Sample size considerations for livestock movement network data.

Area of Science:

• Veterinary Epidemiology
• Network Science
• Computational Biology

Background:

• Animal movements between farms are key drivers of infectious disease spread in livestock populations.
• Social network analysis (SNA) is increasingly used to study these movements and identify high-risk premises.
• Limited understanding exists regarding the impact of sampling and incomplete network data on SNA studies.

Purpose of the Study:

• To develop and present a simulation algorithm for estimating required sampling proportions in animal movement network analysis.
• To enable a priori determination of sample sizes for network analyses using incomplete or sampled data.
• To ensure population estimates derived from network analyses possess known precision.

Main Methods:

• A simulation algorithm was developed to estimate sampling proportions based on network size, density, and degree distribution.
• The algorithm's repeatability was assessed for networks with at least 1000 nodes (farms).
• Simulated networks were constructed to mimic real-world livestock movement networks.

Main Results:

• The simulation algorithm provides estimates for required sampling proportions for network analysis.
• Sample size requirements for network degree measures were found to vary depending on the sampling method used.
• The algorithm demonstrated consistent output repeatability for networks of 1000 or more nodes.

Conclusions:

• The presented algorithm offers a practical approach for determining appropriate sample sizes in livestock movement network studies.
• It facilitates the design of studies with known precision for analyzing disease dissemination.
• The tool aids in tailoring surveillance and control strategies by understanding network structures and disease spread dynamics.