Epidemic spreading in a weighted pig trade network

Area of Science:

• Veterinary Epidemiology
• Network Analysis
• Disease Transmission Modeling

Background:

• Trade networks play a crucial role in the spread of animal diseases.
• Understanding factors influencing epidemic size is vital for effective disease control.
• Edge weights in network analysis can provide deeper insights into complex processes.

Purpose of the Study:

• To evaluate the effect of different network versions (unweighted, trade contacts, livestock volume, livestock per contact) on an epidemiological model's outcome.
• To assess the impact of transmission probability and primary outbreak farm type on final epidemic size.
• To analyze the interactions between network versions, transmission probability, and farm type.

Main Methods:

• Utilized a pig trade network dataset from Northern Germany (2013-2014) with 678 farms and 1,018 directed edges.
• Implemented an epidemiological model incorporating edge weights to influence disease spread probability.
• Employed a generalized linear mixed model to analyze the significance of fixed effects and interactions.

Main Results:

• All tested fixed effects (network version, transmission probability, farm type) and their interactions were statistically significant (p ≤ 0.05).
• Network versions B (sum of delivered livestock) and C (mean delivered livestock per contact) consistently resulted in the highest number of infected farms.
• Farm types at the beginning of the production chain (e.g., breeding farms) showed a significantly higher number of infected farms and greater reach within the network.

Conclusions:

• The inclusion of edge weights significantly affects epidemiological model outcomes.
• Network versions reflecting livestock volume (B and C) are critical for accurately modeling disease spread in trade networks.
• Disease control strategies should consider the type of farm initiating an outbreak and its position within the production chain.