Causal Analysis in Multi-Arm Trials with Truncation by Death

• Activity Number: 523
• Type: Topic Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Linbo Wang* and Thomas S. Richardson and Xiao-Hua Zhou
• Companies: University of Washington and University of Washington and University of Washington
• Keywords: Bayesian analysis ; Causal inference ; Multi-arm trials ; Principal stratification ; Truncation by death

Abstract

It is common that in randomized trials with multiple treatment arms, the outcome of interest is "truncated by death", meaning that it is only observed or well-defined conditioning on an intermediate outcome. In this case, the contrasts between different treatment arms are of simultaneous interest. In this presentation, we propose methods for the simultaneous estimation and testing of causal contrasts in multi-arm trials with truncation by death. We show that our methods provide additional insights compared to existing methods that consider these causal contrasts separately. We illustrate our methods with a dataset from a randomized HIV vaccine trial.