Illuminating Hospital Disparities in Readmissions for Patients with Social Risk Factors: Comparing Hospital Performance Using Two Different Approaches

Abstract

Research ObjectiveSocioeconomic disparities in hospital readmissions for Medicare patients are well established. Medicare currently provides hospitals with within‐hospital disparities in readmission using dual eligibility status as a social risk factor. Community‐based social risk factors, such as the AHRQ SES Index (ASI), may allow capture of different types of information regarding social risk, as compared to income and assets, which are principally used for establishing dual eligibility. This study examines concordance of within‐hospital disparities in 30‐day readmission rates for Medicare patients that arise from dual eligibility status and ASI, and how this concordance varies across patients admitted for three common clinical conditions (AMI, HF, and pneumonia).Study DesignWe conducted a retrospective cohort study of hospitalized Medicare fee‐for‐service (FFS) patients meeting the inclusion criteria for the 30‐day AMI, HF, and PN hospital readmission measures used by Medicare. We used Medicare Part A and B Standard Analytic Files and Enrollment Files to identify the study cohort, clinical risk factors for adjustment, and enrollment information. We calculated the within‐hospital disparity score, for each hospital, using dual eligibility (defined as full‐benefit Medicaid enrollment at time of discharge) and the ASI score for beneficiary 9‐digit ZIP code of residence. Using the method currently used by CMS for estimating within‐hospital disparities, we calculated the hospital‐specific disparity effect for each hospital and each social risk factor for each of the three conditions. Disparity effects are reported as hospital‐specific odds ratios (ORs) and absolute risk differences; we calculated the Spearman rank‐order correlations between the two disparity effects for each condition.Population StudiedHospitalized U.S. Medicare FFS patients aged 65 years and older admitted to U.S. acute care, critical access, and Veterans Affairs hospitals from July 1, 2015, to June 30, 2018.Principal FindingsThe final sample included 3,126,185 Medicare FFS admissions to 4,730 U.S. hospitals. Both dual eligibility (mean ORs for three conditions ranging from 1.04 to 1.08) and AHRQ SES Index (mean ORs ranging 1.03‐1.05) were associated with higher adjusted odds of 30‐day readmission. However, hospital‐specific disparity effects for the two risk factors were uncorrelated for AMI (Spearman’s correlation □ = ‐0.017, P = .28), weakly correlated for HF (□ = 0.10, P < .001), and moderately correlated for PN (□ = 0.49, P < .001). Results were similar for absolute risk differences.ConclusionsDisparities in 30‐day readmission rates for hospitalized Medicare patients may differ across different dimensions of social risk, depending on the reason for admission. Within‐hospital disparities related to different social risks may capture distinct effects, according to patient condition.Implications for Policy or PracticeEfforts to illuminate socioeconomic disparities in hospital outcomes will benefit from examination of different social risk factors; a single social risk factor may not fully capture the impact of disparities in outcomes.Primary Funding SourceThe study was funded by the Centers for Medicare and Medicaid Services.