Event details:
Date: April 22nd 2026
Time: 7Pm CST
Location: / Virtual Link: https://stevens.zoom.us/recording/detail?meeting_id=uZ5mWxI3Tvmf4wKRuU39og%3D%3D
Abstract
Modern confirmatory clinical trials often evaluate multiple treatment arms, test several hypotheses, and assess effects within clinically relevant subgroups-frequently incorporating interim analyses or adaptive features. These designs create complex multiplicity challenges, where control of Type I error and power is typically evaluated using simulation. While flexible, simulation-based approaches can be time-intensive and difficult to align with complex pre-specified testing strategies.
This talk introduces the statistical theory underlying joint distributions of hypothesis test statistics and shows how distributional and correlation structure can be leveraged to greatly reduce reliance on simulation. Key components include multiple treatment comparisons against a common control, sequential testing across stages, and distributional theory for subgroup-specific and overall effects. A concluding case study demonstrates how theory-informed approximations can guide and validate simulation studies under realistic multiplicity plans.
Speaker Bio
Jason Clark is Associate Vice President, Late Stage Statistics at Merck Research Laboratories, where he leads late-stage oncology development for Bladder, Prostate, Women's Cancer, Melanoma, and Cancer Innovative Strategies. He has over 25 years of experience in drug development, with previous roles at AstraZeneca, Incyte, and Johnson & Johnson.
------------------------------
Charu Gupta
AstraZeneca
------------------------------