Identification of the glucose minimal model by stochastic nonlinear-mixed effects methods

Area of Science:

• Pharmacometrics
• Pharmacokinetics/Pharmacodynamics (PKPD)
• Epidemiological Modeling

Background:

• Nonlinear mixed effects models (NLMEM) are widely used in PKPD and epidemiological studies for analyzing individual and population data.
• NLMEM effectively handle sparse individual data by leveraging population-level information, outperforming methods like weighted least squares (WLS).
• Maximizing the likelihood function in NLMEM can be challenging due to model nonlinearity, necessitating various estimation techniques.

Purpose of the Study:

• To evaluate the performance of recently developed estimation methods for NLMEM.
• To extend the simulation study by Denti et al. by applying these new methods.
• To assess the application of these methods on the intravenous glucose tolerance test (IVGTT) glucose minimal model.

Main Methods:

• Comparison of established methods: First Order (FO), First Order Conditional (FOCE), and Expectation Maximization (EM) algorithm.
• Evaluation of newer EM-based methods: Iterative Two Stage (ITS), Monte Carlo Importance Sampling EM (IMP), IMP assisted by Mode a Posteriori estimation (IMPMAP), and Stochastic Approximation EM (SAEM).
• Inclusion of Markov Chain Monte Carlo Bayesian Analysis (BAYES) as a comparative method.

Main Results:

• The study provides an evaluation of the newest population methods for NLMEM.
• Performance assessment is conducted using the IVGTT glucose minimal model.
• Results aim to complement existing simulation studies in the field.

Conclusions:

• The study contributes to understanding the performance of advanced NLMEM estimation techniques.
• Findings are relevant for researchers in PKPD and epidemiology utilizing complex modeling approaches.
• The evaluation on the IVGTT model provides practical insights for model selection and application.