Abstract
BACKGROUND: Physician networks (PNs) are a recent development in Germany, designed to improve the coordination and quality of healthcare. We compared the performance of PNs that use a standardized system of audit and feedback to that of regular care. METHODS: We analysed a large sample of claims data from Bavaria, Germany, using nearest-neighbour propensity score matching. Patients who had ambulatory care-sensitive conditions (ACSCs) and were enrolled in PNs were matched with control patients receiving regular care. We examined potentially avoidable hospitalizations related to the 13 most common ACSCs (primary endpoints), as well as processes-of-care indicators for disease prevention, pharmacotherapy and coordination of care. RESULTS: There were no significant differences in rates of potentially avoidable hospitalizations between the two groups. However, the networks showed higher vaccination rates, increased participation in disease management programmes, and more frequent use of referrals when consulting specialist physicians. On average, network patients visited a greater number of specialists and had lower continuity of care compared to patients receiving regular care. Polypharmacy and PRISCUS-list prescriptions were more prevalent in the networks. CONCLUSIONS: PNs using audit and feedback do not appear to perform better than regular care in preventing hospitalizations due to ACSCs. However, they do perform better in disease prevention measures while showing inconclusive results for care coordination and pharmacotherapy. Further research is needed to understand effective collaboration among providers and its impact on the quality of care within PNs.