Adaptive multiarm multistage clinical trials.

Area of Science:

• Clinical Trial Design
• Biostatistics
• Pharmaceutical Research

Background:

• Adaptive multiarm multistage (MAMS) trials allow for early stopping of ineffective treatments and sample size adjustments.
• Two distinct group sequential frameworks exist for MAMS trial design: stage-wise and cumulative.
• Both methods ensure strong control of the family-wise error rate while adapting trial parameters.

Purpose of the Study:

• To present two methods for designing adaptive multiarm multistage (MAMS) clinical trials.
• To compare the operating characteristics, specifically power, of the stage-wise and cumulative MAMS approaches.
• To evaluate these methods under various treatment selection, sample size reestimation, and early stopping rules.

Main Methods:

• The stage-wise MAMS approach uses multiplicity adjusted P-values at each stage, combined via a function and monitored against group sequential boundaries.
• The cumulative MAMS approach constructs separate test statistics from cumulative data at each stage, monitored against multiplicity adjusted boundaries.
• Power comparisons were conducted for designs with two and three active treatment arms across a wide parameter space.

Main Results:

• Cumulative MAMS designs demonstrated greater statistical power compared to stage-wise MAMS designs in most scenarios.
• A small power advantage was observed for stage-wise MAMS designs only in the specific case of homogeneous treatment effects (equal effects).
• Both methods successfully maintained strong control of the family-wise error rate despite adaptive modifications.

Conclusions:

• Cumulative MAMS designs offer a more powerful approach for adaptive clinical trials compared to the traditional stage-wise method.
• The choice between stage-wise and cumulative MAMS may depend on the expected homogeneity of treatment effects.
• These findings inform the selection of optimal MAMS designs for efficient clinical trial conduct.