New York City was one of the first cities to be infected by COVID-19 in the United States. It has also been the worst hit, with half a million confirmed cases and 36,000 COVID-related deaths (New York Times, April 23, 2020). By comparing COVID-19 test results with demographic and socioeconomic factors by ZIP code, we found that low education levels, crowded housing, and a lack of health insurance are some of the strongest predictors of high COVID-19 positivity rates.
We used a statistical tool called spatial regression to measure how much the differences in COVID-19 positivity rates from ZIP code to ZIP code can be explained by differences in demographic factors. We began with 31 demographic variables and three health outcome variables. By fitting a regression model, removing insignificant variables, and retesting, we see that most of the differences in positivity rates (positive tests as a percent of total tests) can be explained by just a few socioeconomic factors: education, room crowding in homes, lack of health insurance, percent of people over 60 years, and segregation.
