Yang Yang Department of Statistics University of Georgia Speaker's Website Thu, 10/05/2023 - 4:00pm 204 Caldwell Hall Conditional likelihood – the key to analyzing infectious disease transmission in close contact groups Abstract Close contacts groups such as households, schools and hospitals are ideal venues for understanding transmission characteristics of infectious diseases and for evaluating intervention effectiveness, due to the feasibility of tracking individual-level exposure history. However, observation of such transmission dynamics is often incomplete due to limited resources. For example, when outbreaks of an emerging pathogen are detected, only infected individuals are reported, leaving out uninfected individuals. In epidemiological studies, it is a standard practice to observe households with an index case (first infection), leaving out uninfected households. A more extreme example is outbreak surveillance data collected by CDC, where only large outbreaks with a certain number of cases were reported. Selection bias occurs not only in terms of who will be observed, but also in terms of when to observe. For example, we start observation from symptom onset instead of infection (left truncation) and stop before the transmission dynamic ends (right-censoring). The solutions to these challenges share a common methodological ground: conditional likelihood. About the Author Dr Yang’s research focuses on developing statistical methods for transmission dynamics of infectious diseases such as influenza, dengue, Ebola, COVID-19, HIV and cholera at both individual and population levels. These methods are designed to evaluate transmissibility of pathogens, efficacy of interventions, and effects of risk factors in various study settings. Examples are studies with case-ascertained follow-up, competing risks of infection from co-circulating pathogens or strains, and high-dimensional missing data. He has developed methods for assessing vaccine or antiviral efficacy when there was only a small number of laboratory-tested specimens. Dr Yang is also interested in modeling co-circulating of multiple pathogens with cross-reactivity, e.g., dengue viruses and enteroviruses, especially when the underlying exposure history is incompletely observed. In response to the ongoing COVID-19 pandemic, I have been engaged in multiple collaborative projects of analyzing contact-tracing data, case incidence data and lab-testing data. His long-term goals in methodological research are to integrate phylodynamics and AI methods into transmission modeling and to address surveillance bias.
