Penalty and incentive modeling for hospital readmission reduction

Abstract

Nearly 20% of patients are readmitted to hospitals within a specific time period after hospital discharge. High readmission rates place an unnecessary burden on the healthcare system, and new initiatives to reduce preventable hospital readmissions have been established. The United States Hospital Readmission Reduction Program (HRRP) is an example of a health policy reform that links insurance payments to quality of care. Critics of HRRP believe that its punitive mechanism design provides less money to struggling hospitals and, in some cases, fails to provide proper incentives and resources for quality care improvements. An asymmetric penalty-incentive model for hospital readmission reductions was developed and studied for an insurer-led reimbursement system. We formulate a game-theoretic setting involving an insurer and a hospital. We derive the insurer’s optimal policy design and the hospital’s best response in an insurer-led Stackelberg setting with rational agents. The model was analyzed for centralized and decentralized solutions and compared to the do-nothing solution. Most notably, we found that a positive incentive level is necessary for a win-win region to exist. An example from public hospital data for acute myocardial infarction showed that transitioning from the current 3% penalty-only policy to the optimal 5.47% incentive-only policy would result in only a 0.17% increase in insurer costs while inspiring hospitals to maximize level of care and increase hospital profits by 39.7%.