An efficient sample size adaptation strategy with adjustment of randomization ratio.

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Statistical Power

Background:

• Sample size reestimation is crucial in clinical trials to address design uncertainties and ensure adequate statistical power.
• Current methods often maintain proportional increases in sample size, which may not be optimal when initial randomization ratios are not ideal.

Purpose of the Study:

• To propose and evaluate a novel sample size reestimation strategy: nonproportional increase.
• To enhance statistical power by adaptively adjusting the randomization ratio during sample size increases.
• To ensure type I error rate control under the proposed nonproportional increase method.

Main Methods:

• Development of an adaptive randomization ratio change strategy for sample size reestimation.
• Analytical derivation to demonstrate type I error rate control.
• Monte Carlo simulations to validate theoretical findings and assess power gains.

Main Results:

• The nonproportional increase strategy boosts statistical power beyond simple sample size augmentation.
• Efficient allocation of additional subjects to optimize the randomization ratio is demonstrated.
• Analytical and simulation results confirm the control of the type I error rate.

Conclusions:

• Nonproportional sample size reestimation offers an efficient approach to increase statistical power in clinical trials.
• This adaptive strategy is particularly beneficial when initial randomization ratios deviate from optimal.
• The method provides a statistically sound way to improve trial efficiency and success probability.