Statistical inference for clinical trials with binary responses when there is a shift in patient population

Area of Science:

• Clinical trial methodology
• Biostatistics
• Regulatory science

Background:

• Clinical trial protocols are frequently amended, potentially altering study populations and objectives.
• Significant protocol modifications can compromise a trial's ability to answer original research questions.
• Existing statistical methods may not adequately address inference for trials with amendments.

Purpose of the Study:

• To propose a robust statistical inference model for clinical trials undergoing protocol amendments.
• To develop methods for sample size adjustment in the context of protocol modifications.
• To ensure the integrity and validity of statistical conclusions from amended clinical trials.

Main Methods:

• Development of a logistic regression model tailored for binary endpoints in trials with protocol amendments.
• Derivation of statistical inference procedures under the proposed model.
• Formulation of sample size adjustment strategies compatible with protocol amendments.

Main Results:

• The proposed logistic regression model provides a framework for valid statistical inference despite protocol amendments.
• The derived sample size adjustment method accounts for modifications, maintaining statistical power.
• Simulations demonstrate the proposed method's performance in amended trial scenarios.

Conclusions:

• The proposed logistic regression approach offers a statistically sound method for analyzing clinical trials with protocol amendments.
• This methodology supports reliable inference and appropriate sample size determination, preserving research validity.
• Implementing these methods can enhance the quality and interpretability of findings from modified clinical trials.