Estimation of a parameter and its exact confidence interval following sequential sample size reestimation trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmaceutical Research

Background:

• Adaptive designs are crucial for regulatory decision-making in clinical trials.
• Naive point estimates and confidence intervals in adaptive sequential designs can be biased.
• Accurate statistical inference is essential for confirmatory trials.

Purpose of the Study:

• To develop a novel procedure for accurate estimation following tests in sample size reestimation designs.
• To provide unbiased point estimates and exact confidence intervals for treatment comparisons in adaptive trials.
• To improve the power of statistical tests in adaptive group sequential trials.

Main Methods:

• Utilized a general distribution property of a pivot function from Shen and Fisher's Self-designing group sequential clinical trial.
• Developed a modified estimation procedure to account for futility stopping boundaries, reducing bias with small block sizes.
• Introduced a modified weight function to enhance statistical test power.

Main Results:

• Proposed point estimates are consistent and nearly unbiased with practical sample sizes.
• Exact confidence intervals demonstrate accurate nominal probability of coverage.
• The modified weight function improves the power of the statistical test.

Conclusions:

• The new procedure provides accurate and unbiased estimates and confidence intervals for adaptive group sequential trials.
• The method ensures correct statistical inference crucial for regulatory decision-making.
• The developed techniques are computationally straightforward and improve trial efficiency.