Utilizing shared internal control arms and historical information in small-sized platform clinical trials

Area of Science:

• Clinical Trials
• Drug Development
• Biostatistics

Background:

• Patient recruitment for concurrent control arms in pediatric and rare disease clinical trials is challenging.
• Platform trial designs with shared internal control arms can enhance efficiency and speed drug development.
• Historical borrowing, using data from prior trials, offers further improvements in clinical trial efficiency.

Purpose of the Study:

• To discuss the advantages and limitations of platform trial designs.
• To compare various historical data borrowing strategies.
• To propose and evaluate a method for using historical controls in platform trials.

Main Methods:

• Discussion of platform trial designs and historical borrowing strategies.
• Comparison of data pooling, separate analysis, test-then-pool, dynamic pooling, and Bayesian hierarchical modeling (MAP prior).
• Proposal of a procedure for utilizing historical controls in a platform setting, evaluated through simulations.

Main Results:

• Platform trials and historical borrowing offer potential solutions to recruitment challenges.
• Various historical data borrowing methods have different efficiencies and applications.
• Simulations demonstrate the statistical performance and feasibility of the proposed method.

Conclusions:

• Platform trials combined with historical borrowing can significantly improve clinical trial efficiency, especially for rare diseases.
• The proposed method provides a feasible approach for integrating historical controls into platform trials.
• These innovative designs accelerate the drug development process by optimizing resource utilization.