Can case-based payment contain healthcare costs? - A curious case from China

We adopted a difference-in-difference (DID) design to evaluate the impact of a case-based payment pilot in Tianjin, China on hospital admission, utilization of varied therapeutic regimes, and the associated costs. We used claim data of all admissions of angina and acute myocardial infarction during July 2015 to June 2018, 18 months before and after the program. Our analyses were supported by convincing common trends tests and a couple of sensitivity analyses. As intended, for patients who received percutaneous coronary stenting (PCS) and were counted in the case-based payment system, we showed that the program decreased length-of-stay, per-admission spending, and out-of-pocket spending by 20.8%, 14.2%, and 95.5%, respectively, but did not increase readmissions. However, when considering all patients who suffered from the two types of coronary heart diseases, we found that the program otherwise increased per-admission spending by nearly 11%. As a result, the program took a perverse effect in increasing monthly spending for the health insurance scheme and the society by 1005.6 thousand USD (47·5%) and 1095·7 thousand USD (34·7%), respectively. Increases in hospital admissions, and proportion of performing PCS accounted for 66·7% and 39·2% of the rise, respectively. In addition, our analysis provided evidence of health providers' cream-skimming behaviors, including selecting younger patients with lower CCI in the case-based system, up-coding complications, and keeping higher cost patients in the fee-for-service payment system. We draw lessons that case-based payment may make an unintended impact that increases healthcare costs when incentives are not properly designed.