The Impact of Race and Neighborhood Racial Composition on Preventable Readmissions for Diabetic Medicare Home Health Beneficiaries.

BACKGROUND: The recommended home health financial penalty program for preventable readmission does not factor race/ethnicity and neighborhood racial compositions into the determination of preventable readmission rates. Home health agencies may avoid beneficiaries from certain racial/ethnic groups and neighborhoods if these two factors have an effect on preventable readmissions. We examined the association between preventable readmissions with race/ethnicity and neighborhood racial composition.

METHODS: Several 2009 national data were used, such as the Master Beneficiary Summary File, Medicare Provider Analysis and Review File, and Outcome Assessment Information Set. Our sample consisted of diabetic Medicare home health beneficiaries (African-Americans and Whites only). We analyzed predictors of time-to-first 30-day preventable readmission, including short/long-term diabetic complications, chronic obstructive pulmonary disease/asthma, bacterial pneumonia, dehydration, urinary tract infection, hypertension, heart failure, angina without procedure, uncontrolled diabetes, and lower-extremity amputation.

RESULTS: There were 86,567, 17,262, and 11,392 observations in neighborhoods with low (6 % African-Americans), moderate (35 % African-Americans), and high (76 % African-Americans) density of African-Americans, respectively. Using Cox regression models, we found that in neighborhoods with moderate and high density of African-Americans, African-Americans had 21 % (hazard ratio (HR) 1.21; 95 % confidence interval (CI) 1.04-1.39) and 24 % (HR 1.24; 95 % CI 1.01-1.52) significantly higher hazards of 30-day preventable readmissions than Whites, respectively.

CONCLUSION: Race and neighborhood racial compositions are beyond home health providers' control. These two factors should be considered as covariates for the preventable readmissions in the recommended home health financial penalty program.