Modeling stochastic processes in disease spread across a heterogeneous social system

Area of Science:

• Epidemiology
• Complex Systems Science
• Computational Biology

Background:

• Infectious disease spread is complex, influenced by external factors and internal system dynamics.
• Understanding diffusion mechanisms is vital for effective disease control but challenging due to hidden transmission routes.
• Existing models often struggle to capture the dynamic nature of disease spread and human mobility.

Purpose of the Study:

• To develop a new diffusion framework for stochastic processes modeling disease spread in metapopulations.
• To incorporate human mobility as topological pathways within heterogeneous social systems.
• To quantify diffusion dynamics (exogeneity, endogeneity) and estimate cross-regional infection flow using Bayesian inference and Granger causality.

Main Methods:

• Proposed a novel diffusion framework for stochastic processes.
• Utilized Bayesian inference with the stochastic Expectation-Maximization algorithm.
• Incorporated human mobility data to represent topological pathways in social systems.
• Applied Granger causality to estimate cross-regional infection flow.

Main Results:

• Successfully quantified underlying diffusion dynamics, distinguishing between exogeneity and endogeneity.
• Accurately estimated cross-regional infection flow.
• Demonstrated model robustness through simulations with noisy data, including missing or delayed case reporting.
• Validated the model using 15 years of dengue outbreak data from Australia.

Conclusions:

• The proposed diffusion framework effectively models infectious disease spread by integrating human mobility.
• The model provides valuable insights into diffusion dynamics and infection flow, crucial for public health interventions.
• The framework's robustness and validation with real-world data highlight its practical applicability for disease control strategies.