Flexible designs for phase II comparative clinical trials involving two response variables

Area of Science:

• Biostatistics
• Clinical Trial Design
• Medical Research

Background:

• Phase II clinical trials aim to assess treatment promise and safety before further testing.
• Reducing sample size in clinical trials is crucial due to technical and ethical considerations.
• Complex therapeutic schemes necessitate decisions based on multiple endpoints.

Purpose of the Study:

• To propose flexible designs for phase II clinical trials comparing two treatments with two dependent dichotomous responses.
• To enable early termination of trials based on observed favorable or unfavorable outcomes, thus reducing patient numbers.
• To provide a mathematical framework for sample size determination in such trials.

Main Methods:

• Designs are based on bivariate sequences of multi-state trials.
• Stopping rules utilize distributions related to waiting times for specific event counts.
• Markov chain embedding technique is employed for analyzing exact distributions and determining sample size.

Main Results:

• The proposed designs allow for early trial termination, significantly reducing the number of patients required.
• Expected sample size and power tables are provided for the new designs.
• Numerical illustrations demonstrate the strong performance of these flexible designs.

Conclusions:

• The developed flexible designs offer an efficient approach for phase II clinical trials with dual dichotomous endpoints.
• These designs address the need for reduced sample sizes while maintaining statistical rigor.
• The Markov chain embedding technique provides a unified framework for sample size determination and performance evaluation.