Randomization-Based Covariance Analysis for Confidence Intervals of Treatment Comparisons Based on Restricted Mean...

Area of Science:

• Biostatistics
• Clinical Trials Methodology
• Survival Analysis

Background:

• Existing methods for covariate-adjusted restricted mean survival time (RMST) analysis rely on model-based assumptions.
• These assumptions may not align with the complexities inherent in time-to-event data.

Purpose of the Study:

• To introduce a novel randomization-based method for covariate-adjusted RMST comparisons in randomized controlled trials.
• To provide a robust alternative to model-dependent analyses for survival data.

Main Methods:

• Developed a randomization-based analysis of covariance (RB-ANCOVA) for categorized time-to-event data.
• Constrained covariate mean differences between treatment groups to zero to estimate RMST differences.
• Derived confidence intervals offering enhanced precision compared to unadjusted RMST differences.

Main Results:

• The proposed RB-ANCOVA method provides more precise confidence intervals for covariate-adjusted RMST.
• Demonstrated applicability for comparing two or multiple treatment groups across single or multiple time intervals.
• Validated the methodology using data from a clinical trial for amyotrophic lateral sclerosis (ALS).

Conclusions:

• The randomization-based approach offers a valid and more precise method for covariate-adjusted RMST analysis.
• This methodology is suitable for complex survival data encountered in clinical trials.
• The RB-ANCOVA provides a valuable tool for analyzing treatment effects in terms of RMST.