LAUSR

Background In April 2009, the Chinese government launched Zero Markup Drug Policy (ZMDP) to adjust medical institutions' revenue and expenditure structures. Objective This study evaluated the impact of implementing ZMDP (as an intervention) on the drug costs for managing Parkinson's disease (PD) and its complications from the healthcare providers' perspective. Methods The drug costs for managing PD and its complications per outpatient visit or inpatient stay were estimated using electronic health data from a tertiary hospital in China from January 2016 to August 2018. An interrupted time series analysis was conducted to evaluate the immediate change following the intervention (step change, β1) and the change in slope, comparing post-intervention with the pre-intervention period (trend change, β2). Subgroup analyses were conducted in outpatients within the strata of age, patients with or without health insurance, and whether drugs were listed in the national Essential Medicine List (EML). Results Overall, 18,158 outpatient visits and 366 inpatient stays were included. Outpatient (β1 = −201.7, 95%CI: −285.4, −117.9) and inpatient (β1 = −372.1, 95% CI: −643.6, −100.6) drug costs for managing PD significantly decreased when implementing ZMDP. However, for outpatients without health insurance, the trend change in drug costs for managing PD (β2 = 16.8, 95% CI: 8.0, 25.6) or PD complications (β2 = 12.6, 95% CI: 5.5, 19.7) significantly increased. Trend changes in outpatient drug costs for managing PD differed when stratifying drugs listed in EML (β2 = −1.4, 95% CI: −2.6, −0.2) or not (β2 = 6.3, 95%CI: 2.0, 10.7). Trend changes of outpatient drug costs for managing PD complications significantly increased in drugs listed in EML (β2 = 14.7, 95% CI 9.2, 20.3), patients without health insurance (β2 = 12.6, 95% CI 5.5, 19.7), and age under 65 (β2 = 24.3, 95% CI 17.3, 31.4). Conclusions Drug costs for managing PD and its complications significantly decreased when implementing ZMDP. However, the trend in drug costs increased significantly in several subgroups, which may offset the decrease at the implementation.