Title: Causal inference for clinical trials with a time-to-event outcome

Guest Speaker: Gang Li, Ph.D.
                           Professor of Biostatistics
                           Director, UCLA's Jonsson Comprehensive Cancer Center BASE Unit
                           University of California at Los Angeles

Time: 10:30-11:30 a.m., Tuesday, 12 January 2016

Venue: D405

Abstract of the talk:
To estimate the biological treatment effect in a randomized clinical trial, one often has to conduct subgroup analysis on a latent subgroup of subjects. Latency is induced because subgroup membership is observable in one arm of the trial and unidentified in the other. It typically occurs in clinical trials with all-or-none noncompliance when patients in the control arm have no access to active treatment and in oncology trials when a test used to identify the latent subgroup is performed only on subjects randomized to active treatment. In this talk, I will discuss recent developments for latent subgroup analysis with right censored time-to-event data. Illustrations will be given through a multicenter selective lymphadenectomy trial for melanoma.