Presenter: David Benkeser, PhD
Assistant Professor
Department of Biostatistics and Bioinformatics
Emory University

Presentation Date: September 20, 2022

Title: Statistical learning for understanding whether and how vaccines work

Abstract: The advent of safe and effective vaccines is unambiguously one of the greatest public health achievements in history. The stunning success of preventive COVID-19 vaccines has again highlighted how important vaccine science is to global health. However, in spite of the overall success of the development effort, important questions have arisen along the way: why did it take so long to have a vaccine approved for children? When and how often should we be boosted? Will the original vaccines SARS-CoV2 remain effective with the emergence of new strains? Should the strain used in the vaccine be updated? If it should be updated, to what? And moreover, what regulatory framework should be used to guide these decisions? Answering these questions requires an understanding of the protective mechanisms of vaccines. Mechanisms of vaccines can be elucidated using statistical techniques arising from two closely related fields: (i) vaccine sieve analysis, which seeks to understand whether and how pathogen’s genotypic and phenotypic characteristics influence vaccine efficacy; and (ii) vaccine correlates of protection, which seeks to understand the particular immune mechanisms that lead to highly effective vaccines. In this talk, I will describe how modern statistical techniques in causal inference and machine learning can be used to elucidate and answer these questions.