Robust evaluation of longitudinal surrogate markers with censored data

Area of Science:

• Biostatistics
• Clinical Trials
• Epidemiology

Background:

• Evaluating surrogate markers is crucial in clinical research.
• Existing methods often fail with longitudinal markers or time-to-event outcomes.
• Longitudinal data, like repeated glucose measurements, are common but complex to analyze.

Purpose of the Study:

• To develop robust statistical methods for evaluating longitudinal surrogate markers.
• To address challenges posed by censored time-to-event primary outcomes.
• To quantify the proportion of treatment effect explained by a longitudinal surrogate marker.

Main Methods:

• Proposed a novel method to define and estimate the treatment effect proportion explained by a longitudinal surrogate.
• Accommodated censoring for both primary outcomes and surrogate marker measurements.
• Utilized simulation studies to assess finite-sample performance.

Main Results:

• The proposed methods demonstrated good performance in simulations.
• Successfully applied the methods to analyze fasting plasma glucose as a surrogate marker for diabetes.
• Quantified the treatment effect explained by the longitudinal surrogate marker.

Conclusions:

• The developed methods offer a robust approach for analyzing longitudinal surrogate markers.
• These methods are applicable in settings with censored time-to-event outcomes.
• Facilitates better understanding of treatment effects in complex clinical trial data.