Estimating Surrogate Endpoints Defined by Principal Causal Effects

• Activity Number: 234
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Shiro Tanaka*+ and Yutaka Matsuyama and Yasuo Ohashi
• Companies: University of Tokyo and University of Tokyo and University of Tokyo
• Keywords: advanced prostate cancer ; monotonicity assumption ; principal stratification ; surrogate endpoint

Abstract

In clinical trials, comparison of treatments for the outcome of primary interest such as survival time may require a long follow-up before yielding useful results. Because of this, there is increasing interest in the use of surrogate endpoints to make decisions about treatment efficacy. Frangakis and Rubin (2002) proposed a principal surrogate, which was defined by potential values of surrogate endpoints. However, it is impossible to observe joint distribution of potential surrogate endpoints without untestable assumptions. In this talk, we propose an estimation method of the principal surrogate under monotonicity assumption, which states that there is no subject who would not have a response under test arm, but have a response under control arm. We also applied the proposed method to a clinical trial data of advanced prostate cancer.