ANALYSIS OF DEPENDENTLY CENSORED DATA BASED ON QUANTILE REGRESSION

Area of Science:

• Biostatistics
• Survival Analysis
• Biomedical Research

Background:

• Dependent censoring presents significant methodological challenges in biomedical survival analysis.
• Traditional regression models often assume constant covariate effects, limiting their applicability.

Purpose of the Study:

• To develop a new approach for analyzing dependently censored data using quantile regression models.
• To formulate covariate effects on the quantiles of the marginal distribution of event times.
• To accommodate more dynamic relationships between covariates and survival time.

Main Methods:

• Utilizing quantile regression models to analyze event time data with dependent censoring.
• Proposing novel estimation and inference procedures with an efficient algorithm.
• Establishing theoretical properties including uniform consistency and weak convergence of estimators.

Main Results:

• Extensive simulation studies confirmed the good finite-sample performance of the proposed inferential procedures.
• The method demonstrated practical utility in analyzing a multicenter clinical trial comparing warfarin and aspirin.

Conclusions:

• The proposed quantile regression approach effectively handles dependent censoring in survival analysis.
• This method provides a more flexible alternative to traditional models by capturing dynamic covariate effects.
• The approach is validated through simulations and a real-world clinical trial application.