Abstract P352: Racial Disparity in Post-stroke Readmission - a South Carolina Perspective

Abstract

Background: Readmissions after acute hospitalizations are a cause of both risk and expense, and many of them are potentially preventable. Importantly, risk-standardized hospital readmission rates are sometimes used as a yardstick of the quality of care offered. However, racial variability in readmissions might involve factors beyond quality of care and has not been studied extensively. During our pilot investigation using 90-day post-stroke readmissions data at Medical University of South Carolina (MUSC), we found significant disparities between African Americans and Caucasians. Objective: To identify differences in readmissions between African Americans and other races and determine preventable readmissions from a pragmatic viewpoint. Methods: We obtained deidentified data from Health Sciences South Carolina (HSSC) Clinical Data Warehouse (CDW). The data was comprised of three institutions: Medical University of South Carolina (MUSC), Palmetto Health and Greenville Hospital System University Medical Center. The data consisted of on adult admissions with index diagnosis considered as an ischemic stroke (or closely related) using International Classification of Diseases, Ninth and Tenth Revision (ICD-9, ICD-10) codes between January 2011 and April 2017. Of these, we will determine readmission and reason for readmission over 90-day period. Readmission can be hospital or emergency room readmission. Results: Our database contains 32,548 patients who have been provided clinical care for stroke. Out of these patients 8,308 (25.5%), 23,085 (70.9%) and 1,155 (3.5%) are African Americans, Caucasians and others, respectively. We will present weekly readmission trends over 90 days and evaluate if there are disparities across races. We will apply chi-square test and Student’s t-test to determine statistical significance. For weekly readmission trends over 90 days, we will apply Kolmogorov-Smirnov test to identify difference in readmission patterns across races. We will also identify confounders like socioeconomic status and age and their influence in the racial disparity. Conclusions: From a single center retrospective data, we found that 90-days readmission rates involve African Americans in a disproportionate manner. This multicenter data analysis will further shed light on the etiology of readmission, confounders and the care offered.