How provider payment methods affect health expenditure of depressive patients? Empirical study from national claims data in China from 2013 to 2017

Background: This study aimed to investigate the associations between provider payment methods and expenditure of depressive patients, stratified by service types and hospital levels. Methods: We used a 5 % random sample of urban claims data in China (2013-2017), collected by China Health Insurance Research Association. Provider payment methods (fee-for-services, global budget, capitation, casebased and per-diem payments) were the explanatory variables. A generalized linear model was fitted for the associations between provider payment methods and expenditure. All analyses were adjusted for patient"cioeconomic and health-related characteristics. Results: In total, 64,615 depressive patient visits were included, 59,459 for outpatients and 5156 for inpatients. Female patients accounted for 63.00 %. The total and out-of-pocket (OOP) expenditure significantly differentiated by provider payments. Among outpatient services, when comparing with fee-for-services, capitation payment was associated with substantial marginal reduction in total and OOP expenditure (-$34.18, -$9.71) in primary institutes, yet increases ($27.26, $24.11) in secondary hospitals. Similarly, global budget was associated with lower total and OOP expenditure (-$13.51, -$1.61) in secondary hospitals, while higher total and OOP expenditure ($7.43, $32.27) in tertiary hospitals than fee-for-services. For inpatients, total and OOP expenditures under per-diem (-$857.65, -$283.48) and case-based payments (-$997.93, -$137.56) were remarkably smaller than those under fee-for-services in primary and secondary hospitals, respectively. Besides, case-base payment was only linked with the largest reduction in OOP expense (-$239.39) in inpatient services of tertiary hospitals. Limitation: Only urban claims data was included in this study, and investigations for rural population still warrant. And updated data are needed for future studies. Conclusions: There were varying correlations between provider payment methods and expenditure, which differed by service types and hospital levels. These findings provided empirical evidence for optimizing the mixed payment methods for depression in China.