A discrete-continuous mixture quantile function estimator with a practical application to phase II cancer clinical...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Data Analysis

Background:

• Continuous variables or mixed data are often dichotomized for simpler interpretation in biological and clinical research.
• Exact binomial tests are commonly used for analyzing dichotomized data, but may lack optimal statistical efficiency.

Purpose of the Study:

• To introduce and illustrate a novel quantile-based approach for analyzing dichotomized data.
• To compare the relative efficiency and interpretability of the new method against traditional binomial tests.
• To demonstrate the application of this approach in analyzing discrete-continuous mixtures, such as those in Phase II cancer trials.

Main Methods:

• Development of a quantile-based statistical testing framework.
• Comparative analysis of statistical efficiency between the proposed method and exact binomial tests.
• Application and validation using simulated discrete-continuous mixture data and real-world Phase II cancer clinical trial data.

Main Results:

• The quantile-based approach demonstrates improved relative efficiency compared to the exact binomial test.
• The new method maintains a similar level of interpretability as the binomial test.
• The approach is effective for analyzing complex data structures common in cancer clinical trials.

Conclusions:

• The proposed quantile-based method offers a statistically superior and practically applicable alternative to binomial tests for dichotomized data.
• This method enhances the analysis of mixed-type data, particularly relevant for clinical trial settings.
• Further adoption of this method can lead to more robust and efficient conclusions in biological and clinical research.