D-optimal designs for multiarm trials with dropouts.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Longitudinal Data Analysis

Background:

• Multiarm trials are essential for assessing treatment effects over time.
• Participant dropout is a common challenge in long-duration studies, potentially biasing results.
• Standard trial designs often neglect dropout's impact during the planning phase.

Purpose of the Study:

• To develop an optimal design framework for multiarm trials that explicitly incorporates potential participant dropout.
• To provide robust designs for linear mixed models accommodating noninformative and design-dependent dropouts.
• To demonstrate the advantages of the proposed framework through a case study in Alzheimer's disease research.

Main Methods:

• Development of a novel optimal design framework for longitudinal multiarm trials.
• Integration of dropout considerations into the trial design process.
• Application and simulation-based evaluation of the framework using a clinical trial for Alzheimer's disease.

Main Results:

• The proposed framework offers improved statistical accuracy by accounting for dropouts.
• Simulations show benefits of the new design over standard approaches in Alzheimer's trial redesign.
• The framework provides effective designs for linear mixed models with various dropout patterns.

Conclusions:

• Optimal trial design must proactively address participant dropout for reliable conclusions.
• The developed framework enhances the precision and validity of statistical inferences in long-term multiarm studies.
• This methodology is particularly valuable for clinical trials with extended follow-up periods, such as those for Alzheimer's disease.