Expanding capitation accountability and healthcare utilization in an integrated care network: evidence from China

Capitation may alter provider incentives, but its effects depend on which services and expenditures are included within the accountability boundary. We examined how outpatient-only and total-care capitation affected healthcare use inside and outside an integrated care network in an open-access setting. We used claims data for continuously enrolled family doctor contract service residents in Binhai New Area, China, from December 2022 to June 2025. The study period covered three phases: pre-policy, outpatient-only capitation, and total-care capitation. Primary outcomes were monthly per-capita outpatient visits, inpatient days, outpatient expenditure, and inpatient expenditure, measured separately for in-network and out-of-network care. We applied aggregate-level segmented interrupted time-series models with quarterly indicators and autoregressive errors. Supplementary two-part GEE models examined whether changes were driven by service-use probability or conditional intensity among users. Following outpatient-only capitation, the increasing trends in out-of-network outpatient visits and expenditure shifted directionally downward (trend changes: −0.003 visits per enrollee, 95% CI − 0.007 to 0.001, p > 0.10; and − 1.02 CNY per enrollee, 95% CI − 2.44 to 0.40, p > 0.10). Supplementary analyses associated this pattern mainly with a declining probability of out-of-network outpatient use (trend change: 0.954, 95% CI 0.946–0.962, p < 0.001), while no increase in inpatient utilization or expenditure was detected. After inpatient expenditure was included in the capitated budget, out-of-network inpatient days showed an immediate negative level change (− 0.004, 95% CI − 0.014 to 0.005, p > 0.10) and a significant declining trend (− 0.004 per month, 95% CI − 0.006 to − 0.002, p < 0.01), mainly associated with reduced admission probability (level change: 0.793, 95% CI 0.641–0.981, p = 0.033). Inpatient expenditure also declined directionally, whereas outpatient visit and expenditure trends shifted upward; these expenditure and outpatient changes were not statistically significant. The observed patterns were consistent with a boundary-sensitive association between capitation accountability and healthcare utilization. The clearest association was observed for out-of-network inpatient utilization after inpatient expenditure was incorporated into the accountable budget. We did not detect statistically significant compensatory increases across care settings, although the opposing outpatient and inpatient point estimates and smaller spillover effects warrant cautious interpretation.