Closed MCP-Mod for Pairwise Comparisons of Several Doses With a Control

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• The MCP-Mod approach is standard for dose-response analysis in clinical trials.
• The original MCP method detects dose-response signals but doesn't confirm individual dose efficacy.
• There's a need for methods that provide confirmatory p-values for both overall signals and specific dose comparisons.

Purpose of the Study:

• To extend the MCP-Mod approach for confirmatory p-value generation.
• To enable detection of dose-response signals and pairwise comparisons against placebo.
• To ensure robust statistical inference in dose-finding studies.

Main Methods:

• Application of the closed test principle (Marcus et al.) for hypothesis testing.
• Optimization of contrast coefficients under constraints for Type 1 error rate control.
• Utilizing a candidate set of plausible dose-response shapes during trial planning.

Main Results:

• The extended MCP-Mod approach yields confirmatory p-values for dose-response signals.
• The method allows for pairwise comparisons of individual doses against placebo.
• Optimized contrast coefficients ensure strong Type 1 error rate control at a pre-specified significance level.

Conclusions:

• The proposed extension of MCP-Mod provides a statistically sound framework for dose-response analysis in clinical trials.
• This method allows for more definitive conclusions regarding drug efficacy at specific doses.
• The approach is validated through a comprehensive simulation study, demonstrating its practical utility.