Keywords: Sample Size Estimation, Survival endpoint, Crossover, Conditional power, Software
Sample size estimation is an important step in designing a clinical study. When all the standard required information is available, sample size estimation is straight forward either using the software (Proc Power, nQuery, PASS, EAST) or using the sample size formula directly. Often the standard required information is not available and the sample size estimation can be complicated for non-standard cases since more estimations/calculations need to be done before you can use the software. Sample size calculation requires the collaboration of statisticians and clinical/medical team.
We will share our experience for the sample size estimations for special survival endpoint, cross-over designs, and also the sample size re-estimation based on conditional power. All participants can share their experiences and the challenges they have faced in sample size estimations for non-standard cases including implementations, solutions and software.