Optimality criteria for futility stopping boundaries for group sequential designs with a continuous endpoint

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Statistical Inference

Background:

• Group sequential designs permit early trial cessation for efficacy or futility, unlike fixed designs.
• Futility boundaries are crucial for stopping trials unlikely to succeed but often lack theoretical justification.
• Existing methods for optimal futility boundaries primarily address time-to-event endpoints.

Purpose of the Study:

• To extend the concept of optimal futility boundaries to clinical trials with a single continuous primary endpoint.
• To develop an analytic algorithm for deriving these optimal futility boundaries.
• To evaluate the performance of study designs utilizing optimal futility boundaries.

Main Methods:

• Extended Schüler et al.'s optimality criteria for futility boundaries to continuous endpoints.
• Developed an analytic algorithm to derive "optimal" futility boundaries.
• Applied the new concept to a real clinical trial and compared its performance against designs with arbitrary futility boundaries.

Main Results:

• The proposed method controls the probability of premature stopping for futility when a treatment effect is promising.
• The approach effectively manages the loss of statistical power.
• Optimal futility boundaries enhance the accuracy of stopping for futility under the null hypothesis.

Conclusions:

• Thorough investigation of futility boundaries during trial planning is essential.
• Arbitrary futility boundary selection can negatively impact trial performance.
• Employing "optimal" futility boundaries with predefined criteria enhances the efficiency of group sequential designs.