Spatio-temporal Object-Oriented Bayesian Network modelling of the COVID-19 Italian outbreak data

Area of Science:

• Epidemiology
• Public Health
• Computational Modeling

Background:

• The COVID-19 pandemic presented complex spatial-temporal dynamics.
• Understanding the interplay between disease incidence, healthcare capacity, and mortality is crucial for effective public health response.

Purpose of the Study:

• To model the spatial epidemic dynamics of COVID-19 in Italy.
• To explore static and dynamic relationships between weekly incidence rate, intensive care unit (ICU) occupancy, and death rates.
• To provide a tool for validating policy decisions.

Main Methods:

• Object-Oriented Bayesian Network (OOBN) was employed for spatial modeling.
• An autoregressive approach incorporated spatial and temporal components using lagged variables.
• Data analyzed included weekly incidence, ICU occupancy, and death rates in Italy.

Main Results:

• The OOBN effectively captured spatial and temporal dependencies in COVID-19 spread.
• Identified significant relationships between incidence, ICU occupancy, and mortality rates.
• Demonstrated the model's utility in understanding epidemic progression.

Conclusions:

• The developed Bayesian network model offers valuable insights into COVID-19's spatial-temporal epidemiology.
• The model can serve as a decision-support tool for policymakers managing public health crises.
• Further application of such models can enhance epidemic preparedness and response strategies.