Mixture mean residual life model for competing risks data with mismeasured covariates

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Competing risks data present challenges in accurately modeling failure times.
• Covariate measurement error complicates the analysis of time-to-event data.
• Existing methods may not adequately handle both competing risks and measurement error simultaneously.

Purpose of the Study:

• To propose a novel mixture regression model for analyzing competing risks data.
• To develop robust statistical methods for handling covariate measurement error in this context.
• To provide a framework for reliable inference in complex survival data analysis.

Main Methods:

• A mixture regression model combining logistic regression for failure probabilities and mean residual lifetime (MRL) models.
• Derivation of estimating equations (EEs) for separate inference of logistic and MRL components.
• Application of corrected score estimation using EEs to address measurement error without distributional assumptions for covariates.

Main Results:

• The proposed estimators are demonstrated to be consistent and asymptotically normal.
• Simulation studies confirm the method's performance in various scenarios.
• The approach effectively handles competing risks and covariate measurement error.

Conclusions:

• The developed mixture regression model with corrected score estimation offers a robust solution for competing risks data with measurement error.
• The method provides reliable and asymptotically valid inferences.
• The approach is validated through simulations and real-world data applications.