A spatio-temporal statistical model to analyze COVID-19 spread in USASoudeep Deb, Siddharth Rawat
Journal: Journal of Applied Statistics
The Coronavirus pandemic has affected the whole world extensively and it is of immense importance to understand how the disease is spreading. In this paper, the authors provide evidence of spatial dependence in the pandemic data and accordingly develop a new statistical technique that captures the spatio-temporal correlation pattern of the COVID-19 spread appropriately. The proposed model uses a separable Gaussian spatio-temporal process, in conjunction with an additive mean structure and a random error process.
The model is implemented through a Bayesian framework, thereby providing a computational advantage over the classical way. They use state-level data from the United States of America in this study. Their analysis indicates that the spread of the disease remains significantly correlated in space for up to 1000 kilometres, and significantly correlated in time for a long period. It also establishes that a quadratic trend pattern is most appropriate to model the prevalence of the disease.
Interestingly, the population is found not to affect the prevalence rate, whereas more number of deaths in the previous week positively affects the spread of the disease. They conduct a detailed residual diagnostics to show that the model is adequate enough for understanding the spatio-temporal dependence pattern in the data. It is also shown to have superior predictive power, for both short term (one week) and long term (up to three months), than other popular spatial and temporal models. Last but not the least, they discuss the flexibility of the proposed approach to be adapted and extended to other COVID-19 related studies in future.