Network Analysis of Livestock Movements to Estimate Potential Silent Spread of Foot-and-mouth Disease

Introduction Recognizing the importance of livestock movements in the spread of contagious diseases, various countries in the world have developed livestock movement databases. Social network analysis techniques have recently been applied to study such databases as this technique has the advantage of allowing the study of the interactions among all pairs of livestock holdings that are formed following the movement of livestock. As a result, important holdings, which are central in the flow of animals, may be identified. The objective of this paper was to show two examples of uses of network analysis: (1) to estimate potential epidemic size following routine livestock movements during the silent spread phase of a highly contagious disease, and (2) to generate production types to specify the contact structure among livestock holdings in disease simulation models. Materials and methods We used network analysis techniques using monthly networks of adult dairy cow movements among dairy farms in Ontario (years 2004-2006) and beef cattle movements among all livestock holdings in Region XI, Chile (year 2007). Potential epidemic size was calculated in Ontario using an approach called “infection chain”. We used the degree distributions to classify beef farms into production types required in Chile. Results The monthly networks of livestock movements were highly fragmented throughout the year in Canada (mean=0.997, sd=0.001) and in Chile (mean=0.996, sd=0.002). The median monthly maximal potential epidemic size in Ontario included 13-15 farms. Four production types were created to simulate the spread of FMD in Chile: non-sellers, non-buyers, buyers-sellers and markets. Discussion The infection chain provided a biologically plausible estimate of potential epidemic size as it accounted for the direction of shipments and the time sequence of these shipments. Using the degree distributions also allowed modellers to classify farms according to their movement patterns and according to their management practices.