A Useful Design via the Information Fraction in a Group Sequential Clinical Trial with Censored Survival Data

• Activity Number: 610
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Chen-Hsin Chen* and Chih-Yuan Hsu and Ken-Ning Hsu and Ya-Hung Lu
• Companies: Academia Sinica and Academia Sinica and Academia Sinica and Academia Sinica
• Keywords: alpha spending function ; censoring distribution ; entry distribution ; information time ; maximum duration trial

Abstract

Lan and DeMets proposed the alpha spending function for group sequential trials to enable the frequency and timing of interim analyses unspecified in advance during the design stage. Regarding a trial with censored time to the endpoint, they later defined the information time at an interim analysis in a maximum duration trial design. To compare two survival curves under such a design, information times for group sequential logrank test and Wilcoxon-type statistic have been developed by assuming the underlying survival distribution is exponential or Weibull without the censoring distribution. To cope with practical concerns, by assuming the censoring distribution dependent on the patient's recruited time given various entry distributions and extending the underlying survival distribution to be the generalized gamma distribution, we present a new scheme to determine information fractions for group sequential Harrington-Fleming tests, which cover the above-mentioned tests. We also calculate the expected numbers of events and sample size. The enhanced method facilitates design and analysis of group sequential trials with implemental flexibility.