Sample size calculation for paired survival data: A simulation method

This paper proposes a sample size calculation method for rank tests for paired two-sample survival data using simulation method. A paired exponential survival model is specified by the marginal hazard rates and the correlation coefficient. Input variables include type I and II error probabilities, marginal hazard rates under the alternative hypothesis, accrual rate and follow-up period. Using efficient algorithms, our simulation method provides required sample size within minutes. The method is easily extended to sample size calculation for K-sample tests under dependence.

Duke Scholars

Author Sin-Ho Jung Biostatistics & Bioinformatics, Division of Integrative Geno ...