D ATA D R I V E N . P O L I C Y F O C U S E D LDI ResearchBRIEF Research to Improve the Nation’s Health System 2018 . No. 16 CHANGES TO RACIAL DISPARITIES IN READMISSION RATES AFTER MEDICARE’S HOSPITAL READMISSIONS REDUCTION PROGRAM WITHIN SAFETY-NET AND NON-SAFETY-NET HOSPITALS JAMA Network Open – published November 2, 2018 Krisda H. Chaiyachati, Mingyu Qi, Rachel M. Werner KEYFINDINGS After the Medicare Hospital Readmissions Reduction Program began enforcing financial penalties, disparities in readmissions between white and black patients widened at safety-net hospitals for conditions not targeted by the program. Disparities were stable for conditions targeted by the program. At non-safety-net hospitals, disparities were unchanged for both targeted and non-targeted conditions. THE QUESTION THE FINDINGS Medicare’s Hospital Readmissions Reduction Program (HRRP) financially During the implementation phase of the HRRP, disparities in readmission penalizes hospitals with higher-than-expected 30-day readmission rates rates were improving within both safety-net and non-safety-net hospitals, for select clinical conditions (acute myocardial infarction, pneumonia, irrespective of the clinical condition. After the enforcement of the HRRP and heart failure). Since its implementation, readmission rates have penalties, disparities in readmission rates widened within safety-net declined for the HRRP-targeted conditions, and to a lesser degree, for hospitals, specifically for conditions not targeted by HRRP penalties. non-targeted conditions. There are, however, concerns that the program could exacerbate health care disparities by harming “safety-net hospitals” Comparing safety-net and non-safety-net hospitals – institutions that disproportionately serve minority patients. Before the ACA (2007-2010), white and black patients discharged from safety-net hospitals had relatively higher risk-adjusted readmission rates This analysis looks across three time periods: 2007-2010, prior to the than those from non-safety-net hospitals. The racial disparity remained Affordable Care Act (ACA); 2010-2012, during HRRP implementation; stable during this time, in both safety-net and non-safety net hospitals. and, 2012-2015, after the HRRP enforced penalties. It evaluates how readmission disparities between black and white patients changed within Overall, readmission rates for black and white patients across safety-net safety-net and non-safety-net hospitals after the HRRP began enforcing and non-safety-net hospitals declined during HRRP’s implementation financial penalties, and how these trends differed for conditions targeted (2010-2012). In safety-net hospitals, black patients had a steeper decline and not targeted by HRRP penalties. in readmission rates than white patients, resulting in a slight narrowing COLONIAL PENN CENTER | 3641 LOCUST WALK | PHILADELPHIA, PA 19104-6218 | LDI.UPENN.EDU | P: 215-898-5611 | F: 215-898-0229 | @PENNLDI ResearchBRIEF LDI of disparities in readmission rates. Non-safety-net hospitals also saw a narrowing of disparities in readmission rates during this time, but at an even smaller magnitude. THE FINDINGS SUGGEST THAT WHEN Readmissions disparities stopped improving during the HRRP penalty period (2012-2015). During this time, readmission disparities widened in READMISSION PENALTIES TARGET CERTAIN safety-net hospitals (0.04 percentage point per quarter), whereas they CLINICAL CONDITIONS, SCARCE RESOURCES remained stable in non-safety-net hospitals. COULD SHIFT AWAY FROM OTHER Changes in readmission disparities for targeted and non-targeted conditions OPERATIONAL OBJECTIVES AND AFFECT During the pre-ACA and HRRP implementation periods, the observed QUALITY FOR NON-TARGETED CONDITIONS. trends in black and white readmissions did not differ for HRRP-targeted conditions and non-targeted conditions in either safety-net or non- safety-net hospitals. condition. However, these programs may impose higher penalties However, when the authors compared readmissions by condition on more hospitals, penalties to which safety-net hospitals are already during the HRRP penalty period, readmission disparities widened (0.05 particularly sensitive. percentage point per quarter) within safety-net hospitals for conditions not targeted by the HRRP, while disparities for HRRP-targeted conditions did not change. Readmission disparities were stable for both targeted and THE STUDY non-targeted conditions in non-safety-net hospitals during this period. This study looked at nearly 60 million discharges of Medicare fee-for- service beneficiaries with any clinical condition from 3,871 acute-care THE IMPLICATIONS hospitals between January 1, 2007 and September 30, 2015. About one fifth of discharges were from safety-net hospitals, defined as the those This study shows that hospital-based financial incentives may exacerbate with the highest proportion of Medicaid discharges in each state. health care disparities. It highlights how financial penalties for hospital readmissions have potential consequences for conditions not targeted The analysis relies on three data sources: (1) Medicare Provider Analysis by the penalties, and that safety-net hospitals are particularly sensitive and Review files and the Medicare Beneficiary Summary File, which to this risk. The findings suggest that when readmission penalties target include data on demographics and enrollment, and hospital claims for certain clinical conditions, scarce resources could shift away from other fee-for-service Medicare beneficiaries; (2) Medicare Provider of Services operational objectives and affect quality for non-targeted conditions. files, which include hospital characteristics; (3) the American Hospital Association Annual Survey of Hospitals, which includes data on Medicaid Although the increases in racial disparities in hospital readmission rates discharges at each hospital. All estimates were adjusted for patient and observed in this study are relatively modest, conditions not targeted by hospital characteristics. the HRRP represent six times more discharges than targeted conditions (86.5 percent vs. 13.5 percent). This means that even small increases in The authors were primarily interested in whether a patient had an disparities could have consequences for a larger portion of the population. unplanned readmission within 30 days of hospital discharge. They assessed the patient’s race (white or black), the quarter in which the The HRRP has reduced Medicare payments to safety-net hospitals by patient was discharged, the clinical condition, and whether or not the one to three percent. If payments are reduced, safety-net hospitals have discharging hospital was a safety-net hospital. They measured changes less revenue to invest in quality-improvement programs, thus potentially in disparities in safety-net and non–safety-net hospitals after the HRRP widening existing disparities. These findings point to the HRRP’s potential penalties were enforced and compared with prior trends. longer-term consequences on readmission disparities between white and black patients, especially for those served by safety-net hospitals. Reducing disparities in readmission rates will require an understanding of Chaiyachati KH, Qi M, Werner RM. Changes to Racial Disparities in hospital strategies that lead to persistent reductions among vulnerable Readmission Rates After Medicare’s Hospital Readmissions Reduction populations. The Centers for Medicare & Medicaid Services and the Program Within Safety-Net and Non–Safety-Net Hospitals. JAMA Netw National Quality Forum are considering hospital-wide readmission Open. 2018;1(7):e184154. doi:10.1001/jamanetworkopen.2018.4154 measures, which may help to mitigate differences in trends by clinical LEAD AUTHOR DR. KRISDA CHAIYACHATI Krisda Chaiyachati, MD, MPH, MSHP, is an Assistant Professor of Medicine at the University of Pennsylvania. He designs, studies, and implements innovative strategies for improving health care accessibility, addressing social barriers to care, and managing population health goals. His research seeks to understand how to improve health outcomes for low-income and minority populations.