Associations of Individual and Neighborhood Factors with Disparities in COVID-19 Incidence and Outcomes.

Introduction: The disproportionate impact of coronavirus 2019 (COVID-19) on Black and Hispanic communities has been widely reported. Many studies have used neighborhood racial/ethnic composition to study such disparities, but less is known about the interplay between individual race/ethnicity and neighborhood racial composition. Therefore, our goal in this study was to assess the relative contributions of individual and neighborhood risk to disparities in COVID-19 incidence and outcomes. Methods: We performed a cross-sectional study of patients with emergency department (ED) and inpatient visits to an academic health system (12 hospitals; February 1-July 15, 2020). The primary independent variable was race/ethnicity; covariates included individual age, sex, comorbidity, insurance and neighborhood density, poverty, racial/ethnic composition, education and occupation. The primary outcome was severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) positivity; secondary outcomes included admission and death after COVID-19. We used generalized estimating equations to assess whether race/ethnicity remained significantly associated with COVID-19 after adjustment for individual and neighborhood factors. Results: There were 144,982 patients; 5,633 (4%) were SARS-CoV-2 positive. Of those, 2,961 (53%) were admitted and 601(11%) died. Diagnosis of COVID-19, admission, and death were more common among non-Hispanic Black, Hispanic, Spanish-speaking patients, and those with public insurance. In the base model (adjusting for race/ethnicity, age, sex, and comorbidities), race/ethnicity was strongly associated with COVID-19 (non-Hispanic Black odds ratio [OR] 4.64 [95% confidence interval (CI) 4.18-5.14], and Hispanic OR 6.99 [CI 6.21-7.86]), which was slightly attenuated but remained significant after adjustment for neighborhood factors. Among patients with COVID-19, there was no significant association between race/ethnicity and hospital admission, other than for patients with unknown race. Conclusion: This data demonstrates a persistent association between race/ethnicity and COVID-19 incidence, with Black and Hispanic patients at significantly higher risk, which was not explained by measured individual or neighborhood factors. This suggests that using existing neighborhood factors in studies examining health equity may be insufficient, and more work is needed to quantify and address structural factors and social determinants of health to improve equity.