Lessons from being challenged by COVID-19

Abstract

We present results of different approaches to model the evolution of the COVID-19 epidemic in Argentina, with a special focus onthe megacity conformed by the city of Buenos Aires and its metropolitan area, including a total of 41 districts with over 13 millioninhabitants. We first highlight the relevance of interpreting the early stage of the epidemic in light of incoming infectious travelersfrom abroad. Next, we critically evaluate certain proposed solutions to contain the epidemic based on instantaneous modificationsof the reproductive number. Finally, we build increasingly complex and realistic models, ranging from simple homogeneous modelsused to estimate local reproduction numbers, to fully coupled inhomogeneous (deterministic or stochastic) models incorporatingmobility estimates from cell phone location data. The models are capable of producing forecasts highly consistent with the officialnumber of cases with minimal parameter fitting and fine-tuning. We discuss the strengths and limitations of the proposed models,focusing on the validity of different necessary first approximations, and caution future modeling efforts to exercise great care in theinterpretation of long-term forecasts, and in the adoption of non-pharmaceutical interventions backed by numerical simulations.