Estimating and explaining the spread of COVID-19 at the county level in the USA

The basic reproduction number, R 0 , determines the rate of spread of a communicable disease and therefore gives fundamental information needed to plan public health interventions. Estimated R 0 values are only useful, however, if they accurately predict the future potential rate of spread. Using mortality records, we estimated the rate of spread of COVID-19 among 160 counties and county-aggregates in the USA. Most of the high among-county variance in the rate of spread was explained by four factors: the timing of the county-level outbreak (partial R 2 = 0.093), population size (partial R 2 = 0.34), population density (partial R 2 = 0.13), and spatial location (partial R 2 = 0.42). Of these, the effect of timing is explained by early steps that people and governments took to reduce transmission, and population size is explained by the sample size of deaths that affects the statistical ability to estimate R 0 . For predictions of future spread, population density is important, likely because it scales the average contact rate among people. To generate support for a possible explanation for the importance of spatial location, we show that SARS-CoV-2 strains containing the G614 mutation to the spike gene are associated with higher rates of spread ( P = 0.016). The high predictability of R 0 based on population density and spatial location allowed us to extend estimates to all 3109 counties in the lower 48 States. The high variation of R 0 among counties argues for public health policies that are enacted at the county level for controlling COVID-19.