). The public cares little about academic conflicts, but is alarmed by financial conflicts (10). Financial conflicts are invisible unless disclosed. Disclosure has focused on financial relationships because money is quantifiable, exchangeable, and easier to regulate by impartial rules (1). True, it is not possible to implement a policy that enforces disclosure of every nonfinancial interest. But that limitation does not justify the failure to implement a policy that rules on financial interests (1). The Case for Ignoring Conflicts Some critics claim that disclosure policies are detrimental to science. They argue that an article should be judged solely on its merit and that disclosure diverts a reader's attention from the science. Rothman (29) denounced policies on conflicts of interests as a new McCarthyism, inviting ad hominem judgments and risking damage to the careers of innocent scientists through false innuendo. As recently as 1997, Nature (30) boasted of its opposition to disclosing financial interests, saying that the journal stubbornly believed that "research as we publish it is indeed research, not business" and that the arguments about "financial correctness" were less than compelling. Evidence has been steadily mounting that financial relationships with industry can have significant effects on research: authors are not only more likely to favor a sponsor's product (15–24), but companies have insisted on suppression of unfavorable results or delayed their publication (16, 31–35). The 1999 death of Jesse Gelsinger, a volunteer in gene transfer experiments, was especially disturbing to the public because the university and one of the scientists had equity in a company that was expected to profit from the research (36). Criticism of disclosure policies has abated. Rothman (37) now supports disclosures. Nature introduced a disclosure policy in October 2001 "in the interests of transparency and to help readers to form their own judgments of possible bias" (38). Management and Disclosure To ensure public trust and to avoid government-imposed regulations, journals need to minimize the harm that might arise from a conflict of interest. Complete avoidance of any financial relationship is the simplest approach, but not practical. Prohibition would exclude many knowledgeable authors because they are in high demand as consultants by industry. Relying solely on the good character of the author is the least intrusive, but many authors fail to disclose significant relationships (15, 26, 27). A third option is to supplement individual responsibility with community regulation. Disclosure of financial relationships is the tool most widely used by journals for managing conflicts. Disclosure does nothing to lessen a conflict, it simply publicizes it (2). Although not sufficient to resolve conflicts, disclosure is a necessary step in fostering trust in the material published by a journal. Readers can bear in mind the potential for bias when reading an article by a conflicted author and modify their opinion on the credibility of the statements (2). Response of Authors When confronted with an allegation of conflict of interest, authors reply, "I didn't think that the policy applies to the type of relationship I had," "The amount I received wasn't significant enough to merit declaring it," "This is an invasion of my privacy, and I don't want the information publicized," and "Your accusation is unjustified—show me the evidence that I was biased." Infringement of privacy would apply if the author were writing in a private diary as opposed to publicizing a claim in a journal and hoping that readers believe it. "Whenever a man assumes a public trust," wrote Thomas Jefferson, "he should consider himself as public property." Without readers trusting the words of an author, he or she has nothing. Many confuse the terms "bias" and "conflict of interest," especially when either term is preceded by "potential" (39). Some authors have a financial relationship with a manufacturer—which constitutes a potential for bias. Because the authors believe that this potential for bias was never converted into an actual bias, they conclude that they had only a potential conflict of interest and not an actual conflict. But an author who has received any money from a manufacturer with an interest in the subject of his or her manuscript has an actual conflict of interest. It is for the reader—not the author—to decide whether or not the conflict resulted in bias. A New Form for Authors Responsibility for disclosing a conflict relies on the conscience of the authors. But journals have an obligation to do all that they can to ensure that the information is made available to readers. The evidence shows that asking authors to tick a box is not sufficient to extract the required information (27). The failure of authors to declare a conflict does not necessarily mean that they don't have one (15, 26). To prevent ambiguity, AJRCCM is now requiring authors to state explicitly whether they have or do not have a financial relationship with a manufacturer that has an interest in the subject matter of their manuscript. At the time of submitting a manuscript, authors must upload a form providing explicit answers to a series of concrete questions on financial relationships (form available in online supplement). The form will be accessible to reviewers in the online supplement during peer