Abstract:
|
Phase 3 trials in immuno-oncology can require consideration of multiple endpoints, biomarker-identified and overall populations as well as interim analyses. Formally controlling and explaining Type I error control in this setting can be a challenge. We will address extending the work of Maurer and Bretz (2013, SBR) using graphical methods and group sequential design to further justify use for time-to-event endpoints. General suggestions on strategies for interim analysis as well as Type I error allocation and reallocation will be considered, including examples from the literature.
|