Impact of diagnosis-related group-based bundled payment on hospitalization costs and clinical outcomes in total hip arthroplasty: an interrupted time series analysis from Wenzhou, China.

BACKGROUND: China's nationwide rollout of diagnosis-related group (DRG) bundled payment, implemented from January 2021, is one of the most consequential structural reforms to hospital reimbursement in recent decades. Total hip arthroplasty (THA)-a high-volume, implant-intensive elective procedure with marked inter-institutional cost variation-provides an analytically tractable context in which to evaluate whether fixed-payment incentives alter provider resource utilization while preserving clinical quality. METHODS: This retrospective longitudinal study included 2,621 consecutive THA admissions at a tertiary teaching hospital in Wenzhou, Zhejiang Province, between January 2015 and April 2026 (pre-DRG: n = 1,041; post-DRG: n = 1,580). Segmented linear regression with Newey-West heteroscedasticity- and autocorrelation-consistent (HAC) standard errors was applied to 136 monthly aggregated observations (72 pre-intervention; 64 post-intervention). Primary outcomes were monthly mean total hospitalization cost, drug cost, and orthopedic implantable device cost. A composite adverse discharge outcome rate (in-hospital mortality, failure to achieve clinical improvement, and unclassified discharge) was the clinical quality endpoint. Four pre-specified sensitivity analyses addressed consumer price index (CPI) inflation adjustment, COVID-19-related confounding (exclusion of 2020), Prais-Winsten first-order autoregressive [AR(1)] correction, and case-mix adjustment for monthly patient age and sex composition. RESULTS: Mean total hospitalization cost declined from CNY 53,476 ± 14,898 to CNY 37,475 ± 15,273 (-29.9%; p < 0.001). Orthopedic implantable device cost fell from CNY 35,826 ± 9,607 to CNY 17,727 ± 10,097 (-50.5%; p < 0.001), and drug cost from CNY 6,685 ± 5,058 to CNY 5,741 ± 3,311 (-14.1%; p < 0.001). Interrupted time series (ITS) modeling identified strong pre-existing downward trends in both total (β1 = -89.79 CNY/month; p = 0.003) and device costs (β1 = -104.81 CNY/month; p = 0.001) prior to DRG implementation. Post-DRG slope changes for total and device costs did not reach statistical significance (p = 0.071 and p = 0.061, respectively). Drug cost exhibited a significant immediate level increase (β2 = 3,006 CNY; p = 0.003) followed by a statistically significant sustained monthly decline (β3 = -27.80 CNY/month; p = 0.045). The composite adverse discharge rate did not differ significantly between periods (1.06% vs. 0.63%; χ2 = 0.944; p = 0.334). Sensitivity analysis excluding 2020 yielded a statistically significant post-DRG slope change for total cost (β3 = -146.92 CNY/month; p = 0.047), suggesting that COVID-19-related disruptions partially attenuated this effect in the primary analysis. CONCLUSIONS: Following DRG implementation, total and orthopedic implantable device costs continued their pre-existing downward trajectories, with observed expenditure sustained well below the pre-intervention counterfactual. The primary ITS analysis did not identify a statistically significant DRG-specific acceleration of this decline for total or device costs, a finding that predominantly reflects the dominant contribution of antecedent volume-based procurement reform. Drug cost exhibited the most clearly DRG-attributable effect-an initial transitional surge followed by sustained statistically significant decline-compatible with institutional adaptation after implementation of fixed-payment incentives. No evidence of in-hospital deterioration in adverse discharge outcomes was detected; this underpowered composite endpoint did not capture post-discharge events such as 30-day readmission or surgical-site infection.