review. Once a manuscript has been accepted, the form will be deleted and a disclosure footnote will be published in the article. The BMJ introduced such a form in 1998 in response to the recommendation of Stelfox and coworkers (15). The following year, the rate of disclosures increased to 22% as compared with 2% in 1996 (40). The Association of American Medical Colleges and the United States Public Health Service define a significant financial interest as the receipt of more than $10,000 per company per year; consulting fees, honoraria, and other in-kind compensation should be combined. The guidelines have been criticized in that amounts approaching but less than $10,000 per company per year are sufficiently large to bias an author in favor of a sponsor's products (41). The limit also pertains to a single company. Some authors receive money from several companies, all making a single class of products (say, inhaled glucocorticoids, cephalosporins, or calcium channel antagonists). Payment from individual companies may be less than $10,000, but the aggregate could be much more. Authors writing about those agents as a class could have an incentive of much more than $10,000 to write in their favor, yet are not obliged to declare the conflict (42). AJRCCM recommends that authors be explicit and detailed when listing money received from a manufacturer that has an interest in the subject of their manuscript. We require authors to list receipt of more than $10,000 per company per year in any one of the 3 years preceding submission of a manuscript. Because the threshold of $10,000 is necessarily arbitrary, we also encourage authors to list lesser amounts on the disclosure form. If authors are uncertain about how much to disclose, it is best to err on the side of disclosing everything rather than too little. When deciding how much detail to provide, we advise the "red face" test. If the nondisclosed information were subsequently publicized, would you read the notice with serenity or with a red face? The policies of some journals leave it for editors to decide whether or not to publish disclosure statements. Such policies have aroused charges of paternalism (42). Critics have asked how the publication of a disclosure statement could interfere with a reader's ability to judge the potential for bias in an article (43). At AJRCCM, we believe it is readers who make the greatest investment in an article after it has been published—investing that most precious resource, time, in reading it. Before deciding to trust the contents of the article, readers should be told whether the authors have, or do not have, a financial conflict. The reader is the best person to judge whether $9,999 is a less significant amount of money than $10,001. Consequences of Failing to Disclose When readers suspect authors of failing to disclose a financial relationship, they write to a journal (in the same way that readers raise questions about duplicate publication [44]). The American Medical Association recommends (45) that queries be forwarded to authors for a written explanation. If the authors are found to have an undisclosed financial relationship, the American Medical Association recommends that the journal publish the information in its correspondence columns. The BMJ has a similar policy (46). Some journals impose stricter penalties for failing to disclose a financial relationship. When "trust has been significantly compromised by an author's actions," Nature (38) "will seek to redress the matter by an appropriate combination of sanctions and communications to readers and employers." More extreme is Gastroenterology, the editor of which (47) will "preclude from consideration any further manuscripts by the involved authors." At AJRCCM, we will forward queries about undisclosed financial relationships to authors and publish new disclosures in our correspondence columns. Conclusion Editorial policies on conflicts of interest are easily misinterpreted, in part because the subject has a sanctimonious air about it. The issue is not about dishonesty or guilt, but about transparency and trust. Progress in medicine ultimately depends on the voluntary participation of patients in studies and they must be given no reason to mistrust the process. Maintaining public trust is a community responsibility that relies primarily on the conscience of the investigators. Although editors do not get involved until long after the last data points have been collected, they still have a responsibility in maintaining public trust in published research. Editors have an obligation to provide readers with the information needed to judge the believability of articles they read. Readers do not owe journals trust. Instead, journals need to earn and maintain the trust of their readers. FOOTNOTES Conflict of Interest Statement: M.J.T. is editor of AJRCCM and is responsible for developing its policy on conflict of interest. He receives a fixed stipend from the American Thoracic Society, and will not be affected financially by the success or failure of AJRCCM's policy on conflict of interest. He does not receive financial support for research from pharmaceutical, biotechnology, or medical device companies. He does not serve as a consultant to or on the advisory board of any company. He receives royalties for two books published by McGraw-Hill, Inc. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
