Frailty modelling for survival data from multi-centre clinical trials.

Area of Science:

• Biostatistics
• Clinical Trial Methodology
• Survival Analysis

Background:

• Multi-centre randomized clinical trials (RCTs) can exhibit outcome heterogeneity across centers despite standardized protocols.
• This heterogeneity may complicate the interpretation and reporting of treatment effects.
• Investigating treatment-by-centre interactions is crucial for understanding trial variations.

Purpose of the Study:

• To propose a general frailty modeling approach for analyzing time-to-event data in multi-centre clinical trials.
• To specifically investigate potential treatment-by-centre interactions.
• To provide a robust statistical framework for handling heterogeneity in clinical trial outcomes.

Main Methods:

• Utilized a correlated random effects model to simultaneously model baseline risk and treatment effects across centers.
• Employed the hierarchical-likelihood (h-likelihood) approach for statistical inference.
• Facilitated computation of prediction intervals for random effects with enhanced precision.

Main Results:

• Demonstrated the application of the proposed frailty modeling approach using disease-free survival data from a bladder cancer clinical trial.
• Validated the methodology through a comprehensive simulation study.
• Showcased model selection capabilities using h-likelihood criteria.

Conclusions:

• The proposed frailty modeling approach effectively addresses outcome heterogeneity in multi-centre clinical trials.
• The h-likelihood framework provides a precise method for analyzing treatment-by-centre interactions.
• This methodology enhances the reliability of interpreting treatment effects in complex trial settings.