Abstract 2: Impact of Race and Hospital Racial Makeup on Readmissions for Heart Failure

Abstract

Reducing racial/ethnic disparities in health care is a policy priority, and heart failure (HF) is a particularly important area of focus, not only because HF is the most common cause of hospitalizations and readmissions in the Medicare program, but because black patients suffer a disproportionate burden of this disease. Prior studies have shown that black patients may have higher readmission rates than white patients, but the reasons for these differences are unclear. We sought to determine whether racial disparities in readmission rates for HF were due to the site where black and white patients received care. We used national Medicare data from 2006-2007 to categorize hospitals based on the proportion of black patients they cared for, designating those in the top decile as “minority-serving.” We created risk-adjusted generalized linear regression models, with patient race and hospital racial makeup as our primary predictors, and 30-day readmission as our primary outcome. Forty percent of all black patients and five percent of all white patients were cared for at minority-serving hospitals. In risk-adjusted analyses, black patients had higher 30-day readmission rates than white patients (24.1% vs. 23.3%, p<0.0001). However, in stratified analyses, we found that there were no disparities in readmission rates at non-minority-serving hospitals (Table), and black patients at non-minority-serving hospitals had significantly lower admission rates than white patients at minority-serving hospitals. These results were unchanged after controlling for length of stay, hospital characteristics, and proportion of Medicaid coverage in a hospital. Our findings suggest that racial disparities in readmissions are largely due to the site at which care is provided. Focusing quality improvement efforts at poor-performing, minority-serving hospitals has the potential to improve quality of care for all HF patients while simultaneously reducing disparities in HF care. Minority-serving hospitals Non-minority-serving hospitals Black patients White patients P value for comparison, black v. white Black patients White patients P value for comparison, black v. white Model 1 (Patient characteristics only) 26.2% 25.1% 0.006 23.3% 23.1% 0.33 Model 2 (Model 1 plus length of stay) 26.0% 25.1% 0.014 23.3% 23.2% 0.75 Model 3 (Model 2 plus hospital characteristics) 25.4% 24.3% 0.005 23.0% 22.8% 0.49 Model 4 (Model 3 plus percent Medicaid) 25.0% 24.0% 0.008 22.8% 23.0% 0.51