The diluting effect of medical work groups on feedback efficacy in changing physician's practice.

Feedback intervention has been advocated as a successful method to modify the way that physicians practice medicine. However, most studies concerned with modifying physician profiles have focused on interns and residents. The results presented here concern regular staff and therefore provide a better basis for generalization. Over a 2.5-year period, we analyzed the use of clinical resources by physicians practicing in four medical specialties in two hospitals. Hierarchical multiple regression models were used to control case mix in order to identify the specific effects attributable to feedback. The information failed to modify the physicians' practice profiles. Our results suggest that this failure is related to the organization of inpatient medical practice as a group effort. This mode of practice has a negative effect on feedback efficacy. First, it weakens one of the main feedback mechanisms, that is personal identification with the data. Second, it probably generates mutual adjustment among physicians, thus eliminating practice variations at the source.