Revisit of test-then-pool methods and some practical considerations.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmaceutical Development

Background:

• The test-then-pool method enhances drug development efficiency by incorporating historical data.
• The original method's reliance on nonsignificant differences can lead to incorrect information borrowing, particularly with small sample sizes.
• This inconsistency poses a risk in clinical trials, potentially compromising results.

Purpose of the Study:

• To develop an equivalence-based test-then-pool method for continuous endpoints in drug development.
• To analyze the relationship between the original and equivalence-based test-then-pool methods.
• To provide methods for selecting equivalence margins and adjusting testing levels for robust statistical inference.

Main Methods:

• Development of an equivalence test for assessing consistency between historical and current trial data.
• Derivation of analytical formulas for Type I error and statistical power for both methods.
• Exploration of equivalence margin selection using overlap probability.

Main Results:

• The equivalence-based test-then-pool method offers a statistically sounder approach to data pooling.
• The study provides adjustments to control Type I error rates under true data consistency.
• Analytical derivations allow for precise evaluation of the performance of both test-then-pool methods.

Conclusions:

• The equivalence-based test-then-pool method is a more reliable alternative for borrowing historical information in drug development.
• This refined method improves the efficiency and accuracy of clinical trials, especially those with limited sample sizes.
• The findings offer practical guidance for implementing robust statistical strategies in pharmaceutical research.