Multiple-objective response-adaptive repeated measurement designs in clinical trials for binary responses

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Modeling

Background:

• Response-adaptive designs optimize treatment allocation during clinical trials.
• Previous methods focused on continuous response variables.
• Adapting these designs for binary outcomes presents unique statistical challenges.

Purpose of the Study:

• To extend a multiple-objective allocation strategy for response-adaptive repeated measurement designs to binary responses.
• To investigate the performance of this strategy in the context of nonlinear modeling and information matrices dependent on responses.
• To evaluate the trade-offs between treatment benefit and estimation precision for binary outcomes.

Main Methods:

• Developed a design framework based on success probabilities for binary responses.
• Incorporated various statistical models to handle carryover effects using logits of response profiles.
• Utilized computer simulations to assess the allocation strategy's efficiency and performance.

Main Results:

• The allocation strategy effective for continuous responses also performs well for binary responses.
• Design efficiency, measured by mean squared error, decreases when prioritizing treatment benefit over estimation precision.
• The method successfully allocates more participants to more effective treatment sequences with minimal loss in precision.

Conclusions:

• The proposed response-adaptive allocation strategy is a viable and effective method for binary repeated measurement designs.
• This approach offers a practical way to balance maximizing treatment benefit with maintaining statistical rigor in clinical trials.
• The findings support the use of these adaptive designs for binary outcomes in optimizing clinical trial efficiency and patient outcomes.