Upstream Factors Associated With Hospitalization in Black- and Minority-Serving Hospitals.

Importance: Racial separation of care contributes to racial disparities in health. Objective: To examine the association of place-based contextual factors with hospitalization in a hospital that disproportionately serves Black patients (Black-serving hospital [BSH]) or a hospital that disproportionately serves racial and ethnic minority patients (minority-serving hospital [MSH]) and whether associations are modified by individual race and ethnicity or whether the hospitalization is elective. Design, Setting, and Participants: This cross-sectional analysis of data collected from January 2011 through December 2020 from the nationally representative Medicare Current Beneficiary Survey linked with Medicare claims was conducted from March 2023 through September 2025. Participants included all Medicare-linked survey respondents aged 65 years or older who were hospitalized during the study period. Exposures: Neighborhood disadvantage categorized as 5 groups of the Social Deprivation Index (SDI) and hospital referral region (HRR) residential segregation, with test of effect modification by race and ethnicity and whether the hospitalization was elective. Main Outcomes and Measures: Hospitalization in a BSH (primary outcome) or MSH (secondary outcome) vs not. Results: The final sample included 8735 respondents (mean [SD] age, 79.6 [8.2] years; 4967 [56.9%] female; 756 [8.7%] Black, 517 [5.9%] Hispanic, 7093 [81.2%] non-Hispanic White, and 369 [4.2%] multiracial or other). The 8540 with SDI data were evenly distributed in SDI groups 1 through 4 (20.9%-21.4% in each group), with 15.9% in SDI group 5; 77.9% lived in moderately segregated HRRs and 12.1% in highly segregated HRRs. In models fully adjusted for patient-level, HRR-level, and regional covariates, neighborhood disadvantage (adjusted odds ratio [AOR] for most disadvantaged compared with least: 2.03 [95% CI, 1.56-2.63]) and for residential segregation (AOR for high compared with low: 2.99 [95% CI, 1.98-4.52]) were independently associated with hospitalization in a BSH. In stratified models, in the least disadvantaged or segregated areas all patients were unlikely to be hospitalized in a BSH; however, in the most disadvantaged or segregated areas, White patients had a small increase in probability of BSH hospitalization (AOR, 1.70 [95% CI, 1.24-2.34]; P = .001; or AOR, 1.98 [95% CI, 1.23-3.18]; P = .005, respectively), while Black patients had a larger increase (AOR, 2.77 [95% CI, 1.15-6.60]; P = .02; or AOR, 5.55 [95% CI, 1.86-16.57]; P = .002, respectively). Elective hospitalizations differed in that race and ethnicity and residential segregation were not independently associated with site of care, neighborhood disadvantage was only marginally significant, and interactions were not significant. Conclusions and Relevance: This cross-sectional study found an association between site of care and place-based factors, and that it was modifiable by race and ethnicity and whether the hospitalization was elective. Neighborhood disadvantage and residential segregation were independently associated with receiving care in a BSH for Black patients but odds were lower for White patients, and elective hospitalizations had less place- and race-based separation of care. These findings underscore the need for (1) further research to understand these differences, (2) to address place-based contextual factors, and (3) to invest in Black-serving hospitals.