Excess 30-Day Heart Failure Readmission In Blacks Increases With Neighborhood Deprivation

Abstract

Introduction Neighborhood socioeconomic status (SES) has been associated with incident heart failure (HF) as well as HF readmissions. Neighborhood SES may be particularly impactful for vulnerable populations. Hypothesis: Neighborhood SES modifies the disparity in 30-d HF readmissions between Black and White patients in the Southeastern US. Methods We utilized the Emory Clinical Data Warehouse to create a geo-coded retrospective cohort of all patients with at least one inpatient HF hospitalization at any Emory Healthcare facility from 2010-2018. Quintiles of the Social Deprivation Index (SDI), derived from socioeconomic US Census data, were used to characterize neighborhood deprivation at the census tract level. Linear probability models were estimated with robust errors to examine “excess 30-d HF readmissions” as the absolute difference in probability of readmission between Black and White (referent) patients within quartiles of neighborhood deprivation. Models accounted for patient clustering, and multivariable models adjusted for patient socio-demographics (age, gender, and insurance type), HF type (systolic, diastolic, other), other medical comorbidities (hypertension, diabetes, CKD, CAD, AFib, COPD, PAD, CVA/TIA), vital signs (SBP, HR, RR), laboratory values (serum sodium, creatinine, eGFR, BUN, Hb), discharging specialty, and hospital location. Results There were 30,630 patients in the cohort, with mean age 66.1 ± 15.5 years, 48% female, 53% Black, and 67% insured through Medicare. Compared with White patients, Black patients were younger in age, more likely to be female, more likely to reside in deprived census tracts, have diabetes, hypertension, chronic kidney disease, and had higher comorbidity scores. Blacks had lower coronary artery disease and atrial fibrillation. Between 2010-2018, 21% of Black and 14% of White patients experienced a 30-d HF readmission (p Conclusion Excess 30-d HF readmissions in Black compared with White patients increases with neighborhood deprivation and is not explained by patient sociodemographic factors or medical history. Future studies should focus on specific modifiable features of the built environment that may be leveraged to disrupt the excess risk.