Network analysis of Standardbred horse movements between racetracks in Canada and the United States in 2019: Implications for disease spread and control

Abstract

Horse movements are one of the most important factors for the spread of equine diseases, and past epidemics indicate that contact networks play an important role. Network analysis was used to describe the spatial and temporal characteristics of horse movements between Standardbred racetracks in Canada and the United States during 2019, and to characterize the network to provide a better understanding of the potential racetrack-to-racetrack spread of infectious disease within the network. Networks were constructed and analyzed as an overall network (the entire study period) and monthly networks. There were 254 active Standardbred racetracks in 2019, organized in 24 geographically clustered communities. Movements and subsequent network measures of cohesiveness and centrality exhibited strong seasonal variation. Networks were more highly connected during the summer and early autumn, coinciding with peak racing activities. Monthly networks showed evidence of small-world properties, whereby disease introduction into a racetrack within a local cluster could result in the rapid spread to other racetracks within that cluster, and to other topologically distant clusters through few additional movements. Using centrality measures, a small subset of racetracks were identified as highly influential in the network and could be considered high-risk for disease introduction and disease spread to other racetracks. Enhancement of disease prevention strategies might be most appropriate if targeted to the months associated with peak racing season, and particularly to influential racetracks. The networks produced in this study were not a true representation of the entire contact network as the information contained within the race records only allowed for the consideration of between-racetrack movements. Other non-recorded movements represent further contacts in the network that can have a substantial effect on the spread of disease within a network, and the exclusion of this information can result in incorrect network measure estimates. While likely not an easy task, given the initial findings of this study and experiences from past horse industry infectious disease outbreaks, it could be beneficial for the Standardbred industry to put a movement recording strategy in place. One benefit would be enhanced ability to respond rapidly and efficiently in the event of an outbreak, thereby limiting potential animal health and economic impacts. Additional movement data could also enable further characterization of the network to inform optimal disease prevention and control strategies.

Keywords: Centrality; Horse movement; Network analysis; Racetrack; Small-world; Standardbred.