An Extended Framework of Multiple Testing in Group Sequential Design.

Area of Science:

• Biostatistics
• Clinical Trials Methodology
• Statistical Inference

Background:

• Group sequential trials are essential for adaptive clinical research, allowing early stopping for efficacy or futility.
• Existing methods for multiple hypothesis testing in these trials face challenges in maintaining statistical power and error control.
• Graphical methods offer a structured approach to managing multiple comparisons in sequential settings.

Purpose of the Study:

• To extend existing graph-based group sequential procedures for multiple hypothesis testing.
• To introduce a modified weighted Simes test within a graphical framework for group sequential trials.
• To ensure the proposed procedure effectively preserves the familywise error rate.

Main Methods:

• Development of a novel graph-based group sequential procedure.
• Application of a modified weighted Simes test to control for multiple comparisons.
• Utilizing simulations to evaluate the performance and power of the proposed method.
• Illustrating the procedure with a practical numerical example.

Main Results:

• The proposed modified weighted Simes test procedure effectively preserves the familywise error rate.
• Simulations indicate favorable performance characteristics compared to existing methods.
• The numerical example demonstrates the practical applicability of the new procedure.

Conclusions:

• The enhanced graph-based group sequential procedure offers a robust approach for multiple hypothesis testing in clinical trials.
• This method provides a valuable tool for researchers seeking to maintain statistical rigor while adapting trial conduct.
• The modified weighted Simes test represents a significant advancement in sequential statistical methodologies.