Bayesian latent-class modelling of quarantine testing procedures for American Bison (Bison bison) in the Greater Yellowstone Area to determine Brucella abortus freedom.

OBJECTIVE: American bison (Bison bison) quarantine protocols were established to prevent transmission of brucellosis outside the Greater Yellowstone Area, while allowing for distribution of wild bison for conservation and cultural purposes. Quarantine standards require rigorous testing over 900 days which has led to the release of over 200 bison to Native American tribes. Standards were evaluated using 15 years of laboratory and management data to minimize the burden of testing and increase the number of brucellosis-free bison available for distribution. ANIMALS: All bison (n = 578) from Yellowstone National Park were corralled by the National Park Service and United States Department of Agriculture. PROCEDURES: A statistical and management evaluation of the bison quarantine program was performed. Bayesian latent-class modeling was used to predict the probability of nondetection of a seroreactor at various time points, as well as the probability of seroconversion by days in quarantine. RESULTS: At 300 days, 1 in 1,000 infected bison (0.0014 probability) would not be detected but could potentially seroconvert; the seroconversion model predicted 99.9% would seroconvert by day 294, and 12.8% of bison enrolled in quarantine would seroconvert over time. Using a 300-day quarantine period, it would take 30 years to potentially miss 1 seroreactor out of over 8,000 bison enrolled in the quarantine program. CLINICAL RELEVANCE: Reducing the quarantine program requirements from over 900 days to 300 days would allow management of quarantined bison in coordination with seasonal movement of bison herds and triple the number of brucellosis-free bison available for distribution.

The Importance of Livestock Demography and Infrastructure in Driving Foot and Mouth Disease Dynamics.

Transboundary animal diseases, such as foot and mouth disease (FMD) pose a significant and ongoing threat to global food security. Such diseases can produce large, spatially complex outbreaks. Mathematical models are often used to understand the spatio-temporal dynamics and create response plans for possible disease introductions. Model assumptions regarding transmission behavior of premises and movement patterns of livestock directly impact our understanding of the ecological drivers of outbreaks and how to best control them. Here, we investigate the impact that these assumptions have on model predictions of FMD outbreaks in the U.S. using models of livestock shipment networks and disease spread. We explore the impact of changing assumptions about premises transmission behavior, both by including within-herd dynamics, and by accounting for premises type and increasing the accuracy of shipment predictions. We find that the impact these assumptions have on outbreak predictions is less than the impact of the underlying livestock demography, but that they are important for investigating some response objectives, such as the impact on trade. These results suggest that demography is a key ecological driver of outbreaks and is critical for making robust predictions but that understanding management objectives is also important when making choices about model assumptions.

Modeling nation-wide U.S. swine movement networks at the resolution of the individual premises.

The spread of infectious livestock diseases is a major cause for concern in modern agricultural systems. In the dynamics of the transmission of such diseases, movements of livestock between herds play an important role. When constructing mathematical models used for activities such as forecasting epidemic development, evaluating mitigation strategies, or determining important targets for disease surveillance, including between-premises shipments is often a necessity. In the United States (U.S.), livestock shipment data is not routinely collected, and when it is, it is not readily available and mostly concerned with between-state shipments. To bridge this gap in knowledge and provide insight into the complete livestock shipment network structure, we have developed the U.S. Animal Movement Model (USAMM). Previously, USAMM has only existed for cattle shipments, but here we present a version for domestic swine. This new version of USAMM consists of a Bayesian model fit to premises demography, county-level livestock industry variables, and two limited data sets of between-state swine movements. The model scales up the data to simulate nation-wide networks of both within- and between-state shipments at the level of individual premises. Here we describe this shipment model in detail and subsequently explore its usefulness with a rudimentary predictive model of the prevalence of porcine epidemic diarrhea virus (PEDv) across the U.S. Additionally, in order to promote further research on livestock disease and other topics involving the movements of swine in the U.S., we also make 250 synthetic premises-level swine shipment networks with complete coverage of the entire conterminous U.S. freely available to the research community as a useful surrogate for the absent shipment data.

Assessing intrastate shipments from interstate data and expert opinion.

