Effects of Pay-for-Performance for Primary Care Physicians on Preventable Diabetes-Related Hospitalization Costs among Adults in New Brunswick, Canada: A Quasiexperimental Evaluation.

OBJECTIVES: In New Brunswick, Canada, 13.6% of the population 35 years of age and older is living with type 1 or type 2 diabetes mellitus. To address public health and clinical challenges, pay-for-performance (P4P) for family physicians was introduced in 2010 to enable comprehensive diabetes management. This study assesses the impacts of the P4P scheme on excess health-care costs.

METHODS: We used a quasiexperimental study design drawing on linked population-based administrative data sets of physician billings, hospital discharge abstracts and provider and resident registrations. Prospective cohorts of patients with diabetes were identified through a validated algorithm tracing individuals' interactions with the health-care system. We applied propensity-score difference-in-differences estimation for the effects of P4P on preventable diabetes-related hospitalization costs according to patients' exposures to physicians' uptake of the incentive.

RESULTS: Coverage of incentivized care peaked at less than half (44%) of adults with diabetes, who tended to be younger and less often presenting comorbid conditions compared to those whose providers did not claim incentives. The introduction of P4P was attributed to significantly lower diabetes hospitalization costs among newly diagnosed patients (-0.083; p<0.01) and improved compensation for physicians. No cost avoidance was established among medium-term and longer-term patients or for hospitalizations for conditions concordant with diabetes.

CONCLUSIONS: The effects of New Brunswick's P4P for diabetes care are mixed. Results reflect the deficient evidence base on the effects of P4P on patient-oriented and policymaker-important health outcomes. The high risk for multiple morbidities among patients with diabetes and the heterogeneity of physician responses to performance incentives may be hindering the effectiveness of P4P in improving diabetes outcomes.