A Fresh View of the Efficiency of Sample Size Adaptive Designs

• Activity Number: 298 - Confirmatory Adaptive Designs in Action: Recent Advances in Statistical Frameworks and Operational Models
• Type: Topic Contributed
• Date/Time: Wednesday, August 5, 2020 : 10:00 AM to 11:50 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Lu Cui*
• Companies: UCB Biosciences
• Keywords: Adaptive design; Sample size adaptative design; Power; Efficiency; FDA guidance; Group sequential design

Abstract

After many years of efforts in understanding and implementing sample size adaptative design (SSAD) in clinical trials, the approach is gaining full attention. Recent FDA draft guidance (2018) on adaptive designs for industry sponsors clearly describes SSAD, targeting a range of plausible treatment effect sizes, as a method with enhanced statistical power and sample size efficiency. This presentation is to demonstrate the advantage of SSAD over the conventional group sequential design (GSD) based on the concept of ‘uniformly more efficient’ design (Cui and Zhang, Stat. in Med., 2019). In the presentation, we will introduce SSAD in its more flexible form, and show that GSD, in fact, is a special case of SSAD excluding the early accumulated data in the final analysis. Therefore, a better efficiency is expected to be achieved by more flexible SSAD. Indeed, as shown in our simulation examples, given a GSD, a more efficient SSAD can be constructed. The fresh view and ideal in constructing more efficient SSAD presented hereby may further help applications of SSAD to industry sponsored clinical trials.