Live animal shipments are a potential route for transmitting animal diseases between holdings and are crucial when modelling spread of infectious diseases. Yet, complete contact networks are not available in all countries, including the USA. Here, we considered a 10% sample of Interstate Certificate of Veterinary Inspections from 1 year (2009). We focused on distance dependence in contacts and investigated how different functional forms affect estimates of unobserved intrastate shipments. To further enhance our predictions, we included responses from an expert elicitation survey about the proportion of shipments moving intrastate. We used hierarchical Bayesian modelling to estimate parameters describing the kernel and effects of expert data. We considered three functional forms of spatial kernels and the inclusion or exclusion of expert data. The resulting six models were ranked by widely applicable information criterion (WAIC) and deviance information criterion (DIC) and evaluated through within- and out-of-sample validation. We showed that predictions of intrastate shipments were mildly influenced by the functional form of the spatial kernel but kernel shapes that permitted a fat tail at large distances while maintaining a plateau-shaped behaviour at short distances better were preferred. Furthermore, our study showed that expert data may not guarantee enhanced predictions when expert estimates are disparate.

Realistic assumptions about spatial locations and clustering of premises matter for models of foot-and-mouth disease spread in the United States.

Spatially explicit livestock disease models require demographic data for individual farms or premises. In the U.S., demographic data are only available aggregated at county or coarser scales, so disease models must rely on assumptions about how individual premises are distributed within counties. Here, we addressed the importance of realistic assumptions for this purpose. We compared modeling of foot and mouth disease (FMD) outbreaks using simple randomization of locations to premises configurations predicted by the Farm Location and Agricultural Production Simulator (FLAPS), which infers location based on features such as topography, land-cover, climate, and roads. We focused on three premises-level Susceptible-Exposed-Infectious-Removed models available from the literature, all using the same kernel approach but with different parameterizations and functional forms. By computing the basic reproductive number of the infection (R0) for both FLAPS and randomized configurations, we investigated how spatial locations and clustering of premises affects outbreak predictions. Further, we performed stochastic simulations to evaluate if identified differences were consistent for later stages of an outbreak. Using Ripley's K to quantify clustering, we found that FLAPS configurations were substantially more clustered at the scales relevant for the implemented models, leading to a higher frequency of nearby premises compared to randomized configurations. As a result, R0 was typically higher in FLAPS configurations, and the simulation study corroborated the pattern for later stages of outbreaks. Further, both R0 and simulations exhibited substantial spatial heterogeneity in terms of differences between configurations. Thus, using realistic assumptions when de-aggregating locations based on available data can have a pronounced effect on epidemiological predictions, affecting if, where, and to what extent FMD may invade the population. We conclude that methods such as FLAPS should be preferred over randomization approaches.

Effects of regional differences and demography in modelling foot-and-mouth disease in cattle at the national scale.

Foot-and-mouth disease (FMD) is a fast-spreading viral infection that can produce large and costly outbreaks in livestock populations. Transmission occurs at multiple spatial scales, as can the actions used to control outbreaks. The US cattle industry is spatially expansive, with heterogeneous distributions of animals and infrastructure. We have developed a model that incorporates the effects of scale for both disease transmission and control actions, applied here in simulating FMD outbreaks in US cattle. We simulated infection initiating in each of the 3049 counties in the contiguous US, 100 times per county. When initial infection was located in specific regions, large outbreaks were more likely to occur, driven by infrastructure and other demographic attributes such as premises clustering and number of cattle on premises. Sensitivity analyses suggest these attributes had more impact on outbreak metrics than the ranges of estimated disease parameter values. Additionally, although shipping accounted for a small percentage of overall transmission, areas receiving the most animal shipments tended to have other attributes that increase the probability of large outbreaks. The importance of including spatial and demographic heterogeneity in modelling outbreak trajectories and control actions is illustrated by specific regions consistently producing larger outbreaks than others.

Spatio-temporal patterns and characteristics of swine shipments in the U.S. based on Interstate Certificates of Veterinary Inspection.

Domestic swine production in the United States is a critical economic and food security industry, yet there is currently no large-scale quantitative assessment of swine shipments available to support risk assessments. In this study, we provide a national-level characterization of the swine industry by quantifying the demographic (i.e. age, sex) patterns, spatio-temporal patterns, and the production diversity within swine shipments. We characterize annual networks of swine shipments using a 30% stratified sample of Interstate Certificates of Veterinary Inspection (ICVI), which are required for the interstate movement of agricultural animals. We used ICVIs in 2010 and 2011 from eight states that represent 36% of swine operations and 63% of the U.S. swine industry. Our analyses reflect an integrated and spatially structured industry with high levels of spatial heterogeneity. Most shipments carried young swine for feeding or breeding purposes and carried a median of 330 head (range: 1-6,500). Geographically, most shipments went to and were shipped from Iowa, Minnesota, and Nebraska. This work, therefore, suggests that although the swine industry is variable in terms of its size and type of swine, counties in states historically known for breeding and feeding operations are consistently more central to the shipment network.

