Biomathematics Seminar Series - Richard Zhao (Jan. 20)

Speaker: Richard Zhao (McMaster University)

Title: Improving infectious disease prevalence estimation and parameter inference using number of tests and positivity data

Abstract: Prevalence of infection is a critical variable for modeling infectious dynamics and public health decision-making. However, estimating true prevalence from surveillance data remains challenging. Here we discuss three novel attempts to model testing mechanism using the beta-distribution, hazard ratios and odds ratios respectively. These methods aim to link prevalence with the number of tests, test positivity, and test characteristics at each data point in a robust, flexible, and theoretically justified manner. We further present a data-fitting framework based on the odds ratio approach and demonstrate its performance using simulated datasets as a proof of concept.