A General Framework for Designing and Evaluating Active-Controlled Trials with Non-Inferiority Objectives.

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Public Health Interventions

Background:

• Active-controlled non-inferiority trials are crucial when effective treatments exist but new options are needed.
• Challenges include the lack of a placebo arm and the need to estimate active control effects from historical data.
• Highly effective active controls can make traditional non-inferiority criteria difficult to meet, necessitating alternative approaches.

Purpose of the Study:

• To propose a general framework for designing and evaluating non-inferiority trials.
• To integrate existing analytical methods and accommodate both traditional and alternative success criteria.
• To enable systematic comparison of methods regarding type I error, power, and robustness.

Main Methods:

• Developed a comprehensive framework for non-inferiority trial design and evaluation.
• Integrated traditional and alternative criteria based on sufficient efficacy relative to a hypothetical placebo.
• Systematically compared methods using simulations to assess type I error, power, and robustness to misspecification.

Main Results:

• The framework successfully identified methods offering greater efficiency and robustness in an HIV prevention trial example.
• Demonstrated practical advantages of alternative non-inferiority criteria when active controls are highly effective.
• Highlighted the framework's utility in comparing different analytical approaches.

Conclusions:

• The proposed framework provides a comprehensive toolkit for rigorous non-inferiority trial design.
• It supports informed method selection and robust evaluation of new interventions, especially with highly effective active controls.
• Facilitates the adoption of alternative non-inferiority criteria for improved trial efficiency and relevance.