Estimating and exploring the proportions of inter- and intrastate cattle shipments in the United States.

Mathematical models are key tools for the development of surveillance, preparedness and response plans for the potential events of emerging and introduced foreign animal diseases. Creating these types of plans requires data; when data are incomplete, mathematical models can help fill in missing information, provided they are informed by the data that are available. In the United States, the most complete national-scale data available on cattle shipments are based on Interstate Certificates of Veterinary Inspection, which track the shipment of cattle between states; data on intrastate cattle shipments are lacking. Here we develop four new datasets on intrastate cattle shipments in the U.S., including an expert elicitation survey covering 19 states and territories and three state-level brand inspection data sets. The expert elicitation survey provides estimates on the proportion of shipments that travel interstate over multiple regions of the U.S. These survey data also identify differences in shipment patterns between regions, cattle commodity types, and sectors of the cattle industry. These survey data cover more states than any other source of intrastate data; however, one limitation of these data is the small number of participating experts in many of the states, only seven of the 19 responding states and territories had a group size of three or larger. The brand data sets include origin and destination information for both intra- and interstate shipments. These data, therefore, also provide detailed information on the proportion of interstate shipments in three Western states, including the temporal and geographic variation in shipments. Because the survey and brand data overlap in the Western U.S., they can be compared. We find that in the Western U.S. the expert estimates of the overall proportion of cattle shipments matched the brand data well. However, the experts estimated that there would be larger differences in beef and dairy shipments than the brand data show. This suggests the cattle industries in the West may be sending similar proportions of commodity specific cattle shipments over state lines. We additionally used the expert survey data to explore how differences in the proportion of interstate shipments can change predictions about cattle shipment patterns using the example of model-guided suggestions for targeted surveillance in Texas. Together these four data sets are the most extensive and geographically comprehensive information to date on intrastate cattle shipments. Additionally, our analyses on predicted shipment patterns suggest that assumptions about intrastate shipments could have consequences for targeted surveillance.

Model-guided suggestions for targeted surveillance based on cattle shipments in the U.S.

Risk-based sampling is an essential component of livestock health surveillance because it targets resources towards sub-populations with a higher risk of infection. Risk-based surveillance in U.S. livestock is limited because the locations of high-risk herds are often unknown and data to identify high-risk herds based on shipments are often unavailable. In this study, we use a novel, data-driven network model for the shipments of cattle in the U.S. (the U.S. Animal Movement Model, USAMM) to provide surveillance suggestions for cattle imported into the U.S. from Mexico. We describe the volume and locations where cattle are imported and analyze their predicted shipment patterns to identify counties that are most likely to receive shipments of imported cattle. Our results suggest that most imported cattle are sent to relatively few counties. Surveillance at 10 counties is predicted to sample 22-34% of imported cattle while surveillance at 50 counties is predicted to sample 43%-61% of imported cattle. These findings are based on the assumption that USAMM accurately describes the shipments of imported cattle because their shipments are not tracked separately from the remainder of the U.S. herd. However, we analyze two additional datasets - Interstate Certificates of Veterinary Inspection and brand inspection data - to ensure that the characteristics of potential post-import shipments do not change on an annual scale and are not dependent on the dataset informing our analyses. Overall, these results highlight the utility of USAMM to inform targeted surveillance strategies when complete shipment information is unavailable.

Shifting brucellosis risk in livestock coincides with spreading seroprevalence in elk.

Tracking and preventing the spillover of disease from wildlife to livestock can be difficult when rare outbreaks occur across large landscapes. In these cases, broad scale ecological studies could help identify risk factors and patterns of risk to inform management and reduce incidence of disease. Between 2002 and 2014, 21 livestock herds in the Greater Yellowstone Area (GYA) were affected by brucellosis, a bacterial disease caused by Brucella abortus, while no affected herds were detected between 1990 and 2001. Using a Bayesian analysis, we examined several ecological covariates that may be associated with affected livestock herds across the region. We showed that livestock risk has been increasing over time and expanding outward from the historical nexus of brucellosis in wild elk on Wyoming's feeding grounds where elk are supplementally fed during the winter. Although elk were the presumed source of cattle infections, occurrences of affected livestock herds were only weakly associated with the density of seropositive elk across the GYA. However, the shift in livestock risk did coincide with recent increases in brucellosis seroprevalence in unfed elk populations. As increasing brucellosis in unfed elk likely stemmed from high levels of the disease in fed elk, disease-related costs of feeding elk have probably been incurred across the entire GYA, rather than solely around the feeding grounds. Our results suggest that focused disease mitigation in areas where seroprevalence in unfed elk is high could reduce the spillover of brucellosis to livestock. We also highlight the need to better understand the epidemiology of spillover events with detailed histories of disease testing, calving, and movement of infected livestock. Finally, we recommend using case-control studies to investigate local factors important to livestock risk.

