Comparisons of global tests on intersection hypotheses and their application in matched parallel gatekeeping...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Inference

Background:

• Clinical trials frequently involve multiple endpoints and dose comparisons, leading to statistical multiplicity issues.
• Hierarchical hypothesis structures further complicate multiplicity adjustments in clinical trial analysis.
• Closed test procedures, including the Simes test and average method, are established approaches for managing multiplicity.

Purpose of the Study:

• To extend existing matched parallel gatekeeping procedures for clinical trial multiplicity.
• To integrate the average method into gatekeeping procedures, offering improved power over the Simes test in specific scenarios.
• To adapt these procedures for trials with multiple endpoints and multiple active doses.

Main Methods:

• Extension of the matched parallel gatekeeping procedure, originally based on the Simes test.
• Incorporation of the average method for testing intersection hypotheses, particularly beneficial when dose-response relationships are non-linear.
• Generalization of the procedure to accommodate more than two endpoints and more than two active doses.

Main Results:

• The enhanced gatekeeping procedure allows the use of the average method, increasing statistical power compared to the Simes test when treatment effects of different doses are similar.
• The methodology is successfully extended to handle complex trial designs with numerous endpoints and active treatment arms.
• Provides a more flexible and powerful framework for multiplicity adjustment in advanced clinical trial settings.

Conclusions:

• The proposed extension offers a statistically robust and more powerful method for multiplicity control in clinical trials.
• This approach is particularly advantageous for trials comparing multiple doses of an investigational drug against a control.
• The generalized procedure enhances the ability to detect treatment effects in complex clinical trial designs.