The use of the triangular test with response-adaptive treatment allocation

Area of Science:

• Clinical Trials
• Biostatistics
• Medical Research

Background:

• Sequential clinical trials often employ stopping rules like the triangular test to compare treatments.
• Response-adaptive randomization aims to allocate more patients to the better-performing treatment during a trial.

Purpose of the Study:

• To evaluate the impact of adaptive treatment allocation rules on the performance of the triangular test.
• To compare the drop-the-loser rule with other adaptive designs in conjunction with the triangular test.

Main Methods:

• The study analyzes the triangular test combined with various response-adaptive allocation rules, including the drop-the-loser rule.
• Simulation methods are used to assess the performance metrics such as allocation proportion variability and treatment failures.
• Asymptotic properties of the adaptive rules are summarized.

Main Results:

• The drop-the-loser rule demonstrated a less variable allocation proportion and fewer treatment failures compared to other adaptive rules.
• Combining the triangular test with the drop-the-loser rule resulted in fewer average treatment failures than equal allocation.
• The adaptive approach provided comparable statistical power and similar expected sample sizes.

Conclusions:

• The drop-the-loser rule is a promising adaptive design for use with the triangular test in clinical trials.
• This combination offers benefits over equal allocation by reducing treatment failures while maintaining trial efficiency.
• An AIDS clinical trial example illustrates the practical application and benefits of this approach.