A Bayesian platform trial design with hybrid control based on multisource exchangeability modelling.

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Rare Disease Research

Background:

• Enrolling patients to standard of care (SOC) arms in clinical trials, particularly for rare diseases, faces significant challenges including limited resources, patient availability, and ethical concerns.
• Historical trials often provide well-documented evidence of SOC therapeutic effects, presenting an opportunity to leverage this data.

Purpose of the Study:

• To propose a novel Bayesian platform trial design with a hybrid control arm that utilizes historical control data.
• To employ the multisource exchangeability modeling (MEM) framework for efficient computation and formal evaluation of data source exchangeability.

Main Methods:

• The proposed design incorporates a Bayesian platform trial with a hybrid control arm.
• Multisource Exchangeability Modelling (MEM) is used to assess the compatibility of historical and concurrent data, informing data borrowing decisions.
• Extensive simulation studies were conducted to evaluate the performance of the proposed hybrid design.

Main Results:

• The proposed MEM-based hybrid design demonstrated a significant reduction in sample size for the internal control arm.
• The design effectively borrows more information compared to other Bayesian approaches when historical and internal data are compatible.
• MEM provides a computationally efficient method for evaluating exchangeability between different data sources.

Conclusions:

• The proposed Bayesian platform trial design with MEM offers an efficient and effective strategy for rare disease clinical trials by leveraging historical data.
• This approach can lead to substantial sample size reductions and improved data utilization in the internal control arm.
• The MEM framework facilitates informed data borrowing decisions, enhancing the reliability of trial results.