Abstract P26: 30-Day Risk Standardized Mortality and Readmission After Ischemic Stroke by Hospital Minority Volume

Background: Risk-standardized 30-day mortality (RSMR) and readmission (RSRR) rates in part reflect hospital-based and immediate post-discharge care. Differences in rates across hospitals based on the race-ethnic composition of their patient populations may indicate disparities in care. We examined whether hospital 30-day RSMRs and RSRRs after ischemic stroke vary by the proportions of African-Americans (AA) admitted at hospitals in 1994, 2000, and 2006. Methods: We used Medicare claims data to determine the hospital-level RSMR for any cause 30 days after the index admission and RSRR for any cause within 30 days of discharge among surviving fee-for-service Medicare patients >= 65 years of age discharged with a primary diagnosis of ischemic stroke (ICD-9 433, 434, 436) in 1994, 2000, and 2006. Hierarchical regression models calculated the RSMR and RSRR for hospitals, adjusting for patient demographics and clinical characteristics. Hospitals were grouped into quartiles based on the proportion of AA admitted. Mixed models and quantile regression models examined the relationship between AA volume and outcomes. Results: The RSMR decreased as the proportion of AA increased, whereas the RSRR increased as the proportion of AA increased (figure). These relationships were consistent within each year, and the variability lessened over time. Quantile regression confirms that the relationship persists across the distribution of RSMR and RSRR. Conclusions: RSMR and RSRR varied by hospital race-ethnic volume, but differences in outcome lessened over time. Identifying factors contributing to these differences may identify opportunities to improve stroke care among the elderly.