What do we really know about conflicts of interest in biomedical research?

Rationale. Conflicts of interest in research have long been seen to pose serious threats to the integrity of research. Indeed, the past two decades have revealed increased attention to the possible influence of conflicts of interest as the number of research studies funded by private industry have increased. Objectives. We review empirical findings directly pertinent to issues concerning financial conflicts of interest in biomedical research, and we provide an overview of the current issues and state of understanding concerning such conflicts in research. Methods. We searched Medline from January 1992 to January 2002 to locate current studies that have reported primary or secondary empirical data pertaining to conflicts of interest in research. Results. The rate of potential conflicts of interest for researchers appears to be at least 30% in some situations although the accuracy of such a rate is untested, and the rate of disclosure of conflicts of interest is as low as 2%. Furthermore, some evidence exists to indicate that researchers with conflicting interests may indeed offer different professional opinions and judgments than those for whom such conflicts do not exist. The effectiveness of various disclosure and management methods for conflicts of interest is unknown. Conclusions. We summarize the current empirical literature, concluding that relatively little is known based on such data. We suggest a number of questions that need to be answered by future research, and we offer recommendations for policy makers to consider and evaluate in the future in identifying, revealing, and managing conflicts of interest.