Decisionmaking in Post Clinical Trials

• Activity Number: 107
• Type: Invited
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Committee on Applied Statisticians
• Author(s): Heping Zhang*+
• Companies: Yale University
• Keywords: decision tree ; clinical trials ; treatment effect ; Kullback-Leibler divergence

Abstract

Controlled randomized clinical trials are the standard design to compare the effectiveness of treatments. In those trials, the effects of the treatments are compared among study groups in terms of the within-group average effects. It is possible that the most effective treatment based on the average effect is not as effective as other treatments in particular groups of patients. Identifying such potential subgroups is important for making clinical decisions. We present the decision tree approach to answering this important question. Specifically, we employ the Kullback-Leibler divergence as the node splitting criterion, and determine the tree size according to whether the same clinical decision is reached.