Assessing Methods for Dealing with Crossover in Active-Control Trials

• Activity Number: 177
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jihong Chen*+ and Jay Yang and Xiaosha Sherman Zhang and Andrew Strahs
• Companies: Astellas and Astellas Pharma Global Development Inc. and Astellas Pharma Global Development, Inc. and AVEO
• Keywords: overall survival ; crossover ; RPSFT

Abstract

In some randomized Oncology trials, patients on the control arm, after disease progression, have the opportunity to cross over to the experimental arm. A high percentage of such one-way crossover could lead to biased OS estimate when using an ITT analysis. In the literature, various methods have been used to adjust for the bias caused by crossover. These methods were discussed and developed in the setting that the control arm uses placebo. We're studying a new situation where an active drug is used as control. With crossover, patients on the control arm receive two active drugs while patients on the experimental arm only receive one active drug. We conducted a simulation study to assess the impact of crossover and evaluate the performance of existing methods under this situation. Moreover, we generalized the Rank Preserving Structural Failure Time Model (RPSFT) to a two parameter setting to handle this scenario.