An efficient Gehan-type estimation for the accelerated failure time model with clustered and censored data

Area of Science:

• Biostatistics
• Medical Data Analysis
• Survival Analysis

Background:

• Traditional Gehan-type estimators are sensitive to covariate outliers and ignore within-cluster correlations in censored clustered/longitudinal data.
• Existing methods may not adequately handle outliers in covariates or account for complex data structures common in medical studies.

Purpose of the Study:

• To develop a robust statistical method for parameter estimation in accelerated failure time models for clustered data.
• To address limitations of traditional estimators by accounting for within-cluster correlations, varying cluster sizes, and covariate outliers.

Main Methods:

• Proposed weighted Gehan-type estimating functions for parameter estimation in the accelerated failure time model.
• Developed methods for clustered data with censored observations and outliers in covariates.
• Provided asymptotic properties of the proposed estimators.

Main Results:

• Simulation studies demonstrated the proposed method's robustness to covariate outliers.
• The new method yields more efficient estimators when strong within-cluster correlations are present.
• The approach successfully handled varying cluster sizes and censored observations.

Conclusions:

• The proposed weighted Gehan-type estimating functions offer a reliable and efficient approach for analyzing complex medical data.
• This method enhances the accuracy of parameter estimation in accelerated failure time models for clustered and longitudinal studies.
• Application to medical datasets yielded more convincing results, highlighting its practical utility.