Preventable adverse events affect 13.5% of Medicare hospitalizations and cost more than $17 billion each year; similarly, unplanned hospital readmissions affect 20% of Medicare discharges and cost $15 billion annually. To address these problems, Hospital Value-Based Purchasing (HVBP) and the Hospital Readmissions Reduction Program (HRRP) were launched under the Patient Protection and Affordable Care Act (ACA) in 2012. These pay-for-performance (P4P) policies authorize the Centers for Medicare and Medicaid Services (CMS) to make progressive reductions in Medicare reimbursements based on certain quality measures. Under historical Medicare payment reforms, decreased reimbursements have been associated with operational and staffing reductions, cutbacks in quality initiatives, and an increase in patients discharged in unstable conditions. The potential for such unintended consequences may be heightened when hospitals face increased financial pressure from multiple reimbursement policies simultaneously. Emerging literature further suggests teaching and safety-net hospitals are more likely to be penalized under HVBP and/or HRRP than their counterparts. While these policies have different risk-adjustment methodologies, they both disregard potentially important community characteristics, such as the quality of nursing homes, which can influence the quality of health care delivery. It remains unclear, therefore, if the hospitals that seem to be disproportionately penalized are truly low-performing, or if these P4P policies are improperly calibrated. The proposed research examines the role of community influences, such as the quality of nursing homes, on hospital penalties and the intended and unintended ways hospitals may respond to these P4P incentives. The analyses will use multiple years of hospital data from CMS, Hospital Compare, and the Dartmouth Atlas of Health, in conjunction with patient-level Medicare claims. Specifically, the plan has three aims: 1) quantify the degree to which hospital and community factors contribute to differential penalties through a multi-level Fairlie decomposition, 2) evaluae the effect of HVBP and HRRP on hospital performance and operational measures, and 3) investigate the possibility of unintended consequences, such as assigning readmitted Medicare patients to observation status to avoid being penalized. The last two aims exploit how the penalty factors published by CMS have a discrete sorting threshold in a regression discontinuity design (RDD). The use of quasi- experimental methods to address the potentially endogenous relationship between Medicare revenue and hospital performance will provide critical early feedback and can inform future P4P policies, particularly in the aging population. The proposed research and training plan is a logical progression of the PI's emerging research in the methodological and measurement issues related to patient safety. She will leverage this proposal to expand her methods skill set and develop into an independent investigator, with emphasis on exploring practical questions that result in actionable information to guide policy-makers.
Recently launched policies require Medicare reimbursement rates to reflect provider performance on certain quality metrics, such as adherence to evidence-based processes of care or unplanned 30-day readmissions. The proposed research will examine which hospital and community factors are associated with being penalized and investigate potential ways hospitals may respond to decreased Medicare reimbursements. Results from this study will provide critical early feedback on how these policies may affect the well-being of the aging population and the efficiency of our health care system.
|Aswani, Monica S; Kilgore, Meredith L; Becker, David J et al. (2018) Differential Impact of Hospital and Community Factors on Medicare Readmission Penalties. Health Serv Res 53:4416-4436|