Three-outcome designs for external pilot trials with progression criteria

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmaceutical Research

Background:

• Pilot studies are crucial for informing decisions on progressing to definitive trials.
• Progression criteria and pilot sample size determination lack comprehensive research.
• Existing methods often rely on pre-specified quantitative progression criteria with three possible outcomes.

Purpose of the Study:

• To extend three-outcome trial designs to external pilot studies.
• To propose a unified framework for optimizing pilot trial progression criteria and sample size.
• To compare the efficiency of three-outcome designs against two-outcome alternatives.

Main Methods:

• Reviewing three-outcome designs from phase II oncology.
• Developing a unified framework based on univariate hypothesis tests and frequentist error rate control.
• Applying the framework to an example and comparing with a two-outcome design.

Main Results:

• Three-outcome designs are applicable to pilot studies but not always more efficient than two-outcome designs.
• They allow for incorporating additional information for borderline results, albeit with larger pilot sample sizes.
• These designs can facilitate pre-definitive trial adjustments if effects are predictable.

Conclusions:

• The proposed three-outcome framework optimizes pilot trial progression criteria and sample size for desired operating characteristics.
• It is applicable with or without anticipated post-pilot adjustments.
• Generally, three-outcome designs necessitate larger sample sizes than two-outcome alternatives.