Optimizing surveillance for livestock disease spreading through animal movements

Abstract

The spatial propagation of many livestock infectious diseases critically depends on the animal movements among premises; so the knowledge of movement data may help us to detect, manage and control an outbreak. The identification of robust spreading features of the system is however hampered by the temporal dimension characterizing population interactions through movements. Traditional centrality measures do not provide relevant information as results strongly fluctuate in time and outbreak properties heavily depend on geotemporal initial conditions. By focusing on the case study of cattle displacements in Italy, we aim at characterizing livestock epidemics in terms of robust features useful for planning and control, to deal with temporal fluctuations, sensitivity to initial conditions and missing information during an outbreak. Through spatial disease simulations, we detect spreading paths that are stable across different initial conditions, allowing the clustering of the seeds and reducing the epidemic variability. Paths also allow us to identify premises, called sentinels, having a large probability of being infected and providing critical information on the outbreak origin, as encoded in the clusters. This novel procedure provides a general framework that can be applied to specific diseases, for aiding risk assessment analysis and informing the design of optimal surveillance systems.

Epidemiological modelling for the assessment of bovine tuberculosis surveillance in the dairy farm network in Emilia-Romagna (Italy)

“…A priori information on the network structure of cattle exchanges during non-emergency periods can help orienting control strategies to prevent epidemics in areas characterized by a high number of industrial farms (Bajardi et al, 2012;Gilbert et al, 2005). Our framework, based on a mathematical modelling approach, provided decision-makers with a powerful cost-effective tool to assess the effectiveness of the current bTB surveillance system in ER, by highlighting strengths and weaknesses its different components.…”

“…We simulated disease dynamics on a static network, ignoring seasonality in cattle movement and the dynamical nature of cattle trade. Dynamical networks can effectively represent the time-varying structure of the network -which is crucial to simulate the dynamics of acute and highly contagious diseases, such as Food-and-Mouth Disease (Bajardi et al, 2011(Bajardi et al, , 2012Vernon and Keeling, 2009). However, since bTB is characterized by slow infection dynamics and it is able to remain unnoticed for a long period of time, we believe that it is not essential to track the small structural changes occurring on a short time scale for suitably describe bTB dynamics.…”

Optimal surveillance network design: a value of information model

“…For this structured anisotropy related to the nonrandom and non regular structure of the mobility network, we believe that random and regular surveillance network are suboptimal surveillance networks. Vice versa, small-world surveillance networks such http://www.casmodeling.com/content/2/1/6 as power-law networks capture the small-world features of mobility networks (Bajardi et al 2012), thus they are able to monitor outbreak evolution over space and time very accurately. Figure 6 also shows the result of GSUA in which uncertainty is considered for SIR and mobility model input factors.…”

“…All explored surveillance networks correspond to different reported outbreak patterns. Other approaches for designing surveillance have been developed in the past; for instance (Bajardi et al 2012) developed a model for robust outbreak cluster detection based on network features. While such model has an optimization component, deep uncertainty in outbreak reporting (that is uncertainty related to report cases at a selected site and time) is not considered and the detection of clusters do not detect outbreak sources.…”

Cattle movement network, herd size, and bovine brucellosis in the State of Mato Grosso, Brazil

“…The first of such studies were conducted by Bigras-Poulin et al (2006), Christley et al (2003), Corner et al (2003), and Webb and Sauter-Louis (2002). The network analysis currently applied to veterinary epidemiology goes beyond a simple description of the pattern of animal movements and is applied: to the mathematical modelling of the spread of disease within a network (BAJARDI et al, 2012); to provide support for risk analysis and risk-based sampling based on the detection of livestock production zones (GRISI-FILHO et al, Cattle movement network, herd size, and bovine brucellosis in the State of Mato Grosso, Brazil 2013; LENTZ et al, 2011); and to the surveillance and control of infectious diseases SCHÄRRER et al, 2015).…”