Some open problems in causal mediation pathway analyses

Seminar Series

Friday, October 8, 2021 - 12:00
Zoom
Peter Song, PhD

Abstract: Mediation analysis is a pervasive methodology in biomedical studies to help understand the mechanistic role of mediators as part of the exposure-response relationship.  Through the directed acyclic graphic (DAG) modeling, such an analysis has recently been extended for its capability within the field of causal inference.  The existing mediation analysis methodologies are inadequate to satisfy needs of practitioners; for example, the Sobel test is known to be underpowered on testing for mediation effects.  In this talk, I would like to first overview a few statistical problems in the context of causal mediation analysis that are of practical importance, and then present some preliminary solutions and insights for future work to solve these open problems. They include a new adaptive bootstrap method to overcome the conservatism in hypothesis testing for mediation effects, a new copula formulation for causal mediation effects in the presence of categorical mediators, and a mixed-integer programming approach to study high-dimensional mediation effects.  Motivating biomedical examples, as well numerical illustrations, will both be used in the presentation.

Speaker: Peter Song, PhD
Professor, Biostatistics
University of Michigan

Short Bio: Dr. Song is Professor of Biostatistics at the Department of Biostatistics, School of Public Health in the University of Michigan, Ann Arbor.  He received his PhD in Statistics from the University of British Columbia, Vancouver, Canada in 1996.  He has published over 190 peer-reviewed papers and graduated 22 PhD students.  Dr. Song's current research interests include data integration, distributed inference, high-dimensional data analysis, longitudinal data analysis, mediation analysis, spatiotemporal modeling, and precision health. He collaborates extensively with researchers from nutritional sciences, environmental health sciences, chronic diseases, and nephrology. He is IMS Fellow, ASA Fellow and Elected Member of the International Statistical Institute. Dr. Song now serves as Associate Editor of the Journal of American Statistical Association, the Canadian Journal of Statistics, and the Journal of Multivariate Analysis.

For Zoom information, please contact Terry Hales at terry.hales@duke.edu