December 14, 2023 Webinar
Adaptations and Type I Error Control in Platform Trials
Ekkehard Glimm, PhD
Abstract
In the wake of recent initiatives to speed up drug development, platform trials have received increased attention. Specifically, multi-sponsor trials, where multiple companies investigate their experimental treatments within a unified framework under a single master protocol, are appealing. However, these design also present statistical challenges. Typically, not all drugs enter the platform simultaneously, leading to (a) comparisons of non-concurrent treatments, (b) changing randomization ratios as new treatment arms enter or old ones are terminated, (c) the need to make decisions about efficacy and futility of treatments while the study is still ongoing.
Currently, there is a debate within pharmaceutical statistics about how to address these issues. This debate impacts areas such as type I error control, the necessity of limiting comparisons to concurrent controls or adjusting for time trends in treatment effects, and the acceptability of response-adaptive randomization in confirmatory clinical trials. In this presentation, we will provide an overview of recent discussions on these topics and explore how adaptive design methodology and multiple comparison procedures can be applied in the context of platform trials.
Short Bio
Ekkehard Glimm, PhD, is Senior Director in Biostatistics at Novartis Pharma in Basel, Switzerland. He has a PhD in mathematics from the University of Magdeburg in Germany on a topic from multivariate statistics. Ekkehard joined Novartis in 2005, working first in Oncology Biostatistics and in the Statistical Methodology group since 2006. Since 2021, he is also an adjunct professor for biostatistics at the Medical Faculty of the University of Magdeburg, Germany.
Since joining Novartis, Ekkehard's work has focused on adaptive clinical trial designs, methods for multiple endpoints in clinical trials and more recently platform and master protocol trials. He has authored and co-authored around 50 papers published in peer-reviewed scientific journals and is an associate editor of the Biometrical Journal. Since 2023, he is also one of the three editors-in-chief of Pharmaceutical Statistics.