Monday, April 17, 2023, 12 – 1pm

In observational studies, potential confounders may distort the causal relationship between an exposure and an outcome. However, under some conditions, a causal dose-response curve can be recovered by using the G-computation formula. When the exposure is continuous-valued, nonparametric inference on this curve can be challenging since it cannot be estimated at regular rates. We show that if the causal dose-response curve is known to be monotone, asymptotically valid inference can be carried out using causal isotonic regression, an extension of classical isotonic regression. We discuss the statistical properties and implementation of causal isotonic regression. We also discuss extensions of causal isotonic regression for use when the data-generating mechanism includes various sampling complications. These extensions are critical for contemporary applications, including the assessment of correlates of protection in vaccine trials. We illustrate our proposed methods by studying controlled vaccine efficacy curves using data from COVID vaccine trials.

—

Marco Carone is an Associate Professor of Biostatistics and Adjunct Associate Professor of Statistics at the University of Washington, and an Affiliate Investigator in the Vaccine and Infectious Disease Division at the Fred Hutchinson Cancer Research Center. He is also the Norman Breslow Endowed Faculty Fellow. He completed his doctoral studies in Biostatistics at the Johns Hopkins Bloomberg School of Public Health under the guidance of Daniel Scharfstein. Before joining UW, he was a postdoctoral fellow and lecturer at UC Berkeley, where he worked with Mark van der Laan. His research centers on causal inference, high-dimensional statistics, survival analysis, longitudinal data, and more. 

Zoom Participation. See announcement.

Event Type: Seminars
Room Number: Virtual Presentation - ET
Building: Remote Access -Zoom
Speaker's Name: MARCO CARONE
Speaker Website: faculty.washington.edu…
Speaker's Professional Title: Associate Professor of Biostatistics, Adjunct Associate Professor of Statistics, University of Washington, Affiliate Investigator, Fred Hutchinson Cancer Research Center
Talk Title: Causal isotonic regression: theory and practice
For More Information: edward@stat.cmu.edu
Affiliations: Computer Science Department (CSD), Machine Learning Department (MLD)
Organization(s): Department of Statistics & Data Science