How the Neighborhood Environment Shapes COVID-19 Burden: Early Findings from the COVID Neighborhood Project.

Abstract

Introduction A lack of fine-scale, spatially-resolute case data has limited examination of the distribution of COVID-19 across neighborhoods within U.S. states. This hinders our ability to ascertain the true COVID-19 burden in communities.

Methods We collected spatially-referenced COVID-case data from April 2020 through April 2022 at the census tract or ZIP code level for 21 states as part of the COVID-19 Neighborhood Project (CONEP). Case data were linked with neighborhood-level socioeconomic status (SES), specifically neighborhood affluence and disadvantage, using data from the National Neighborhood Data Archive (NaNDA). For each state, we calculated the cumulative COVID-19 case count per 100,000 population and median and interquartile range (IQR) of neighborhood cases. We then conducted linear regression models estimating the mean increase in neighborhood-level cumulative COVID-19 case count per 100,000 associated with a one-unit increase in neighborhood-level SES, controlling for neighborhood population density and county-level political partisanship.

Results In descriptive analyses, we observed substantial differences in state IQRs, demonstrating that COVID-19 spread varied state by state. Oregon had the smallest IQR with 2467 cases, suggesting a more homogenous spread of COVID-19 across neighborhoods in that state; Nevada had the largest IQR with 22300 cases, suggesting a heterogenous spread. In linear regression models, we also found that in many states, the neighborhood-level COVID-19 case count was associated with neighborhood-level SES, neighborhood population density, and county-level political partisanship.

Conclusions There are differences in the COVID-19 burden distribution across neighborhoods within states, so exploring local contexts is critical in understanding how COVID-19 will impact communities long-term.