A Bayesian shared parameter model for joint modeling of longitudinal continuous and binary outcomes.

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Biomarker Research

Background:

• Joint modeling of biomarkers enhances clinical decision-making by improving parameter estimation efficiency.
• Longitudinal studies often involve biomarkers with non-equivalent observation times and disparate missing data patterns.
• Accurate analysis of such complex data is crucial for reliable clinical insights.

Purpose of the Study:

• To propose a novel joint model for associated continuous and binary longitudinal responses.
• To address challenges posed by different missing data patterns and unequal observation times.
• To enhance the accuracy and efficiency of biomarker analysis in clinical studies.

Main Methods:

• A conditional model for joint modeling of continuous and binary responses was developed.
• Two shared random effects models were employed to handle intermittent missingness.
• Parameter estimation and model implementation were performed using a Bayesian approach with Markov Chain Monte Carlo (MCMC).

Main Results:

• Simulation studies validated the performance and robustness of the proposed joint model.
• The model demonstrated effectiveness in handling longitudinal data with mixed response types and complex missingness.
• The proposed method provided reliable parameter estimates for associated biomarkers.

Conclusions:

• The developed joint model offers a robust framework for analyzing mixed biomarkers in longitudinal studies with complex data structures.
• This approach improves the efficiency of parameter estimates, leading to better clinical decision-making.
• The model's application to bariatric surgery data highlights its practical utility in real-world clinical research.