Transmission Dynamic Modeling for COVID-19 Data in US and the World

• 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): Haoyu Zhang* and Chaolong Wang and Xihong Lin
• Companies: Harvard T.H. Chan School of Public Health and Huazhong University of Science and Technology and Harvard T.H. Chan School of Public Health
• Keywords: COVID-19; Transmission; Ascertainment; Reproductive number

Abstract

Understanding COVID-19 transmission dynamics, such as the temporal trend of the transmission rate, i.e., time-varying effective reproductive numbers (Rt), is important for controlling the disease, estimating the prevalence and the total number of infections, and the fatality. Modeling COVID-19 transmission dynamics faces significant challenges. A large number of cases are likely to be un-ascertained/undetected due to insufficient testing. Many of these un-ascertained cases are asymptomatic and mildly symptomatic but still infectious. For a given region, the unknown ascertainment rate often varies over time due to evolving testing capacity. Furthermore, case counts are often subject to delayed reporting. Epidemic models that ignore unascertained cases and reporting delays can lead to biased estimates of the transmission rate and the prevalence of the disease. To address these problems, we develop an over-dispersed Poisson Partial Differential Equation transmission dynamic model to estimate the region-specific Rt and ascertainment rates as piece-wise constants by (1) using daily case counts of positive tests and accounting for both reporting delay and over-dispersion; (2) incorporati