A Semiparametric Regression Model with Online Estimation Algorithm for Dynamics of Disease Transmission

• Activity Number: 300 - Addressing Statistical Challenges of the COVID-19 Crisis
• Type: Invited
• Date/Time: Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
• Sponsor: ENAR
• Author(s): Jing Huang* and Jiasheng Shi and David Rubin and Jeffrey S. Morris
• Companies: University of Pennsylvania Perelman School of Medicine and University of Pennsylvania Perelman School of Medicine and Children's Hospital of Philadelphia and University of Pennsylvania Perelman School of Medicine
• Keywords: instantaneous reproduction number; observation-driven model; Poisson process; Quasi-score method; time series

Abstract

Reproduction number (R) plays a central role in predicting the evolution of an infectious disease outbreak like the outbreak of this destructive SARS-CoV-2, which however varies by location and by time due to multiple factors.

In order to study the dynamics of disease transmission, we modeled the instantaneous reproduction number Rt, t? 0, which is allowed to vary over time. We proposed an online algorithm to iteratively estimate Rt using an observation-driven Poisson regression model with a latent Markov process, and to study the impact of covariates on its variation. Our estimators allow a close monitor and dynamic update on the knowledge of Rt whenever new data were available, and allow a forecasting of future Rt under different conditions to provide guidance for policy making.

We conduct analysis on a national dataset with more than 500 counties and 3 million cases in United States, where the intriguing results reveal that among the county-level factors been studied, implementation of social distancing is the most significant in reducing transmission, population density and percentage of elders are positively associated with Rt, while temperature has limited impact.