Payment mechanism and medical quality: new evidence from China healthcare reform

Reform of payment methods promote to achieve the public value of health insurance through incentive-compatible mechanisms. Payment differentials, defined as the difference between DRG/DIP payment amounts and total medical expenses, together with the rational design of surplus retention mechanisms, are crucial for controlling healthcare costs, improving service quality, ensuring the sustainability of health insurance funds, and guiding medical institutions towards reasonable surplus targets.

This study constructs a one-shot game model under conditions of information asymmetry among hospitals, patients, and health insurance agencies. Using sample data from health insurance settlements in DRG pilot cities for the years 2021–2022, with readmission rates as the quality indicator and payment differentials as the measurement index, the study explores the association of payment differentials on the service quality of medical institutions.

Negative payment differentials are significantly associated with higher readmission rates and lower measured service quality. This may be consistent with insufficient payment incentives under the DRG system, which prompts medical institutions to adopt more aggressive cost-control measures, such as reducing hospital stays and limiting expenditures on medical services, which in turn may be associated with lower service quality. Moreover, the relationship between payment differentials and service quality varies depending on factors such as the type of surgery, hospital tier, and patient age.

There should be a more effective alignment between medical service pricing and health insurance payment standards, may help guide medical institutions toward more sustainable financial performance through the optimization of treatment processes, and promoting the shift towards a value-based payment system, potentially contributing to improved incentive alignment. Noting that as an observational study, our findings reflect correlations rather than causal effects.