Epidemic predictions in an imperfect world: modelling disease spread with partial data

Area of Science:

• Veterinary Epidemiology
• Network Science
• Infectious Disease Modeling

Background:

• Big-data epidemic models inform infectious disease control policies.
• Livestock movement data are vital for realistic transmission modeling.
• Detailed livestock movement data are unavailable in many global regions.

Purpose of the Study:

• To determine the necessary quantity of network data for accurate livestock epidemic predictions.
• To assess the impact of data limitations on epidemiological forecasting.
• To identify efficient data sampling strategies for disease control.

Main Methods:

• Utilized a comprehensive UK cattle trade network database.
• Implemented various network data sampling strategies.
• Analyzed the relationship between data quantity and prediction accuracy.

Main Results:

• Targeting high-movement nodes (key trade hubs) enables accurate predictions of epidemic size and spread.
• Reduced datasets, when strategically sampled, can yield reliable epidemiological insights.
• The study quantifies the data requirements for effective disease surveillance.

Conclusions:

• Strategic data collection, focusing on high-movement nodes, can overcome data limitations in livestock disease modeling.
• This approach enhances epidemiological prediction capabilities for countries with restricted data access.
• Findings support resource-efficient disease control strategies in global livestock management.