Despite widespread enthusiasm about the potential impact of new investments in comparative effectiveness research, recent history suggests that scientific evidence may be slow to change clinical practice. Reflecting on studies conducted over the past decade, we identify five causes that underlie the failure of many comparative effectiveness studies to alter patient care. These are financial incentives, such as fee-for-service payment, that may militate against the adoption of new clinical practices; ambiguity of study results that hamper decision making; cognitive biases in the interpretation of new information; failure of the research to address the needs of end users; and limited use of decision support by patients and clinicians. Policies that encourage the development of consensus objectives, methods, and evidentiary standards before studies get under way and that provide strong incentives for patients and providers to use resources efficiently may help overcome at least some of these barriers and enable comparative effectiveness results to alter medical practice more quickly.