Incorporating External Information or Real-World Evidence into Adaptive Design with Sample Size Re-Estimation

• Activity Number: 511 - In Pursuit of Speed and Efficiency in New Drug Development: Statistical Designs at the Forefront of Pushing the Boundaries of Innovation
• Type: Topic Contributed
• Date/Time: Thursday, August 11, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Ran Liao* and Junjing Lin and Margaret Gamalo
• Companies: Eli Lilly and Company and Takeda Pharmaceuticals and Pfizer Inc.
• Keywords: Adaptive design; Real-world evidence; borrowing

Abstract

The diversity of patient population offers advantages in patient representation and is crucial for showing effectiveness and safety for a broad population in clinical trials. However, population heterogeneity is one of the fundamental challenges in designing clinical trials. Heterogeneity will also introduce extra variation, which may cause a problem in the analysis in terms of lowering signal to noise. Furthermore, subgroups may have differential response to the treatment, while pooled analysis may dilute the true treatment effect. An adaptive design provides a flexible solution to handle the uncertainty of patient population by enriching subgroup of trial participants who have a higher likelihood of benefiting from the new treatment. In this research work, we are exploring how adaptive design can be applied in a situation where part of the control arm can be replaced with external control from real-word data. A two-stage, adaptive design with sample size re-estimation (SSR) approach, in which SSR is implemented at interim analysis. Futility and efficacy stopping boundaries are pre-specified to guide the adaptation strategies proceeding to the second stage.