Abstract Body: Introduction: Residential segregation influences individual health outcomes across the life course. Historic and modern racist lending practices, a form of structural racism, have contributed to residential segregation. Using historic and contemporary measures, we investigated whether longitudinal lending discrimination was associated with stroke prevalence in the U.S.
Methods: We used a historic redlining index which linked Home Owners’ Loan Corporation (HOLC) maps to 2010 census tract boundaries to measure the degree of redlining (ranging from best [1] to hazardous [4]). Historic redlining data was combined with current lending discrimination (Home Mortgage Disclosures Act) to assess perceived area lending risk over time. Historic redlining index was dichotomized at the 25^th percentile, and paired with modern lending discrimination measure to create a 4-level categorical measure of lending trajectory: (1) persistent discrimination- high levels of redlining & current discrimination; (2) growing investment- high levels of redlining & no current discrimination; (3) declining investment- low levels of redlining & current discrimination; and (4) no discrimination- low levels of redlining & no current discrimination. Census tract stroke prevalence was estimated using 2021 CDC PLACES data. ANOVA was used to measure the differences between categories of longitudinal lending discrimination.
Results: Total of 68172 census tracts had stroke prevalence estimates; of those 10351 were HOLC graded. Nearly half (47%) of graded census tracts had persistent discrimination and tracts with persistent discrimination were younger, less non-Hispanic white, and had higher proportions of diabetes, obesity and hypertension (Table). Stroke prevalence was 1.59% (95% CI:1.50-1.69) higher in tracts with persistent discrimination compared to tracts with no discrimination. Stroke prevalence was also higher in tracts with declining investment (1.40%; 95% CI: 1.18-1.63%) however, tracts that had growing investment had only slightly higher stroke prevalence (0.27%; 95% CI: 0.18-0.37%) compared to tracts with no discrimination. Visual geospatial correlation is shown between stroke and persistent discrimination in New York City (Figure).
Conclusions: Areas with persistent lending discrimination for the last 80-90 years are associated with higher stroke prevalence. The influence of the historical and modern social environment in which one resides influences the burden of stroke in a community.