Changing physicians' behavior: What works and thoughts on getting more things to work

Health services research consistently demonstrates a gap between research-based best clinical practice and what doctors actually do. Traditionally, the profession of medicine has behaved as if dissemination of research findings in peer-reviewed journals will eliminate this gap, even though professionals typically have less than 1 hour per week to read. This problem is complicated by the fact that physicians have not been trained generally to appraise published research, which is of variable quality in any event. Physicians interested in changing their practices also encounter organizational, peer group, and individual barriers at the same time as they face information overload and patient expectations. In a word, physicians' abilities to manage information is overwhelmed. This article both summarizes initiatives to improve physicians' information management through efforts to synthesize available evidence and describes the current evidence base of effectiveness and efficiency of dissemination and implementation strategies. We conclude that there is an imperfect evidence base to support decisions regarding strategies that are likely to be appropriate and effective under varying circumstances. Since this problem is compounded by the lack of a theoretical base for conceptualizing physician behavior change, we suggest exploring the applicability of behavioral theories to the understanding of professional behavior change. We also suggest exploring the use of theory-based process evaluations alongside randomized trials of dissemination and implementation strategies to further test theories and to explore causal mechanisms. Further research is required to explore determinants of provider behavior to better identify modifiable and non-modifiable effect modifiers, to develop methods of identifying barriers and facilitators to change, and to estimate the efficiency of dissemination and implementation strategies in the presence of different barriers and effect modifiers.