Mathematical Models of Epidemics: Tracking Coronavirus using Dynamic Survival Analysis

Grzegorz Rempala (March 24, 2020)

Please install the Flash Plugin

Abstract

As the outbreak of COVID-19 in the city of Wuhan appears to be the beginning of a global pandemic, there is much public interest in predicting both the dynamics and the size of the ongoing regional outbreaks in different countries. It is also important to ascertain the potential effects of early interventions such as school closures and mandatory or self-imposed quarantines. To answer some of these questions, we propose a general framework for analyzing the ongoing outbreak trend using data from a partially observed epidemic curve under minimal assumptions that are clearly speci- fied. In particular, this framework does not assume any specific infectious or recovery periods (which are often unknown) or observable prevalence of the disease (allowing, for instance, for silent infectives). We show that this analysis can help anticipate both the likely temporal trends of an ongoing epidemic as well as its final size in a commu- nity with or without social distancing. We use our approach to predict the trajectory of the epidemic curve from Wuhan city in Hubei province, which is most detailed one available to date from the COVID-19 outbreak.