24/03/2026 - Doyen 22 - 12:45
(PSE)
will give a presentation on
Abstract
This study evaluates the causal impact of a pay-for-performance program on diabetes monitoring quality in primary care. Introduced in 2014 in the Province of Verona (Italy), the program offered general practitioners capitation-based bonuses for improving the completeness of electronic health records for diabetic patients aged 65 and above. Using administrative data on approximately 200 GPs across three Local Health Units over 2010–2017, we employ two complementary identification strategies: a staggered event study exploiting the temporal rollout of the policy and a difference-in-discontinuities design leveraging the age-65 eligibility threshold.
We address three research questions using five outcomes. First, we examine whether the incentive improved diabetes monitoring on the extensive margin (share of patients receiving at least one annual HbA1c test) and the intensive margin (average number of HbA1c tests per patient). Second, we assess its effect on diagnostic documentation quality, measured as the share of diabetic patients with a formal diagnosis recorded in the GP's clinical records. Third, we investigate potential spillover effects on broader healthcare utilisation, captured by average GP consultations and antidiabetic prescriptions (ATC code A10) per patient.
Both approaches yield consistent results. The incentive produced immediate and significant increases in HbA1c monitoring, raising the share of patients tested by approximately 13–14 percentage points and increasing test frequency. These effects concentrate in the first two post-treatment years and emerge precisely at the age-65 threshold. Importantly, we find no significant effects on GP consultations, diagnostic recording, or prescription intensity, indicating improved adherence to monitoring protocols without distortions to broader care patterns. Gender-stratified analyses reveal similar responses for male and female patients. Robustness checks—including donut-hole designs, spillover tests, and placebo cutoff analyses—support the identifying assumptions and suggest that mild pre-trends in some specifications reflect beneficial within-physician spillovers rather than anticipatory behavior. Our findings demonstrate that even modest administrative incentives tied to electronic record quality can meaningfully improve chronic disease monitoring in publicly contracted primary care systems.
More information about the speaker