Cost shifting or cost cutting by hospitals as a response to reimbursement reform? The case of diagnosis-related groups (DRG) scheme in China

Background Diagnosis-related group (DRG) systems for healthcare reimbursement were recently introduced among hospitals in China, raising concerns about cost-shifting, where hospitals may increase charges for self-financing patients to offset reimbursement cuts by DRG. In 2018, both Nanchang and Ganzhou Cities in Jiangxi Province installed DRG information systems, but only Nanchang fully implemented the DRG system during the 2019–2020 pilot period. Materials and methods Drawing from a healthcare administrative dataset of 14,310 patients' records, this study investigates the hospitalization costs associated with Intraocular Lens (IOL) implantation procedures in Jiangxi Province, China, from 2017 to 2020. By applying the quantile difference-in-differences (DID) and difference-in-difference-in-difference (DDD) methodologies, the research examines the impacts of DRG implementation on hospitalization costs, with a particular focus on self-financing patients. Results Upon the implementation of DRG in hospitals in Nanchang, the IOL cost ratio for patients in Nanchang was lower on average by 0.047 ( p < 0.001 ), with the largest reduction observed at higher quantiles (e.g., Q75: −0.105, Q90: −0.091) ( p < 0.001 ). The DDD model revealed a decrease of 0.226 in the IOL cost ratio for self-financing patients in DRG hospitals post-implementation. These results remained robust across various healthcare cost ratios. Conclusion The study indicates that implementing the DRG system in Chinese hospitals effectively constrained healthcare costs without causing unintended cost-shifting among self-financing patients. The system's heightened impacts on higher quantiles indicate its efficacy in addressing high-cost outliers. These outcomes are likely attributed to the centralized medical procurement and healthcare information technology integrated into the Chinese healthcare system.

Abstract licence: CC BY

69 more references, whose wording the paper’s licence keeps private.

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.