Randomized two-stage Phase II clinical trial designs based on Barnard's exact test

Area of Science:

• Clinical Trials
• Biostatistics
• Oncology Research

Background:

• Two-stage clinical trial designs are preferred over one-stage designs in oncology.
• They allow for early termination due to lack of efficacy.
• This leads to significant sample size savings.

Purpose of the Study:

• To present exact two-stage designs utilizing Barnard's exact test for differences in proportions.
• To compare these new designs with existing methods (Kepner, 2010; Jung, 2010).
• To provide decision rules for trial planning under various scenarios.

Main Methods:

• Development of exact two-stage designs based on Barnard's exact test.
• Comparative analysis against established two-stage design methodologies.
• Generation of comprehensive tables for practical trial planning.

Main Results:

• The proposed exact two-stage designs demonstrate substantial sample size savings compared to existing methods.
• The use of Barnard's exact test provides a robust framework for early efficacy assessment.
• Decision rules are provided for various assumed underlying data distributions.

Conclusions:

• The presented exact two-stage designs offer an efficient alternative for clinical trials, particularly in oncology.
• Significant sample size reductions can be achieved, optimizing resource allocation.
• The methodology is recommended for its practical utility and statistical rigor in trial design.