MEGH

MEGH implements a Mixed-Effects General Hazard model for analysis of clustered survival data, accounting for fixed and random effects across clusters such as medical centers or geographical/administrative regions.

Key Features:

• Parametric Mixed-Effects Model: Extends survival models to include parametric fixed and random effects for handling between-cluster variability.
• General Hazard Structure: Generalizes mixed-effects proportional hazards and mixed-effects accelerated failure time models to provide a flexible hazard framework.
• Likelihood-Based Parameter Estimation: Uses a likelihood-based algorithm for parameter estimation within general subclasses of the model.
• Diagnostic Tools: Implements diagnostics to assess random effects and their distributional assumptions.
• Performance Evaluation: Performance has been evaluated via theoretical studies, simulation experiments, and real-world applications, including analysis of leukaemia data.
• Implementation: Implemented as an R package called "MEGH".

Scientific Applications:

• Epidemiological Studies: Analysis of clustered survival data in epidemiology, including multi-center or regionally grouped datasets.
• Clinical Trials: Analysis of survival outcomes in multi-center clinical trials accounting for between-center variability.
• Biomedical Research: Application to biomedical datasets with regional or administrative clustering, exemplified by leukaemia data analysis.

Methodology:

Parametric mixed-effects modelling within a general hazard framework that includes mixed-effects proportional hazards and mixed-effects accelerated failure time subclasses; likelihood-based parameter estimation for general model subclasses; diagnostic procedures for random effects and their distributional assumptions; evaluation via theoretical studies and simulation experiments and application to real-world data (leukaemia).