|
|
| |||||||||||||||||||||||||||||||||
What we did not appreciate until it was too late was that Dr. Manson and Dr. Faich had both been paid consultants for companies that stood to gain from the sale of one of the antiobesity agents studied. In 1995, Dr. Manson consulted for Interneuron Pharmaceuticals and for Servier Amérique, the manufacturer and a distributor of dexfenfluramine (Redux), one of the most common of the appetite suppressants studied by Abenhaim et al. Her coauthor, Dr. Faich, had also consulted for Servier Amérique, beginning in 1994, and for Wyeth–Ayerst, the North American distributor of Redux, beginning in April 1996 and continuing until as recently as the month of publication. We did not become aware of the essential features of these associations until three days before the publication date, when the first of many reporters phoned to ask us about the conflict of interest. The ensuing publicity was as intense as any we have experienced at the Journal, but despite the generally accurate media coverage, we owe it to our readers to set the record straight ourselves.
Since 1990 the Journal has had a policy that prohibits editorialists and authors of review articles from having any financial connection with a company that benefits from a drug or device discussed in the editorial or review article.3 This policy was an extension of our earlier policy, announced in 1984, that required authors merely to disclose their financial connections with industry.4 In the case of scientific reports, disclosure seemed adequate. Scientific reports are self-contained. They present original data, and readers can judge for themselves whether the authors' interpretations are supported by the data. Editorials and review articles are different. They are not self-contained, and there are no primary data. Instead, editorialists and authors of review articles evaluate an issue on the basis of what they select from the literature as relevant. In the case of editorialists, their task is to use the study they have been asked to editorialize about as a springboard for an open-ended consideration of an important issue. It is expected that they will provide an unbiased and authoritative opinion about the matter. That is why we insist that editorialists have no financial ties to products that figure prominently in their work.
There is no better demonstration of the reason for our policy than the Manson and Faich editorial. Although Abenhaim et al. found a significantly increased risk of primary pulmonary hypertension in subjects who had taken antiobesity drugs (the adjusted odds ratio for those who had taken appetite suppressants for more than three months was 23.1, with a 95 percent confidence interval of 6.9 to 77.7), Manson and Faich calculated the offsetting risk of obesity itself and found it to be greater. They concluded their editorial with the sentence, "Although physicians and patients need to be informed, the possible risk of pulmonary hypertension associated with dexfenfluramine is small and appears to be outweighed by benefits when the drug is used appropriately." When considered as the opinion of unbiased experts, this is just the sort of practical summary we want from editorialists. But when considered as the conclusion of people who were paid consultants for companies that sell dexfenfluramine, it raises troubling questions. Did the authors give sufficient attention to the possibility that the risks associated with long-term use would be even greater? Were they too quick to attribute the risks associated with obesity to obesity itself? Were they too dismissive of other, nonpharmacologic treatments of obesity? Mere disclosure of their conflicts of interest would not have answered these questions for readers not expert in the field.
Critics of our policy might contend that it is unnecessary, because experts cannot be influenced by financial ties. That is probably true for the great majority of them, including Drs. Manson and Faich. But some no doubt can be influenced, and readers should not have to go through the exercise of assessing editorials to decide which to trust. Furthermore, even if very few experts would knowingly tailor their opinions for gain, close association with a company can engender good will that may manifest itself as unconscious bias, a condition that is probably much more frequent and even more difficult for readers to assess. Editorialists who have worked with a company are also accustomed to seeing issues from that company's perspective, and this viewpoint can affect the focus of an editorial. As we pointed out earlier, a conflict of interest is not synonymous with bias. If Drs. Manson and Faich had not consulted for companies that stood to gain from the use of Redux, they might very well have written exactly the same editorial. We have no reason to believe 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 sometimes highly technical issue.
On June 24, 1996, before the editorial was written, we sent Dr. Manson the standard letter to prospective editorialists. In it is a paragraph set in boldface: "Because editorials involve interpretation and opinion, we ask that authors not have ongoing financial associations (including equity interest, regular consultancies, or major research support) with a company that produces a product (or its competitor) discussed in the editorial. If there are any questions about this policy, please phone us." Because Dr. Manson's consultancy was not "ongoing" at the time she was asked to write the editorial, she evidently believed she was not precluded by this paragraph from writing the editorial. Dr. Faich's consultancy was ongoing, but he did not interpret it as "regular." Accordingly, in the covering letter they sent us with their editorial was the statement: "Drs. Manson and Faich participated in the FDA review of dexfenfluramine as scientific consultants on the health risks of obesity. They have no financial interest or equity in any pharmaceutical company producing anti-obesity agents." Nothing was said in their letter about being paid as consultants by companies that manufactured or marketed Redux. Dr. Manson had disclosed a connection with Interneuron and Servier in a scientific report she coauthored in the Journal last year,5 but we gathered from a telephone conversation that it was not relevant to the editorial.
The purpose of the words "ongoing" and "regular" in our standard paragraph quoted above was to allow for the common circumstance of editorialists writing about a product made by a company with which they have had a past, no longer relevant financial association. Dr. Manson's connection with Interneuron and Servier Amérique did not fit that description. Although her consultancy was not ongoing, it did concern the very drug she was being asked to consider in her editorial. It was in a gray area not explicitly covered by the language of our policy and therefore required further discussion. Dr. Faich's connection, however, was in no such gray area, and if we had known the facts of his consultancy we would not have permitted him to coauthor an editorial on the subject. In the Correspondence section of this issue of the Journal (page 1064), Drs. Manson and Faich present their view of the matter.
So where do we stand? No one can say for certain whether the views of Manson and Faich were in any way influenced by their associations with Interneuron, Servier, and Wyeth–Ayerst. Readers will have to decide for themselves, difficult though that is. But the episode should be put in perspective. To our knowledge, this was the first violation of our policy in the six years it has been in effect. That is not a bad record. Disturbing as the episode was, we see no reason not to continue to trust prospective editorialists to honor the policy. Trust is essential in scientific research, and we believe it is rarely betrayed. Prospective authors need to remember the reason for our conflict-of-interest policy; if that is clear, then the exact wording of the statement in our letter becomes less important. Nevertheless, we intend to make the statement more inclusive by modifying it slightly to read: "Because editorials involve interpretation and opinion, we require that authors be free of financial associations (including equity interest, consultancies, or major research support) with a company 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. Otherwise, please sign the attached form attesting that you have no such associations."
After the Journal instituted its 1984 policy requiring disclosure of conflicts of interest, other major medical journals adopted similar policies, but so far we stand alone in our contention that disclosure is not enough in the case of editorials and review articles. In our view, the increasing involvement of researchers in commercial activities makes this policy all the more important. Readers must be able to rely on editorialists to be disinterested. We hope that we will soon be joined in our policy by our sister journals.
Marcia Angell, M.D.
Jerome P. Kassirer, M.D.
References
| |||||||||||||||||||||||||||||||||
Related Letters:
Editorials and Conflicts of Interest
Salzman R., Stone L. R., Douglas E. F., Blank N. K., Erisman M. D., Angell M., Kassirer J. P.
Extract |
Full Text
N Engl J Med 1997;
336:728-729, Mar 6, 1997.
Correspondence
Disclosure of Authors' Conflicts of Interest — A Follow-Up
McMasters K. M., Recker R. R., Goldberg B., Angell M., Utiger R. D., Wood A. J.J.
Extract |
Full Text
N Engl J Med 2000;
343:146-147, Jul 13, 2000.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
