Combining different sources of movement data to strengthen traceability and disease surveillance

Abstract

Recent concerns with food safety in the United States have highlighted the importance of traceability systems in animal production chains. Yet, adoption of these systems presents various challenges. Interstate Certificates of Veterinary Inspection (ICVIs) are required for most interstate animal movements and are considered the most representative source of livestock movement data. However, exceptions exist, where Owner Shipper Statements (OSSs) are often used as an alternative but previously unexplored. Calf movement networks are understudied, yet important in understanding animal and human disease transmission dynamics. The objective of this study was to use movement records to describe calf networks within a US region and explore how the inclusion of OSSs impact the structure of calf networks built using ICVIs. Calf movement records to and from Ohio were obtained through ICVIs and OSSs from June 2021 to June 2022. To explore and compare movement patterns, network analysis was performed individually for an ICVI-based network and a network combining both document types. Zip codes were considered nodes and calf movements (cattle up to 4 months) were considered links. Whole-network and node-level parameters were calculated, and Mann-Whitney U tests were performed to evaluate statistical differences by network type. Community detection was performed to investigate the underlying structure of calf networks in Ohio. The frequency of animal movements recorded through OSS (n = 766, 49.8 %) and ICVIs (n = 772, 50.2 %) was similar. Most animal movements included mixed sex (60.0 %), dairy breeds (81.6 %) and animals up to one week old (74.1 %). There were major differences in the network structure with OSSs compared to ICVIs exclusively. Movements recorded through OSSs showed larger median number of animals per movement (60; IQR 23–105) compared to ICVIs (49; IQR 16–80); reaching up to 696 calves per batch of transported calves. Failing to include OSSs would have resulted in an incomplete network, excluding 40.3 % of the zip codes (n = 206) represented in this database. The ICVI-based network involved fewer zip codes across states, whose connections were sparser than in the combined network. The two analyzed networks revealed contrasting centrality results, especially for out-going geographical regions (P < 0.01), suggesting a discrepancy in their potential to influence disease transmission dynamics. Moreover, including OSSs resulted in a network with lower closeness centrality scores (P < 0.01). Results suggested heterogeneous patterns of calf movements, depending on the source of records, and emphasized the importance of incorporating multiple sources of movement data for the development of targeted disease surveillance strategies, particularly using community detection analysis.