An Agent-Based Model to assess COVID-19 spread and health systems burden in Telangana state, India

Abstract

Objectives: To assess the transmission dynamics and the health systems burden of COVID-19 using an Agent Based Modeling (ABM) approach using a synthetic population. Study design: The study used a synthetic population with 31,738,240 agents representing 90.67 percent of the overall population of Telangana state, India as per 2011 Census of India. Lockdown phases as per Indian scenario considering the effects of post-lockdown, use of control measures and immunity on secondary infections were studied. District-level localized parameters were assigned to agents as local models prove to be much helpful for policymakers. Methods: The counts of people in different health states were measured separately for each district of Telangana. The model was run for 365 days and six scenarios with varying proportions of people using control measures (100%, 75% and 50%) and varying immunity periods of recovered patients (90 and 180 days). Results: Results indicate that the peak values were attained soon after the lockdown was lifted. The risk estimates indicate that protection factor values are higher when more proportion of people adopt control measures such as use of face mask and social distancing. Population Attributable Risk values measured longitudinally indicated higher values like 60.41% and 47.18% when 75 percent of people followed control measures during lockdowns. Conclusions: ABM approach helps to analyze grassroot details compared to compartmental models. Risk estimates allows the policymakers to determine the protection offered, its strength and percentage of population shielded by use of control measures.