Evidence to action. Dealing with conflict of interest in moving from systematic reviews to guidelines

Article type
Authors
Guyatt G, Akl E1, Kearon C2, Crowther M2, Nathanson I3, Zelman-Lewis S4, Schunemann H2
1University of Buffalo, Buffalo, New York, United States
2McMaster University, Hamilton, Ontario, Canada
3University of Central Florida College of Medicine, Orlando, Florida, United States
4Health & Science Policy, American College of Chest Physicians, Northbrook, Illinois, United States
Abstract
Background: The Cochrane Handbook specifies that Cochrane reviews should not make recommendations. Ultimately, however, the impact of Cochrane reviews depends on their use in guiding clinical practice. Incorporation in clinical practice guidelines enhances the impact of Cochrane reviews. Conflict of interest can corrupt the process of appropriate incorporation of review findings into guidelines. The American College of Chest Physicians guidelines for anti-thrombotic therapy represent a prominent guideline facing the challenges of developing recommendations on the basis of systematic reviews. Methods: We conducted an iterative process to develop an approach to guideline development that included input from a wide variety of individuals involved in the anti-thrombotic guidelines to develop a process designed to achieve two goals: 1) Result in recommendations free of conflict of interest. 2) Allow the full benefit of input from clinical experts. Results: Primary responsibility for each chapter rests with a methodologist without important conflicts of interest. A committee of academic physicians reviews a potential panel member’s financial conflicts and decides if they are acceptable, unacceptable, or acceptable provided future industry involvement is restricted. Experts who are approved during this review but are judged to have important financial or intellectual conflict of interest can participate in collecting and interpreting evidence. Only panel members without important conflicts can, however, participate in the development of recommendations, a process from which conflicted participants are excluded. Conclusion: This process is innovative in that it puts far more emphasis on intellectual conflict of interest than previous approaches, that it places final responsibility for the guidelines with an unconflicted methodologist, and that it represents a strategy for comprehensive input from experts while – hopefully – keeping recommendation uninfluenced by conflicts. These approaches may facilitate optimal use of Cochrane and other systematic reviews in guiding clinical practice.