We don’t know whether conflicts of interest influence researchers, but we certainly know they might
By Marcia Angell, M.D.
Boston Globe, April 11, 2016
A front-page story by Charles Ornstein in the Globe reports increasing concern about editorial decisions at The New England Journal of Medicine (NEJM), one of which is to defend the practice of medical researchers having personal financial ties to drug companies whose products they are testing. The NEJM’s national correspondent referred to critics of such conflicts of interest as “pharmascolds.”
I was on the editorial staff of NEJM from 1979 until 2000, serving as executive editor from 1988 until 1999, and then as editor in chief before I stepped down in 2000. So I have firsthand knowledge of the evolution of its conflict-of-interest policies and the reasons for them.
In 1984, the NEJM became the first medical journal to require authors of papers submitted to it to disclose all financial ties to companies that could financially benefit from the results. The editor at the time, Arnold S. Relman, was concerned by the growing financial associations — in addition to research sponsorship — between drug companies and supposedly independent researchers, and by the influence of sponsoring companies on the way that research was conducted and reported.
For example, researchers increasingly served as paid consultants or members of speakers bureaus for the same companies whose products they were evaluating. Moreover, companies were asserting a right to be involved in all aspects of studies they sponsored, including writing the papers and deciding if, when, and where they would be published.
Within a few years, other major medical journals followed NEJM’s lead, and instituted similar disclosure policies. Six years later, Relman extended the policy by banning altogether any industry ties for authors of editorials and other opinion pieces, because these types of articles contain no data that readers can judge for themselves. Thus, by 1990, the NEJM had the most stringent conflict-of-interest policy of any medical journal. Relman’s successor, Jerome P. Kassirer, continued the policies, as did I. But with the arrival of the current editor in chief, Jeffrey M. Drazen, the outright ban was ended, and the days of the “pharmascolds” were over.
Financial conflicts of interest matter. They can best be defined as any financial association (or promise of one in the future) that would give researchers an incentive to distort their work. For example, if a researcher is doing a study comparing drug A with drug B, and has a large equity interest in company A, he or she would benefit if drug A is shown to be better. That doesn’t mean that the researcher would not be scrupulously honest, only that there would be an incentive to shade the work in favor of drug A.
So a conflict of interest is not synonymous with distorted work; it simply raises the question. But how are physicians who read a paper by authors with conflicts of interest supposed to know whether the work was disinterested or not? Unless bias is blatant, there is simply no sure way to tell. Adding to the difficulty is the fact that bias can be unconscious. The researcher may seek to rise above his or her financial interest, but it would be human nature to lean in the desired direction without meaning to. So we don’t know whether conflicts of interest influence researchers, but we certainly know they might.
Unfortunately, there is much evidence that financial conflicts of interest are in fact distorting medical research. Studies of drugs sponsored by manufacturers are more likely to conclude the drugs are beneficial than research sponsored by the National Institutes of Health. Worse, industry-sponsored research that does not show benefit is very often not published at all. For example, a review (published in NEJM, Jan. 17, 2008) of 74 company-sponsored clinical trials of antidepressants found that 37 of 38 positive studies — that is, studies that showed effectiveness — were published. But of the 36 negative studies — those that failed to show effectiveness — 33 were either not published, or published in a form that wrongly conveyed a positive outcome. The cumulative result of this sort of thing is that both the public and physicians come to believe that drugs are better and safer than they are.
Judges with equity interest in Company A would not hear a case involving a dispute between Company A and Company B. Similarly, a reporter for The Boston Globe would not be permitted to write a story about Company A if he or she had a financial interest in it. We instinctively know why that should be so. Why should it be different for medical researchers?
Marcia Angell, M.D., is former editor in chief of The New England Journal of Medicine and is on the faculty of Global Health and Social Medicine at Harvard Medical School.