Bayesian regression model for recurrent event data with event-varying covariate effects and event effect

Area of Science:

• Biostatistics
• Cardiovascular Epidemiology
• Statistical Modeling

Background:

• Cardiovascular disease events like stroke and heart failure frequently recur in hypertensive patients.
• Existing models often struggle to account for complex correlations and changing risk factors associated with recurrent events.

Purpose of the Study:

• To develop and present a novel Bayesian regression model for recurrent cardiovascular events.
• To explicitly model subject-specific heterogeneity, event dependence, event-varying covariate effects, and the impact of prior events on future risks.

Main Methods:

• Proposed a Bayesian regression framework to analyze recurrent event data.
• Incorporated methods to address correlation from heterogeneity and event dependence.
• Modeled event-varying covariate effects and the direct effect of prior events.

Main Results:

• The proposed model effectively quantifies how recurrent event incidences alter covariate effects and future event risks.
• Demonstrated the model's utility in analyzing the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT-LLT) data.
• Outperformed commonly used recurrent event models in specific analytical scenarios.

Conclusions:

• The novel Bayesian model provides a robust approach for analyzing recurrent cardiovascular events in hypertension.
• This method enhances understanding of disease progression and treatment effects by accounting for complex event correlations and time-varying factors.
• The model offers improved quantification of risk for recurrent cardiovascular events, aiding clinical decision-making.