Proportion of treatment effect mediated by surrogate endpoints

Area of Science:

• Biostatistics
• Clinical Trial Design
• Epidemiology

Background:

• Surrogate endpoints are crucial for efficient randomized clinical trials.
• Existing surrogacy measures have limitations including imprecise estimation and ignoring treatment interactions.

Purpose of the Study:

• To propose a novel surrogacy measure, the proportion of treatment effect mediated (PMS), to address limitations of current methods.
• To evaluate the performance and advantages of PMS.

Main Methods:

• Developed PMS based on decomposing treatment effects into direct, indirect, and interaction components.
• Validated PMS using Monte Carlo simulations.
• Applied PMS to empirical data from the ORIENT trial.

Main Results:

• PMS offers a more robust and interpretable measure of surrogacy compared to existing methods.
• Simulations demonstrated the advantages of PMS in various scenarios.
• Empirical application confirmed the utility of PMS in real-world trial data.

Conclusions:

• PMS provides a valuable tool for assessing surrogate endpoints in clinical trials.
• This new measure enhances the reliability and efficiency of trial design and analysis.
• PMS facilitates a better understanding of treatment effects and mediation pathways.