Associations Between Neighborhood-Level Characteristics and COVID-19 Incidence and Mortality in Southeastern Pennsylvania

Abstract

Rationale: The COVID-19 pandemic has disproportionately impacted racial/ethnic minority and socioeconomically disadvantaged groups in the United States, including at-risk populations within Southeastern Pennsylvania. We sought to determine whether neighborhood-level health, demographic, and socioeconomic characteristics in Southeastern Pennsylvania were associated with COVID-19 incidence and mortality at the zip code and municipality level, thereby establishing whether neighborhood-level disparities mirror individual-level ones. Methods: Cumulative zip code- and municipality-level data on COVID-19 cases and deaths were obtained from the public data hubs of 5 counties in Southeast Pennsylvania, and those of individual long-term care facilities (LTCFs) were obtained from the Pennsylvania Department of Health. For corresponding geographic areas, demographic and socioeconomic status variables were obtained from the American Community Survey, and data on the health status and behaviors of local residents were obtained from the Southeastern Pennsylvania Household Health Survey. COVID-19 cases and deaths reported by LTCFs were excluded from area-aggregated counts. Multivariable quasi-Poisson models with offsets for population counts were created to determine whether neighborhood-level variables were associated with COVID-19 incidence and mortality. Before adjusted incidence rate ratios were calculated, such models included individual predictors that were significantly associated (p<0.05) with COVID-19 outcomes and excluded highly collinear terms as determined by having variance inflation factors greater than 3. Results: Among 208 zip codes and municipalities that had complete data, the COVID-19 cumulative incidence through July 24, 2020 ranged from 0 to 331.9 per 10,000 residents, and the COVID-19 mortality rate ranged from 0 to 1.0 per 10,000 residents. Among 45 neighborhood-level variables considered, 5 were independently associated with COVID-19 incidence (p<0.01): 1) the proportion of residents aged 65 years or older (incidence rate ratio [IRR] = 1.341, 95%-CI: 1.147-1.567 for a 10% increase), 2) population density (IRR = 1.002, 95%-CI: 1.001-1.003 for a 100 people/square kilometer increase), 3) the proportion of individuals eating 3 or more servings of fruits/vegetables daily (IRR = 0.891, 95%-CI: 0.836-0.950 for a 10% increase), 4) average median house value (IRR = 0.989, 95%-CI: 0.980-0.994 for a $10,000 USD increase), and 5) the proportion of 2-person households (IRR = 0.997, 95%-CI: 0.995-0.999 for a 10% increase). The proportion of individuals aged 65 years or older was the only factor independently associated with COVID-19 mortality (IRR = 2.59, 95%-CI: 1.55-4.41 for a 10% increase). Conclusions: Neighborhood-level data can help identify specific needs of vulnerable populations and inform policies to address health disparities related to COVID-19.