Heterogeneity matters: Contact structure and individual variation shape epidemic dynamics

Area of Science:

• Epidemiology
• Mathematical Biology
• Computational Science

Background:

• Mathematical modeling is vital for evaluating non-pharmaceutical interventions (NPIs) during pandemics like COVID-19.
• Current models often rely on average parameters, potentially overlooking crucial population heterogeneities.

Purpose of the Study:

• To investigate the impact of population heterogeneity in connectivity and viral load on COVID-19 epidemic dynamics using stochastic simulations.
• To compare the outcomes of heterogeneous models with traditional point-estimate-based models (e.g., ODEs).

Main Methods:

• Translated a COVID-19 ordinary differential equation (ODE) model into a stochastic multi-agent system.
• Utilized contact networks to represent complex population interaction structures.
• Incorporated probabilistic infection rates to account for individual viral load variations.

Main Results:

• Population heterogeneity significantly influences epidemic dispersion and evolution, effects not captured by average-based models.
• Different types of heterogeneity (e.g., network hubs vs. individual infectivity) lead to distinct epidemic trajectories.
• Network hubs increase initial dispersion and effective reproduction number but lower the herd immunity threshold (HIT).

Conclusions:

• Stochastic multi-agent models are essential for accurately capturing the effects of population heterogeneity in infectious disease modeling.
• Policy-making based solely on average parameters may be insufficient; considering heterogeneity is critical for effective pandemic response and understanding herd immunity.