A Hierarchical Epidemic Model to Estimate the Time-Varying Reproduction Numbers and Prevalence of COVID-19 Infections Accounting for Under-Reporting and Delayed Reporting of Cases

• Activity Number: 191 - Statistical Research in Rapid Response to COVID-19 Pandemic: Forecasts, Risk Factors, Therapeutics, and Vaccine Trials
• Type: Invited
• Date/Time: Tuesday, August 10, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Rounak Dey* and Haoyu Zhang and Corbin Quick and Xihong Lin
• Companies: Harvard T.H. Chan School of Public Health and Harvard T.H. Chan School of Public Health and Harvard T.H. Chan School of Public Health and Harvard T.H. Chan School of Public Health
• Keywords: COVID-19; Epidemic model; Under-reporting; Reproduction number

Abstract

As the COVID-19 pandemic continues to spread globally, estimating the time-varying reproduction number (Rt) is crucial to monitor the disease spread and inform policy decisions for controlling the disease. However, modelling COVID-19 transmission dynamics faces substantial challenges. Many cases are likely to be un-ascertained due to insufficient testing, and the ascertainment rate can also vary over time due to evolving testing capacity. Moreover, case counts are often subject to delayed reporting. Epidemic models that ignore un-ascertainment and reporting delays can lead to biased estimates of the transmission rate. To address these problems, we propose a hierarchical epidemic model to estimate the time-varying Rt curve and the total number of infected individuals, using daily case counts of a given region by explicitly modelling the time-varying ascertainment rate as a function of the daily number of tests. We further adjust the daily case counts using both fixed and random lags to adjust for the reporting delays. We apply our method on simulated epidemic data as well as the state-level data in the United States to provide the time-varying Rt curves and prevalence of COVID-19.