Adaptation of the COVASIM model to incorporate non-pharmaceutical interventions: Application to the Dominican Republic during the second wave of COVID-19

Abstract

[EN] We adapt the Covasim agent-based model for predicting new COVID-19 cases by tuning the transmissibility rate with information on the impact of the most common non-pharmaceutical interventions (NPIs) obtained through machine learning models. Such impact has been estimated thanks to the information on applying pools of NPIs worldwide from the Oxford COVID-19 Government Response Tracker. This approach permits the simulation of a whole country or a smaller region, providing information about asymptomatic, recovery, severe, and critical new cases and enabling governments and authorities to set NPIs plans to cope with the pandemic.