Attenuation caused by infrequently updated covariates in survival analysis

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Time-dependent covariates in survival analysis present challenges for regression models.
• Infrequent updating of quantitative covariates can attenuate regression coefficients.
• The Cox proportional hazards model and additive hazard models are commonly used.

Purpose of the Study:

• To evaluate the attenuation of regression coefficients due to infrequent covariate updating.
• To propose and compare simple adjustment methods for infrequent covariate updating.
• To assess the impact of different stochastic processes and trends on covariate attenuation.

Main Methods:

• Simulation studies using data mimicking the CSL1 liver cirrhosis trial.
• Evaluation of Cox proportional hazards and additive hazard models.
• Comparison of proposed simple adjustment methods with existing techniques.

Main Results:

• Attenuation of regression coefficients can be substantial, particularly for Ornstein-Uhlenbeck processes.
• Covariate trends and non-synchronous updating also contribute to attenuation.
• Simpler adjustment methods show advantages in large datasets with infrequent updating.

Conclusions:

• The degree of attenuation depends on the covariate's underlying stochastic process.
• Simple adjustment methods are effective and practical for large survival datasets with infrequently updated covariates.
• Existing complex methods may be less advantageous than simpler techniques under sparse updating conditions.