A dynamic trajectory class model for intensive longitudinal categorical outcome.

Area of Science:

• Gerontology
• Biostatistics
• Longitudinal Data Analysis

Background:

• Longitudinal studies of older adults often involve frequently measured responses like activities of daily living (ADL).
• Analyzing distinct temporal patterns and transitions between these patterns in longitudinal data presents analytical challenges.

Purpose of the Study:

• To present a novel dynamic latent class model for longitudinal categorical responses.
• To enable periodic updating of trajectory class membership based on evolving predictors.
• To jointly model longitudinal responses and informative events like death without assuming conditional independence.

Main Methods:

• Developed a dynamic latent class model for frequently measured longitudinal categorical responses.
• Modeled longitudinal responses within trajectory classes using class-specific generalized linear mixed models.
• Jointly modeled an informative event (e.g., death) with the longitudinal response using shared random effects.

Main Results:

• The proposed method effectively identifies distinct temporal patterns (trajectory classes) in activities of daily living over extended periods.
• Periodic updating of trajectory classes allows differentiation between genuine pattern changes and local response fluctuations.
• Jointly modeling events like death alongside ADL trajectories impacts parameter estimates.

Conclusions:

• The novel dynamic latent class model provides a robust framework for analyzing longitudinal categorical data with frequent measurements.
• The method's ability to periodically update class membership and avoid conditional independence assumptions enhances its applicability in aging research.
• This approach offers valuable insights into the dynamic nature of functional trajectories in older populations.