Early Stage Prediction of US County Vulnerability to the COVID-19 Pandemic

Mihir Mehta; Juxihong Julaiti; Paul Griffin; Soundar Kumara

The dataset is comprised of 37% urban and 63% rural counties based on the urban and rural county definition for year 2013. 49 In order to determine if there is an association between urbanicity and vulnerability, we performed a set of one-sided t-tests. The null hypothesis -the 10% least vulnerable counties would have the same proportion of rural counties as the actual proportion of rural counties in the dataset -was rejected for every day from March 14 th to March 26 th . Additionally, the null hypothesis -the actual positive instances counties would the same proportion of urban counties as the actual proportion of urban counties in the dataset -was also rejected for every day over the analysis period. It can therefore be concluded that there is a positive association between urban and most vulnerable counties as well as rural and least vulnerable counties. The continuous decreasing trend in the confidence interval of the urban counties proportion estimate within actual positive instance counties can be used to infer that COVID-19 is propagating from urban counties to rural counties.