Optimal Two-Stage Phase II Designs with Long-Term Endpoints

Phase II trials are often designed with an interim analysis so they can be stopped early if a drug is ineffective. One problem with interim analyses is incomplete follow-up for some patients. Standard designs such as Simon (1989) require suspension of accrual until patient follow-up is completed. Case and Morgan (2003) proposed a two-stage design for a Phase II oncology trial with a long-term endpoint that does not suspend accrual when the interim analysis is conducted. We review the Case and Morgan design and propose modifications to ensure protection of Type I error rate and improved robustness to nonoptimal conditions likely to be encountered in practice. Software (an R package) is described to create the optimal designs and easily simulate their properties to check asymptotic approximations and the robustness of the performance of the designs in operational conditions.