A joint model for longitudinal continuous and time-to-event outcomes with direct marginal interpretation

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Survival Analysis

Background:

• Joint modeling of longitudinal and time-to-event data is crucial for understanding complex biological processes.
• Existing methods often lack flexibility in handling associations between different data types or accommodating overdispersion.

Purpose of the Study:

• To propose a unified and flexible marginalized joint model framework for analyzing continuous longitudinal and repeated time-to-event outcomes.
• To develop a similar framework for bivariate repeated time-to-event outcomes.
• To ensure model parameters have direct marginal interpretations.

Main Methods:

• Development of a marginalized joint model building upon the generalized linear mixed model framework.
• Incorporation of flexible association structures for within- and between-outcome measurements.
• Accommodation of overdispersion and censoring for time-to-event outcomes.

Main Results:

• The proposed model framework successfully integrates various features, including flexible associations and overdispersion.
• Simulation studies confirm the model's properties and performance.
• The models are effectively applied to real-world data from chronic heart failure and comet assay studies.

Conclusions:

• The developed marginalized joint models provide a unified, flexible, and interpretable approach for joint analysis of longitudinal and time-to-event data.
• The models are computationally feasible and can be fitted using standard statistical software.
• This framework offers significant advantages for analyzing complex biomedical data.