Semiparametric methods for survival data with measurement error under additive hazards cure rate models

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Measurement error significantly impacts data analysis, biasing estimates and reducing statistical power.
• Survival analysis with error-contaminated data is challenging, especially for populations with a cured fraction.
• Existing methods often focus on relative risks (proportional hazards model), neglecting absolute risk differences.

Purpose of the Study:

• To develop robust estimation methods for survival data with measurement error in the presence of a cured fraction.
• To extend the additive hazards model to handle error-contaminated covariates and time-dependent covariates.
• To provide a flexible framework for estimating absolute risk differences in complex survival data.

Main Methods:

• Utilizing an additive hazards model for non-cured individuals' lifetimes.
• Modeling the measurement error process with an additive model.
• Developing estimation techniques for scenarios both with and without measurement error.
• Incorporating time-dependent covariates for enhanced model flexibility.

Main Results:

• The proposed methods effectively handle survival data with measurement error and cured fractions.
• The additive hazards model allows for the estimation of absolute risk differences, offering a complementary perspective to relative risks.
• Theoretical and numerical evaluations demonstrate the validity and utility of the developed methods.
• A real-life data application confirms the practical applicability of the methodology.

Conclusions:

• The developed methods provide a valuable tool for analyzing complex survival data affected by measurement error and cured fractions.
• The approach enhances the understanding of covariate effects by enabling the estimation of absolute risk differences.
• This work contributes to the field of survival analysis by offering flexible and robust statistical techniques.