When we assumed stewardship of this journal in 2001, one of the first topics the editors discussed was our policy on conflict of interest. Four years later, we are still discussing it.
As you might infer, we are not of one mind. While we recognize that financial interests can corrupt the integrity of the scientific process, the remedy is less clear. Just because prominent clinical journals have plunged into this voyage is no proof of safe passage. Meanwhile, the epidemiology journals have been quiet on this topic. What prompts us now to venture into these troubled waters?
One factor is the manifold sins of the tobacco industry.1 For decades the industry has obscured and corrupted scientific evidence. Would their assault have been slowed if financial conflicts had been disclosed? Perhaps.
Another reason is our growing belief that conflicts of interest are more damaging when they emerge later than if they are acknowledged earlier. As has been said of democracy, disclosure may be an imperfect solution, but the alternatives seem worse.
A third consideration is the aggressive stance taken by the major clinical journals. Lancet, JAMA, the New England Journal of Medicine and others threw down the gauntlet in simultaneous editorials2 published on September 12, 2001 (not, as it turned out, a good day for press coverage). Those editors introduced policies on conflict of interest that were subsequently incorporated into the “Uniform Requirements for Manuscripts Submitted to Biomedical Journals”3 (better known as the Vancouver Guidelines). These policies call for transparency on matters of funding, potential financial gain, and other matters of personal interest. Such standards apply not just to manuscript authors but to reviewers, letter-writers, and editors. Epidemiology subscribes to the Vancouver guidelines in other respects. Should we make an exception for conflict of interest?
We believe this topic needs to be aired among epidemiologists. Epidemiologists, like other researchers, are not necessarily motivated by curiosity alone. Conflicts of interest can affect many stages of data collection and analysis. Epidemiologic results, in turn, are broadly used. They have direct relevance to public health policy, and they enter into governmental and nongovernmental decision-making.
We have faith that the system of science is self-correcting in the long run (truth will stand and error will fall). In the meantime, the process of replicating epidemiologic results is tedious. Science may prevail in the end—but what about public health in the meantime? Is it constructive to withhold information about an investigator’s conflict of interest (or those of the critics) from people who use the evidence?
Perceived conflict of interest (which is subjective by definition) is also a matter of concern. Such conflicts are not necessarily a bad thing. Indeed, they are to some degree an inevitable part of epidemiologic research, and less damaging (especially when acknowledged). The question is how much to make explicit.
Starting with an assumption that it is better to acknowledge conflicts early than to discover them later, we present the following draft policy for Epidemiology. We emphasize the word DRAFT. This is not a declaration of policy, and it is not even a statement with which all the editors agree. It is rather the beginning of a conversation with the epidemiologic community.
Epidemiology proposes a policy on conflict of interest that would follow 2 general guidelines. First, we would require disclosure only of financial conflict of interest (real or apparent). Second, we would disclose this information routinely, but not use it as the basis for editorial decisions.
Specifically, we would ask our authors and reviewers (and ourselves) to provide relevant information on the following:
* Sources of research funding.
* Involvement of research funders in the conduct of the work or presentation of results.
* Financial arrangements with parties interested in the study findings; for example, through legal testimony based on study results.
* We would also encourage authors to inform us of any nonfinancial conflicts (real or apparent) that might be relevant.
We know the opinions among our readers must be at least as diverse as those among ourselves. We would like to hear your opinions. Most of you are aware of what other journals require (particularly the clinical journals). Are their policies succeeding, or do such policies constitute a “new McCarthyism”?4 Are the threats from conflicts of interest as serious as we think, or is the cure worse than the disease? Give us your opinions (email@example.com), and help us frame the journal’s next steps.
1. Stanton A, Glantz SA, Slade J, Bero LA, Hanauer P, Barnes DE, eds. The Cigarette Papers. Berkeley: University of California Press; 1998.
2. Davidoff F, DeAngelis CD, Drazen JM, et al. Sponsorship, authorship and accountability. New Engl J Med. 2001;345:825–827.
3. International Committee of Medical Journal Editors. Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Writing and Editing for Biomedical Publication. http://www.icmje.org/#conflicts
. Accessed February 1, 2005.
4. Rothman KJ. Conflict of interest. The new McCarthyism in science. JAMA. 1993;269:2782–2784.
© 2005 Lippincott Williams & Wilkins, Inc.