Power and sample size for randomized phase III survival trials under the Weibull model.

Two parametric tests are proposed for designing randomized two-arm phase III survival trials under the Weibull model. The properties of the two parametric tests are compared with the nonparametric log-rank test through simulation studies. Power and sample size formulas of the two parametric tests are derived. The sensitivity of sample size under misspecification of the Weibull shape parameter is also investigated. The study can be designed by planning the study duration and handling nonuniform entry and loss to follow-up under the Weibull model using either the proposed parametric tests or the well-known nonparametric log-rank test.