Assessing a Treatment Effect in Light of Rescue Therapy

• Activity Number: 308
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Judy Li*+ and Jerry John Weaver and David I. Ohlssen
• Companies: FDA and Novartis Pharmaceuticals Corporation and Novartis Pharmaceuticals Corporation
• Keywords: rescue therapy ; pattern mixture model ; missing data

Abstract

When clinical trials allow for the use of rescue therapy on ethical grounds, assessing the true treatment effect becomes challenging. Rather than restricting the treatment effect to the traditional "intention-to-treat" definition, we define the true treatment effect to be the one that would have been obtained had rescue therapy not been allowed. The main difficulty in estimating this quantity is that rescue therapy is a non-randomized intervention, and therefore could be related to the randomized treatment assignments. One possible approach is to take into consideration the proportion of patients having certain patterns of rescue therapy use, and appropriately apply them as weights in the pre-specified analysis method. This approach is akin to pattern mixture modeling, which is often used to address missing data that is non-ignorable. In this talk, we will discuss the application of the pattern mixture model in relationship to a clinical trial that incorporated rescue therapy by using simulated data sets to assess the true treatment effect for the primary endpoint.