COVID-19 Incidence Rate in Minnesota: A Bayesian Spatial Analysis

Since the beginning of 2020, Coronavirus disease-2019 (COVID-19) has spread across the world with unprece-dented speed. Almost a half-billion COVID-19 cases have been recorded globally, with almost 1.5 million in Minnesota alone. Because of limited knowledge surrounding the transmission of COVID-19 and the immediate need to control outbreaks, modeling the spread of COVID-19 is of great importance. Though studies modeling COVID-19 have been done, there are few tudies that use Bayesian hierarchical spatial models for this purpose. This study utilizes the Bayesian spatial Conditional Autoregressive Leroux Model to model the number of confirmed COVID-19 cases in the state of Minnesota by county. Data for COVID-19 cases from the start of the pandemic to present day were obtained from the Minnesota Department of Health, and relevant covariates for COVID-19 were added to the model. These covariates include vaccination data by county (and stratified by age group), population density, distance from the capital cities (Minneapolis and St. Paul), as well as a number of variables that describe the demographic breakdown of each county (by race, age group, socioeconomic factors, and health conditions). The final model was chosen primarily based on Watanabe–Akaike Information Criterion, analysis of residuals using Moran’s I-Statistic, Geweke’s diagnostic, and inspection of the 95% credible intervals for model coefficients. Results indicate that Bayesian spatial models are an appropriate model choice to model the spread of the novel coronavirus in Minnesota. There are many covariates, such as vaccination rate and median age, that provide valuable information about counties’ COVID-19 incidence rates