Pandemic Risk Assessment and Management in a Bayesian Framework

Credible assessment of a pandemic is critical in decision-making by healthcare providers, local and national government agencies, and international organizations. It facilitates a valuable lead time to policy-makers for efficient planning, resource allocation, and enforcing interventions. This work presents a Bayesian-based semi-mechanistic model for a short-term forecast of pandemic risk. We employed a hierarchical Bayesian network to learn the parameters of the susceptible, exposed, infected, quarantined, recovered, deceased (SEIQRD) model to forecast the risk under evolving conditions due to imposed regulations, varied individual responses, and the advent of the multiple waves of a pandemic outbreak. The model has been validated for the risk forecast of the Coronavirus Disease of 2019 (COVID-19) using the benchmark models reported by the Center for Disease Prevention and Control and real-world data. We have also presented the impact analysis of COVID-19 pandemic risk using Bayesian inference. The proposed hierarchical Bayesian-based semi-mechanistic model (HBN-SEIQRD) resulted in accurate prediction with the lowest leave-one-out (LOO) cross-validation scores among the proposed Bayesian frameworks. The model is helpful in predicting the disease trajectory under evolving conditions due to government regulations, societal responses, and individual practices