Semiparametric regression analysis on longitudinal pattern of recurrent gap times

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Survival Analysis

Background:

• Longitudinal studies often involve recurrent events, where individuals experience multiple occurrences of the same event type.
• Analyzing the time gaps between these recurrent events is crucial for understanding disease progression and treatment effects.
• Existing methods may not fully capture individual heterogeneity or handle censored data effectively.

Purpose of the Study:

• To explore the probability structure of recurrent event gap times under censoring.
• To introduce novel statistical models for analyzing longitudinal recurrent event data with individual heterogeneity.
• To provide robust inference procedures for the longitudinal pattern parameter.

Main Methods:

• Exploration of the probability structure of recurrent gap times in the presence of censoring.
• Development of stratified proportional reverse-time hazards models with unspecified baseline functions.
• Proposal and study of inference procedures using appropriate riskset construction.

Main Results:

• The study elucidates the probability structure of recurrent gap times, accounting for censoring.
• The proposed models effectively accommodate individual heterogeneity in recurrent event analysis.
• Inference procedures demonstrate reliability through simulations and a real-world cohort study.

Conclusions:

• The developed statistical framework provides a powerful tool for analyzing recurrent event gap times in longitudinal studies.
• The methodology is suitable for research involving individual differences and censored data.
• Application to the Denmark schizophrenia cohort study highlights the practical utility of the approach.