Study objectives: To understand the theory and results of how to improve physician performance, as part of overall health-care quality improvement. In particular, to study whether and how guideline production and dissemination affects physician performance.
Design: Review of meta-analyses and structured reviews; review of behavior change theories implicit in interventions to change physician performance.
Setting: Primarily the United States.
Patients or participants: Various patients and physicians, determined by reviews.
Interventions: None.
Measurements and results: There is no unifying theory of physician behavior change tested among physicians in practice. Attempts to affect individual physicians' performance have often met with failure. Mixed results are found for almost all interventions reviewed. Multiple interventions yield better results.
Conclusions: The answer to the question of what works to improve an individual physician's clinical performance is not simple. Emerging theory and evidence suggests that applications of behavior-change methods should not be focused on which tools (don't) always work. Instead, guideline development and implementation methods should be theory driven and evidence based (supported by evidence that proves the theory correct). In particular, the framework of evidence-based quality assessment offers some insight into past failures and offers hope for organizing attempts at guideline implementation.