Association for Public Policy Analysis & Management

In an effort to reduce costly and largely preventable hospital readmissions for Medicare enrollees, the Affordable Care Act (ACA) of 2010 included a provision known as the Hospital Readmissions Reduction Program (HRRP). Initially, the HRRP included financial penalties for hospitals that exhibit a relatively high rate of 30-day Medicare readmissions for three specific conditions: acute myocardial infarction (i. e. heart attack), heart failure, or pneumonia. Beginning in fiscal year 2013, hospitals that failed to meet the readmission goals established by the Centers for Medicare and Medicaid Services (CMS) received up to a 1% reduction in total payments made under the Medicare Inpatient Prospective Payment System. This penalty escalated to 2% in the second year of the program and to 3% in the third year and beyond. By fiscal year 2015, 2,610 hospitals, or approximately half of all U. S. hospitals, were set to receive a readmission penalty under the HRRP.

While the explicit goal of the HRRP was to prevent readmissions among the Medicare population, it is conceivable that hospital efforts to reduce readmissions for Medicare patients suffering from any of the three targeted conditions may also affect the privately insured or Medicare patients hospitalized for a non-targeted diagnosis. These "spillover effects" depend in part on the hospital's strategy to avoid readmission penalties and have potentially significant implications for expenditures on hospital care.

In this paper, we examine the effect of the HRRP on changes in readmission rates for both targeted and non-targeted Medicare beneficiaries as well as for the privately insured. Our empirical methodology improves on earlier efforts to identify the effects of the HRRP that rely on problematic control groups that may themselves be affected by the program. Initially, we document changes in readmission rates to targeted Medicare beneficiaries in order to assess the program's impact on its intended population. Next, we consider changes in readmission rates for the privately insured and non-targeted Medicare patients in order to establish the presence of spillover effects associated with the HRRP. Finally, we expand our analysis to identify differential responses to the HRRP by hospital market share and by ownership status.