The additive nonparametric and semiparametric Aalen model as the rate function for a counting process.

Area of Science:

• Statistics
• Survival Analysis
• Biostatistics

Background:

• The additive risk model is a statistical tool for analyzing event rates in counting processes.
• Current methods often rely on intensity-based models, which can be complex and require extensive data.
• There is a need for models that focus on the rate function using selected covariates.

Purpose of the Study:

• To present a new statistical model for the rate function of a counting process, diverging from intensity-based approaches.
• To evaluate the performance of existing and new estimators for the Aalen model and its semi-parametric variant.
• To introduce a novel test statistic for assessing time-constant effects within these models.

Main Methods:

• Utilizing the additive risk model framework, focusing on the rate function instead of intensity.
• Applying the Aalen estimator for rate functions and proposing an alternative variance estimator.
• Investigating the semi-parametric Aalen model and deriving corrected standard error estimators.
• Developing and implementing a test statistic for time-constant effects.

Main Results:

• The standard Aalen estimator provides nearly unbiased estimates when the rate function follows the Aalen form.
• The conventional martingale-based variance estimator is inaccurate; an alternative is necessary.
• Standard errors computed assuming intensities are incorrect for the semi-parametric Aalen rate model, necessitating a different estimator.
• A new test statistic for time-constant effects was developed and evaluated.

Conclusions:

• The proposed modifications offer more accurate estimation and hypothesis testing for counting processes using the Aalen model.
• The study highlights the importance of using appropriate variance estimators and provides a new tool for analyzing time-constant effects.
• This work contributes to more robust statistical modeling in survival analysis and related fields.