The Balanced Survivor Average Causal Effect

• Activity Number: 283
• Type: Invited
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Committee on Applied Statisticians
• Author(s): Tom Greene*+
• Companies: University of Utah
• Keywords: longitudinal analysis ; truncation by death ; principal stratification

Abstract

Statistical analysis of longitudinal outcomes is often complicated by the absence of observable values in patients who die prior to their measurement. Recently, this problem has been investigated using the framework of principal stratification to define the target estimand as the survivor average causal effect (SACE), which in the context of a 2-group randomized clinical trial is the mean difference in the longitudinal outcome between the treatment and control for the principal stratum of always-survivors. The SACE is often estimated under a monotonicity constraint requiring that the treatment does not reduce survival in any patient, which introduces an asymmetry between the randomized groups. I will introduce an alternative estimand, the balanced-SACE, which is motivated using rank-preserving transformations and is balanced with respect to the potential survival times under the treatment and control. I consider estimators of the balanced-SACE which do not require monotonicity, and examine expressions for the bias of the estimators and strategies for sensitivity analysis when the rank preserving assumption is violated.