Enhancing insight into regional differences: hierarchical linear models in multiregional clinical trials

Area of Science:

• Clinical Trials Methodology
• Statistical Modeling
• Pharmaceutical Research

Background:

• Multi-regional clinical trials (MRCTs) are increasingly used for global drug development.
• International Council for Harmonisation E17 guideline highlights the need to address regional variations in MRCTs.
• Regional differences, stemming from intrinsic and extrinsic patient factors, pose challenges for MRCT design and analysis.

Purpose of the Study:

• To introduce and investigate hierarchical linear models (HLMs) for analyzing MRCTs.
• To enhance HLMs by incorporating random effects in both intercept and slope for greater flexibility.
• To develop methods for sample size calculation under HLMs, considering budget constraints.

Main Methods:

• Utilized hierarchical linear models (HLMs) with covariates for known factors and random effects for unknown factors.
• Extended HLMs to include random effects in both intercept and slope.
• Developed methods for sample size determination considering fixed regions and budgetary limitations.

Main Results:

• HLMs with random intercept and slope effects yield accurate type I error rates and power when the number of regions is adequate.
• Estimating regional variabilities is challenging with a small number of regions.
• Budget constraints influence the number of regions, and patient numbers per region depend on treatment effect variability.

Conclusions:

• A robust framework is presented for managing regional endpoint differences in MRCTs.
• Proposed strategies, including figures and budget-conscious sample size calculations, enhance MRCT design efficiency.