The Unintended Consequences of Conflict of Interest Disclosure

Abstract

The authors of this article address a question that is fundamental to the integrity of scientific publications: Can readers (whether scientists/physicians or the public through media releases) be sure that information provided in any scientific publication is unbiased data free of any conflict of interest influence on the authors? Author disclosure of financial associations with commercial entities that have an interest in the research published in medical journals is a commonly accepted practice and is broadly perceived to have beneficial effects in prevention of bias. It is the contention of the authors; however, that the information provided under the presently used disclosure format does not help readers assess the degree of commercial influence over the data presented and may have the unintended consequence of contributing to and exacerbating bias. Examples from 2 experimental studies showed how disclosure can exacerbate bias. In scenarios presented in these studies, participants were divided into 2 roles: advisor (author/investigator) and estimator (scientist/physician or patient/reader). Advisors were considered to be relative experts and provided advice to less informed estimators. For some scenarios, advice was tainted by conflict, and in others, the advice was free of conflict. The examples presented showed that advisors with conflicts of interest gave more biased advice than those without conflicts of interest and that bias was substantially greater with disclosure. There are other problems with disclosure that threaten scientific integrity. In addition to disclosed conflicts, such as direct payments from pharmaceutical and device companies, some conflicts are not disclosed, such as financial arrangements that give physicians and their institutions incentives for providing high-cost services. Disclosure can be useful in protecting against bias despite its potential to be harmful only if the current policy for disclosing conflicts is considerably improved. The authors believe that with more comprehensive and uniform disclosure, the threat of having to clearly disclose problematic conflicts to patients and others would discourage physicians from entering into relationships that are difficult to justify. There is a rationalization among authors and the medical profession that with disclosure, the obligation to deal with conflicts of interest is adequately addressed. However, apparent transparency under the current disclosure policy cannot substitute for more substantive reform.