Does vertical integration increase the costs for primary care inpatients? Evidence from a national pilot programme in...

To assess the impact of vertical integration (VI) within County-Level Integrated Health Organisations (CIHOs) on the costs of primary care inpatients.

This study assessed Xishui, a national pilot county for CIHOs, using inpatient claims data. The treatment group comprised 10,118 inpatients from 5 vertically integrated township health centres (THCs), while the control group consisted of 21,165 inpatients from 19 non-vertically integrated THCs. The periods from July 2020 to December 2021 and January 2022 to December 2022 were defined as pre- and post-policy intervention, respectively. The primary outcome variables were total health expenditures (THS), out-of-pocket (OOP) expenditures, and the proportion of OOP expenditures. Propensity score matching was employed to align inpatient demographics and disease characteristics between the groups, followed by a difference-in-differences analysis to evaluate the outcomes.

VI significantly increased THS (β = 0.1337, p < 0.01) and OOP expenditures per case (β = 0.1661, p < 0.001), but the increase in the proportion of OOP expenditures per case was not significant (β = 0.0029, p > 0.05). For the basic medical insurance for urban and rural residents, THS per case (β = 0.1343, p < 0.01) and OOP expenditures (β = 0.1714, p < 0.001) significantly increased. For the basic medical insurance system for employees, THS per case also increased significantly (β = 0.1238, p < 0.01), but the change in OOP expenditure proportion per case was not significant (β = 0.1020, p > 0.05). The THS per case led by Xishui County People's Hospital, the leading county medical sub-centre (CMSC), significantly increased (β = 0.1753, p < 0.01), whereas the increase led by Xishui County Traditional Chinese Medicine Hospital was not significant (β = 0.0742, p > 0.05). Increases in OOP expenditures per case were significant in CMSCs led by the People's Hospital and the Traditional Chinese Medicine Hospital (β = 0.1782, p < 0.01 and β = 0.0757, p < 0.05, respectively).

VI significantly increased THS and OOP expenditures for primary care inpatients. However, VI could exacerbate economic disparities in disease burden across different insurance categories.