Moment approximation of infection dynamics in a population of moving hosts

Area of Science:

• Epidemiology
• Computational Biology
• Mathematical Modeling

Background:

• Epidemiological models often use static networks, but real-world host contact structures are dynamic.
• Simulating large populations with individual-based models (IBMs) is computationally intensive.
• Understanding host movement and contact is crucial for disease spread dynamics.

Purpose of the Study:

• To develop and validate a moment approximation (MA) method for approximating stochastic individual-based models (IBMs) in epidemiology.
• To assess the accuracy of the MA method compared to IBMs and mean-field (MF) models.
• To apply the MA method to model avian influenza (H5N1) spread in poultry flocks.

Main Methods:

• Developed a moment approximation (MA) method to create a deterministic model from a stochastic individual-based model (IBM).
• Compared simulation results from the MA model, an IBM, and a mean-field (MF) model.
• Applied models to simulate the spread of avian influenza (H5N1) in a poultry flock.

Main Results:

• The MA model demonstrated close agreement with the IBM simulations.
• The study highlights the necessity of incorporating host displacement and contact processes into epidemiological models.
• The MA method provides a computationally efficient alternative to IBMs for large populations.

Conclusions:

• The moment approximation (MA) method is a viable and accurate approach for modeling dynamic contact processes in epidemiological studies.
• Accurate representation of host behavior, including movement and contact, is essential for effective disease spread modeling.
• This work offers a reusable framework for approximating complex stochastic models in epidemiology and other fields.