Evaluation of Sample Size for Noninferiority Survival Clinical Trials in Endocrine Therapies

• Activity Number: 275
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Naum M. Khutoryansky*+
• Companies: Novo Nordisk Inc.
• Keywords: survival analysis ; sample size calculations ; non-inferiority trials ; cardiovascular events

Abstract

The problem of the sample size and power determination for time-to-event trials has been studied extensively by many authors. Several methods to solve this problem have been presented under various assumptions regarding the survival, entry and censoring distributions, different approaches in formulating the null and alternative hypotheses. The results presented in some papers on sample size calculation for non-inferiority survival clinical trials were based on the simplified approach when the variance used in determination of the critical region was taken from the alternative hypothesis. The influence of such a simplification on the precision of the sample size calculation was not estimated. The goal of this talk is to compare the simplified approach and the precise approach (when the critical region is based only on the null hypothesis and Type I error). In addition, another approach is considered when the critical regions are determined by the confidence intervals calculated using observed hazards (as recommended in the FDA guidance on cardiovascular events). This talk will present practical examples in evaluation of long-term cardiovascular safety of diabetes therapies.