A Comparison of Methods for Treatment Selection in Seamless Phase II/III Design

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Pharmaceutical Research

Background:

• Seamless phase II/III designs combine treatment selection and confirmatory phases for efficiency.
• These adaptive designs allow mid-trial modifications to optimize resource use and reduce patient burden.

Purpose of the Study:

• To comprehensively evaluate statistical approaches for seamless phase II/III trials.
• To assess methods for controlling Type I error rates and maintaining statistical power in adaptive designs.

Main Methods:

• Evaluation of four classes of statistical approaches integrating combination functions and multiple testing methods.
• Extensive simulation studies to assess Type I error control and statistical power.
• Illustration of practical implementation in real clinical trial settings.

Main Results:

• The study provides a detailed comparison of different statistical methods for seamless phase II/III trials.
• Simulation results demonstrate the performance of various approaches in terms of error control and power.
• Practical examples showcase the applicability of these methods in clinical practice.

Conclusions:

• Appropriate statistical methods are crucial for the valid analysis of seamless phase II/III trials.
• The evaluated approaches offer robust control of Type I error rates under different treatment selection rules.
• This work aids in the practical implementation and statistical rigor of adaptive clinical trial designs.