Mixed payment methods and health cost control in China: A machine learning analysis of policy big data

Abstract

Escalating health costs have incentivized many countries to adopt mixed payment systems. This study investigates how different prospective payment methods—including global budgets, capitation, per diem, single disease payment, diagnosis-related groups, and diagnosis intervention packet—interact to affect health costs in China. Using a novel health policy database, we applied supervised machine learning techniques to gauge prefectural efforts in implementing these payment methods. By matching these policy data with nationally representative survey data, we examined the interaction effects of global budgets and other payment methods on individual health spending and access. The results suggest that well-coordinated combinations of global budgets and other payment methods significantly decreased individual health spending and improved access, whereas poorly coordinated payment methods produced the opposite effects. This study introduces an innovative machine learning approach for analyzing geographic policy variations, offering a deeper understanding of the complex interplay among mixed payment methods.