Optimal and Minimax Three-stage Designs for Phase II Clinical Trials Considering Response and Toxicity

摘要

The main objective of a phase II trial is to assess the activity of a treatment or drug. On the other hand, we may also need to consider its safety at the same time. Jin (2007) proposed a two-stage design that controls the type I errors of response and toxicity rate separately. In this article, we extend Jin's ideas to three-stage designs and propose optimal and minimax designs for three-stage phase II trials, which set the two type I errors at different levels. It shows that a lot could benefit from three-stage designs in contrast to two-stage designs.