Simultaneous confidence interval for assessing non-inferiority with assay sensitivity in a three-arm trial with...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Inference

Background:

• Three-arm trials are crucial for evaluating experimental treatments against active references and placebos.
• Assessing non-inferiority (NI) with assay sensitivity is vital for drug development.
• Existing methods often rely on hypothesis tests, with limited focus on confidence intervals for simultaneous assessment.

Purpose of the Study:

• To develop a hybrid approach for constructing simultaneous confidence intervals to assess NI and assay sensitivity in three-arm trials.
• To compare the performance of the proposed hybrid method against normal-approximation and bootstrap-resampling methods.

Main Methods:

• Development of a novel hybrid approach for simultaneous confidence intervals.
• Implementation of normal-approximation-based and bootstrap-resampling-based simultaneous confidence intervals for comparison.
• Conducting simulation studies to evaluate performance metrics like empirical coverage probability and mesial-non-coverage probability.

Main Results:

• The hybrid approach utilizing the Wilson score statistic demonstrated superior performance.
• The hybrid method showed better empirical coverage probability and mesial-non-coverage probability in simulations.
• The proposed methods were illustrated using a practical example.

Conclusions:

• The hybrid confidence interval approach provides a robust method for assessing non-inferiority and assay sensitivity in three-arm trials.
• The Wilson score statistic within the hybrid approach is recommended for its enhanced statistical performance.
• This study contributes a valuable tool for clinical trial biostatisticians.