Weighted estimating equations for linear regression analysis of clustered failure time data.

Area of Science:

• Statistics
• Biostatistics
• Survival Analysis

Background:

• Linear survival models are crucial for analyzing time-to-event data.
• Clustered data presents unique challenges in parameter estimation due to dependencies.
• Existing methods may lack efficiency or be computationally intensive in clustered settings.

Purpose of the Study:

• To develop and evaluate efficient estimation methods for regression parameters in linear survival models with clustered data.
• To compare the efficiency of proposed methods against existing techniques.
• To assess the performance of practical estimators under different working models.

Main Methods:

• Proposed one-step updates from an initial consistent estimator.
• Utilized rank-based scores incorporating weight matrices for improved efficiency.
• Approximated optimal weights using quadratic programming.
• Conducted simulation studies to evaluate estimator performance.

Main Results:

• The proposed one-step updates provide consistent estimation.
• Rank-based scores with weight matrices enhance efficiency.
• Simpler methods demonstrated comparable efficiency to optimal weights, except under strong dependence.
• Simulation results explored the practical performance of various estimators.

Conclusions:

• Efficient estimation of regression parameters in linear survival models is achievable with clustered data.
• Rank-based scores and optimized weights improve estimation efficiency.
• Practical, simpler methods offer a good balance of efficiency and ease of use for clustered survival data analysis.