On the self-triggering Cox model for recurrent event data.

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Survival Analysis

Background:

• Recurrent event data are common in longitudinal studies, where multiple events can occur per subject.
• The dependence between successive events is often overlooked in standard recurrent event analyses.
• Ignoring this dependence can lead to suboptimal model fit and prediction accuracy.

Purpose of the Study:

• To develop and validate a novel Cox-type regression model that incorporates a time-varying triggering effect.
• To improve the accuracy of recurrent event data modeling by accounting for the influence of prior events.
• To provide a statistical test for assessing the significance of triggering effects.

Main Methods:

• Utilized a Cox-type regression model incorporating a time-varying triggering effect based on the number and timing of previous events.
• Employed partial likelihood for parameter estimation and statistical inference.
• Developed a statistical test to evaluate the presence of triggering effects.

Main Results:

• The proposed model demonstrated superior fit and prediction capabilities compared to standard recurrent event models in simulation studies.
• A real-world data analysis on chronic pseudomonas infections in cystic fibrosis patients validated the model's effectiveness.
• The statistical test successfully identified significant triggering effects in the analyzed data.

Conclusions:

• The novel Cox-type regression model with time-varying triggering effects offers a significant advancement in analyzing recurrent event data.
• This approach enhances model performance and predictive accuracy, particularly in clinical and epidemiological research.
• The method provides a robust framework for understanding and modeling event dependencies in longitudinal studies.