Extension of Rank Preserving Structural Failure Time Model and Its Application in Oncology Clinical Trials

• Activity Number: 488
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Ling Wang and Shijie Tang*+ and Connie Lee and Yuanjun Shi
• Companies: Takeda and Infinity Pharmaceuticals and Takeda and Takeda
• Keywords: Oncology ; Clinical Trial ; Crossover ; Survival

Abstract

In oncology clinical trials, it is common for patients continue to receive effective subsequent therapies or be allowed to crossover to active treatment after disease progression. As a result, the estimate of treatment effects for overall survival using the intent to treat principle may be biased. Statistical methods such as inverse probability censoring weighting (IPCW) and rank preserving structural failure time (RPSFT) model have been proposed to better estimate the true treatment effect by adjusting the effect of subsequent therapies or the crossover to active treatment. The original RPSFT method assumes that only patients from the placebo arm will crossover to the active treatment. We proposed an extension to the original RPSFT method to allow the modeling of survival time due to subsequent therapies for both placebo and active arms. All these methods will be illustrated in a real oncology trial example. Results will be discussed in our presentation.