A corrected pseudo-score approach for additive hazards model with longitudinal covariates measured with error

Area of Science:

• Biostatistics
• Survival Analysis
• Longitudinal Data Analysis

Background:

• Characterizing time-to-event data with time-dependent covariates is crucial in medical studies.
• Existing proportional hazards models often rely on normality assumptions and face numerical challenges with error-prone, intermittently measured covariates.
• Limited research exists on additive hazards models for time-dependent covariates measured with error.

Purpose of the Study:

• To propose a novel statistical approach for analyzing time-to-event data with time-dependent covariates measured with error.
• To develop a method within the additive hazards framework that bypasses normality assumptions.
• To provide a consistent and asymptotically normal estimator for regression parameters.

Main Methods:

• A corrected pseudo-score approach is proposed for regression parameters.
• The method does not assume normality for random effects or errors, only requiring variance structure assumptions.
• The proposed estimator yields an explicit form.

Main Results:

• The developed estimator is proven to be consistent and asymptotically normal.
• The approach demonstrates practical applicability through simulations.
• The method is validated using data from an HIV clinical trial.

Conclusions:

• The corrected pseudo-score approach offers a viable and robust alternative for analyzing time-to-event data with complex covariate structures.
• This method overcomes limitations of traditional proportional hazards models, particularly when dealing with measurement error and non-normal data.
• The findings have significant implications for biostatistical analysis in clinical trials and other medical research areas.