Wednesday, Aug 7: 8:30 AM - 10:20 AM
Topic-Contributed Paper Session
Oregon Convention Center
Innovation in clinical trial design may not be enabled by off-the shelf software. This talk will focus on group sequential design with variations such as testing for multiple hypotheses, design for possibly delayed treatment effects and stratified design with differing outcome distributions and treatment effects in different strata. Examples of trials with biomarker/histology subgroups and also multiple experimental treatment groups will be presented. We discuss software specification, testing, issue management and release strategies. Specific trials developed include gsDesign (group sequential design), gsDesign2 (group sequential design with more options, including non-proportional hazards), multiple testing in group sequential design (gMCPLite and WPGSD). Also, we discuss the use of Shiny to enable design without the need to program in R.