Drug Evaluation Committee Study of Simulation for Planning Adaptive Design
Data Science Subcommittee
February 2025
A study on adaptive design is currently underway at ICH-E20. While the flexibility afforded by adaptive design is expected to increase the opportunity for trial participants to receive better treatment and more efficient drug development, as well as to make better use of available resources, it is also expected to increase the probability of type 1 errors, bias in point estimates, and the probability of coverage of confidence intervals if statistical methods are not used appropriately. The use of appropriate statistical methods can lead to undesirable phenomena in terms of statistical validity, such as an increase in the probability of the first type of error, bias in point estimates, and a decrease in the coverage probability of the confidence interval compared to the nominal confidence coefficient.
When actually planning an adaptive design, it is important to evaluate the behavioral characteristics of the design at the planning stage, and the behavioral characteristics will be evaluated through simulations in hypothetical situations. This deliverable summarizes the methods and procedures for simulation to evaluate the behavioral characteristics of a study design when planning an adaptive design, as well as the indicators to be evaluated in the simulation and how to report the results.
The intended audience for this deliverable is biostatisticians involved in planning and conducting clinical trials, clinical members planning clinical trials, and regulatory affairs members responsible for regulatory affairs.
We hope that this publication will help you understand adaptive design and conduct trials appropriately.
Japan Pharmaceutical Manufacturers Association, Committee on Drug Evaluation
Data Science Subcommittee FY2023 Ongoing Task Force 1
Study of Simulation for Planning Adaptive Design
Deliverables of the Drug Evaluation Committee Return to list of all deliverables