Joint analysis of panel count and interval-censored data using distribution-free frailty analysis

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Recurrent and nonrecurrent event data are common in medical research.
• Interval censoring is frequently encountered in longitudinal studies.
• Existing methods often analyze these event types separately, potentially missing important dependencies.

Purpose of the Study:

• To develop a unified statistical framework for jointly analyzing interval-censored recurrent and nonrecurrent event data.
• To incorporate flexible semiparametric models, including Box-Cox transformations.
• To account for dependence between event processes using an unspecified frailty variable.

Main Methods:

• Joint analysis using pseudolikelihood for recurrent (panel count) data and sufficient likelihood for nonrecurrent data.
• Conditioning on the sufficient statistic for the frailty variable.
• Assuming independence of events over examination times.

Main Results:

• Established large sample theory for the proposed joint analysis.
• Developed a computational procedure for practical implementation.
• Demonstrated applicability through analyses of skin cancer and scleroderma lung disease datasets.

Conclusions:

• The proposed methodology provides a robust approach for joint analysis of complex event data under interval censoring.
• The method effectively models dependencies between different event types.
• It offers valuable insights for analyzing longitudinal health outcomes.