gsem: A Stata command for parametric joint modelling of longitudinal and accelerated failure time models

Area of Science:

• Biostatistics
• Statistical Modeling
• Epidemiology

Background:

• Joint modeling of longitudinal and survival data is increasingly used but faces analytical and software challenges.
• Traditional Cox regression for survival sub-models may fail when proportional hazards assumptions are unmet, necessitating alternative approaches.
• Parametric survival models offer a preferable alternative to Cox regression in such scenarios.

Purpose of the Study:

• To describe and demonstrate parametric survival models within the framework of joint modeling.
• To showcase the application of the generalized structural equation model (gsem) command in Stata for fitting these joint models.
• To illustrate the practical use of gsem for parametric joint modeling using a real-world dataset.

Main Methods:

• Utilized the gsem command in Stata, which integrates linear mixed-effects models for the longitudinal sub-model.
• Explored five parametric survival models within gsem: exponential, Weibull, log-normal, log-logistic, and gamma accelerated failure time models.
• Applied the developed methodology to a dataset of 312 patients with primary biliary cirrhosis.

Main Results:

• Successfully described the properties of the gsem command for parametric joint modeling.
• Demonstrated a practical application of parametric joint models using the primary biliary cirrhosis dataset.
• Validated the effectiveness of gsem for fitting complex joint statistical models.

Conclusions:

• Parametric joint models can be effectively implemented using the gsem command in Stata.
• The gsem command is presented as a valuable and potentially unique tool in the literature for fitting parametric joint models.
• The application to primary biliary cirrhosis data underscores the command's utility for health-related research.