Addressing the Health Versus Economy Dilemma in Data-Driven Policymaking During a Pandemic

Abstract

The recent COVID-19 pandemic highlighted a need for tools to help policy-makers make informed decisions on what policies to implement in order to reduce the impact of the pandemic. Several tools have previously been developed to model how non-pharmaceutical interventions (NPIs), such as social distancing, affect the rate of growth of a disease within a population. Much of the focus of the modelling effort have been on projections of health factors, relating them to the NPIs, with only few works addressing the health-economy trade-off. However, there is a particular gap in illustrations of real data-driven solutions in this area. In this paper, we proposed a purely data-driven framework where we modelled health and economic impacts with Bayesian and Recurrent Neural Network (RNN) models respectively, and used NSGA-II to identify policy stringencies over a three-week period. We demonstrate that this framework can produce a range of solutions trading off between health and economy projections based on real data, that may be used by policymakers to reach an informed decision.