Extensions of the 2-in-1 adaptive design

Abstract

The 2-in-1 adaptive design [Chen et al. 2018] allows seamless expansion of an ongoing Phase 2 trial into a Phase 3 trial to expedite a drug development program. An intermediate endpoint can be used for the adaptive decision. Under a mild assumption that is expected to generally hold in practice, both the Phase 2 trial (in case of no expansion) and the Phase 3 trial (in case of expansion) can be tested at the full alpha level without inflating the overall Type I error of the study. Due to its flexibility and robustness, the design has quickly generated a lot of interest since its recent publication.We make three major extensions to the 2-in-1 design in this paper: 1) an increase of adaptive decisions from two to any number; 2) incorporation of the group sequential method for data monitoring; 3) expansion of the univariate design to multivariate. These extensions not only facilitate the application of 2-in-1 design in practice but also stimulate research interest in related statistical issues. Although we mainly use hypothetical trials in oncology and neuroscience for illustration, the application of the 2-in-1 design and its extensions is not limited to the two therapeutic areas.