by Steve Olson
Angell hadn't known before journalists started calling her. But it was too late to do anything about it. "Reporters had already gotten their embargoed copies," she now says. The resulting media storm "was as intense as any we have experienced at the Journal," Angell later wrote with NEJM editor Jerome Kassirer.
Angell recalls that most newspapers handled the story "fairly responsibly." But one of the authors of the editorial remembers the episode quite differently. "We were pilloried in the media in a very unfair way," says JoAnn Manson, an associate professor of endocrinology at Harvard Medical School, who wrote the editorial with University of Pennsylvania researcher Gerald Faich. She claims never to have had a conflict of interest and says that the NEJM and the media distorted her story in ways that caused her great distress. "Our positions were completely misrepresented."
Conflicts of interest, whether real or apparent, have become big news in science. Over the last few years, stories involving breast implants, global warming, and drug efficacy have all revolved around whether scientific results were influenced by the source of funding for the research.
One reason for the spate of stories is the growing thicket of connections between academic researchers and private industry. Encouraged by federal policies and cash-strapped universities, more and more researchers are seeking corporate funding, owning stock in startup companies, or serving as consultants to industry.
According to a survey directed by Sheldon Krimsky at Tufts University, a third of the lead authors in the articles published in 14 major scientific and medical journals in 1992 had a financial interest closely related to their scientific work--that is, they were listed on a patent related to the research or served on an advisory board, were an officer, or had substantial stock holdings in a company commercializing the research. These figures don't even include less substantial financial interests like consultancies and speaking fees, which Krimsky and his colleagues were not able to document from publicly available information.
These financial entanglements are forcing science writers to make some tough decisions. When a journal publishes a disclosure saying that industry paid for part of a scientist's research, does that mean the results are suspect? Should journalists and public information officers simply cite the source of a researcher's funding and leave it at that? Is it safe to assume that the policies at institutions and journals will keep tainted science from getting into print?
The flap at the NEJM demonstrates how thorny these questions can be. Since 1990 the NEJM has had a policy that the authors of editorials and review articles cannot have any financial connection with a company that benefits from a product discussed in the editorial or article. For the authors of research articles, potential conflicts of in-terest need to be disclosed to editors, who then decide whether the article should include a disclosure state-ment describing the sources of research funding.
The distinction between editorials and scientific articles is an important one, Angell emphasizes. Be-cause articles present original data, readers can judge for themselves whether the data support the authors’ conclusions. But editorials make judgments based on information selected from the literature. That’s why the NEJM has decided that anyone with a possible financial bias, like Manson and Faich, should not write editorials.
Angell insists that the NEJM has not claimed that Manson and Faich’s conclusions were swayed by their association with industry. As she has written in dis-cussing the case, “A conflict of interest is not synony-mous with bias. If Drs. Manson and Faich had not consulted for companies that stood to gain from the use of Redux [one of the antiobesity drugs discussed in the editorial], they might very well have written exactly the same editorial. We have no reason to be-lieve otherwise. But no one should have to make that judgment, least of all readers who are counting on editorialists for a balanced perspective on a some-times highly technical issue.”
Manson claims that the NEJM was at fault, not she. She points out that she had done only one week of consulting for industry, when a company paid her to present a summary of her research to an FDA advisory board. She says that the consulting ended long before she wrote the article and that she had no ongoing financial interest in the research. She also says that she made every reasonable attempt to notify the journal about the consulting she had done but that her editor misinterpreted what she was saying. “I very much wanted a disclosure stating that we had served as consultants,” she says. “It would have protected us.”
Angell points out that editorials in the NEJM can-not have disclosure statements, because authors with such conflicts cannot write editorials. She also says that Manson’s consulting was in a gray area and there-fore required further discussion, while Faich’s consultantship was ongoing at the time he wrote the editorial. In an effort to keep such a situation from recurring, the NEJM reworded its policy to read that “We require that authors [of editorials] be free of financial associations (including equity interests, consultancies, or major research support) with a com-pany that stands to gain from the use of a product (or its competitor) discussed in the editorial. If there are any questions about this policy, please phone us.”
It’s ironic that the NEJM was caught up in this affair, because it has the most rigorous conflict-of-interest policies of any major science journal. Many other journals, including Science, The Lancet, the Journal of the American Medical Association, and the Proceedings of the National Academy of Sciences, have looser policies that simply ask all authors to disclose potential conflicts to editors, who then decide what to do with the information.
For example, Science requires that submitted manu-scripts include “information about the authors’ pro-fessional and financial affiliations that may be per-ceived to have biased the presentation.” According to Science’s managing editor Monica Bradford, “It’s an in-between solution. The problem is that if people are going to try to get away with something, they won’t disclose the conflict anyway.”
That’s exactly the point, say the journal editors at the opposite end of the spectrum from the NEJM. These editors insist that disclosure statements cannot uncover real misconduct and serve to undermine the science being presented. One editor, Kenneth Rothman of Epidemiology, goes so far as to compare disclosure policies to McCarthyism, “in that these policies can damage the career of innocent scientists by false innuendo.”
As Nature—the most visible proponent of a “don’t ask, don’t tell” policy—stated in a recent editorial, “This journal has never required that authors declare such affiliations. Until there is evidence that there are serious risks of [fraud, deception, or bias in presentation], this journal will persist in its stubborn belief that research as we publish it is indeed research, not business.”
Faced with this range of opinion among journal editors, most science writers are leaning heavily on straight disclosures to handle the issue. “We don’t have time to turn every journal report into a full-blown investigative piece,” says Cory Dean, the editor of the science section at the New York Times. “Our practice has been to disclose [a conflict of interest] if we think it’s relevant.”
“Personally, I try not to assume dishonesty when I first learn about
something,” says Robert Cooke of Newsday. “I include a
funding source if it’s germane to the topic and if a reader needs to
know it.” He adds that ties with industry have long been common in
chemistry and aeronautics without drawing much media attention. Only with
the advent of biotechnol-ogy, where basic research can quickly yield commer-cial
products, has the issue become so prominent.
Science reporters point out that conflict-of-interest issues often come
up in other parts of newspapers and magazines, where their importance is
judged by the first rule of journalism: It depends. Just how newsworthy
is it, for example, when a major campaign contributor spends a night in
the Lincoln bedroom and shortly thereafter sees a favorable proposal emerge
from the White House?
For their part, many public information officers are leaning toward disclosure in press releases and other documents. “When we issue news releases about published work by an investigator, we routinely dis-close the source of funding and, when appropriate, the investigator’s relationship to any company in-volved in the research,” says Joann Rodgers, the deputy director of public affairs at Johns Hopkins University. “Many of us here are former journalists and we’re sensitive to how the press covers and perceives con-flicts of interest.”
Such disclosures are essential if readers are to inter-pret a science story accurately, say many media watchers. “The first responsibility for the reporter is to get the information clearly and accurately out to the public,” says Tufts' Krimsky. "Disclosure provides the public with a much better sense of skepticism about the research being described."
Yet it is also clear that an unadorned statement of funding sources can plant seeds of doubt in readers' minds, whether those doubts are intended or not. For example, John Bailar, an epidemiologist at the University of Chicago, says that when he reads about research funded by industry, his reaction is "Fine, go ahead and take industry money, but don't expect me to take your paper as seriously."
The problem with a simple disclosure, says Barbara Culliton of Nature Medicine, is that it can imply misbehavior by calling attention to itself. "There can be a presumption of guilt if you took money from industry, and a presumption of innocence if the money came from government or a foundation," she says. "But everyone has biases. Who are we kidding?"
In fact, these other biases can be much more influential than financial conflicts of interest. Some researchers are so professionally wedded to particular points of view that they interpret everything from a particular perspective. Like journalists, scientists have a strong self-interest in publishing, even if it means exaggerating a point to impress an editor.
"The conflicts that are right on the surface don't worry me as much as the ones that are hidden," says Bailar. "Like the investigator who depends on NIH grants and knows that if the present work is not published, the next grant is less likely to come along."
Traditionally, peer review has served as science's bias detector. Many editors say that the best protection against bias, whether from financial interests or other sources, is a Rolodex full of good reviewers.
The National Science Foundation and National Institutes of Health also have policies requiring that researchers inform their institutions if they have financial interests--including consulting arrangements or patent rights--in companies that might be affected by their research. These policies have in turn helped public information officers convince investigators and administrators that it is important to publicize funding sources.
Still, institutional policies cannot cover all circumstances, like the researcher invited by a journal to discuss a particular area of science. Nor are institutional policies--or even peer review--guaranteed to catch the researcher who is determined to deceive.
Better policies and more appropriate rules can help, but they can't solve the problem. Like journal editors, science writers will always have to make judgments about what information is important and how that information should be presented. It comes with the job, says Boyce Rensberger of the Washington Post: "That's one of the things that reporters are supposed to do--be on the lookout for conflicts of interest."
Steve Olson is a freelance writer in Bethesda. MD.