Mapping U.S. cattle shipment networks: Spatial and temporal patterns of trade communities from 2009 to 2011.

The application of network analysis to cattle shipments broadens our understanding of shipment patterns beyond pairwise interactions to the network as a whole. Such a quantitative description of cattle shipments in the U.S. can identify trade communities, describe temporal shipment patterns, and inform the design of disease surveillance and control strategies. Here, we analyze a longitudinal dataset of beef and dairy cattle shipments from 2009 to 2011 in the United States to characterize communities within the broader cattle shipment network, which are groups of counties that ship mostly to each other. Because shipments occur over time, we aggregate the data at various temporal scales to examine the consistency of network and community structure over time. Our results identified nine large (>50 counties) communities based on shipments of beef cattle in 2009 aggregated into an annual network and nine large communities based on shipments of dairy cattle. The size and connectance of the shipment network was highly dynamic; monthly networks were smaller than yearly networks and revealed seasonal shipment patterns consistent across years. Comparison of the shipment network over time showed largely consistent shipping patterns, such that communities identified on annual networks of beef and diary shipments from 2009 still represented 41-95% of shipments in monthly networks from 2009 and 41-66% of shipments from networks in 2010 and 2011. The temporal aspects of cattle shipments suggest that future applications of the U.S. cattle shipment network should consider seasonal shipment patterns. However, the consistent within-community shipping patterns indicate that yearly communities could provide a reasonable way to group regions for management.

Sources of bovine tuberculosis in the United States.

Despite control and eradication efforts, bovine tuberculosis continues to be identified at low levels among cattle in the United States. We evaluated possible external sources of infection by characterizing the genetic relatedness of bovine tuberculosis from a national database of reported infections, comparing strains circulating among US cattle with those of imported cattle, and farmed and wild cervids. Farmed cervids maintained a genetically distinct Mycobacterium bovis strain, and cattle occasionally became infected with this strain. In contrast, wild cervids acted as an epidemiologically distinct group, instead hosting many of the same strains found in cattle, and the data did not show a clear transmission direction. Cattle from Mexico hosted a higher overall richness of strains than US cattle, and many of those strains were found in both US and Mexican cattle. However, these two populations appeared to be well-mixed with respect to their M. bovis lineages, and higher resolution data is necessary to infer the direction of recent transmission. Overall patterns of both host and geographic distributions were highly variable among strains, suggesting that different sources or transmission mechanisms are contributing to maintaining different strains.

The impact of movements and animal density on continental scale cattle disease outbreaks in the United States.

Globalization has increased the potential for the introduction and spread of novel pathogens over large spatial scales necessitating continental-scale disease models to guide emergency preparedness. Livestock disease spread models, such as those for the 2001 foot-and-mouth disease (FMD) epidemic in the United Kingdom, represent some of the best case studies of large-scale disease spread. However, generalization of these models to explore disease outcomes in other systems, such as the United States's cattle industry, has been hampered by differences in system size and complexity and the absence of suitable livestock movement data. Here, a unique database of US cattle shipments allows estimation of synthetic movement networks that inform a near-continental scale disease model of a potential FMD-like (i.e., rapidly spreading) epidemic in US cattle. The largest epidemics may affect over one-third of the US and 120,000 cattle premises, but cattle movement restrictions from infected counties, as opposed to national movement moratoriums, are found to effectively contain outbreaks. Slow detection or weak compliance may necessitate more severe state-level bans for similar control. Such results highlight the role of large-scale disease models in emergency preparedness, particularly for systems lacking comprehensive movement and outbreak data, and the need to rapidly implement multi-scale contingency plans during a potential US outbreak.

A national-scale picture of U.S. cattle movements obtained from Interstate Certificate of Veterinary Inspection data.

We present the first comprehensive description of how shipments of cattle connect the geographic extent and production diversity of the United States cattle industry. We built a network of cattle movement from a state-stratified 10% systematic sample of calendar year 2009 Interstate Certificates of Veterinary Inspection (ICVI) data. ICVIs are required to certify the apparent health of cattle moving across state borders and allow us to examine cattle movements at the county scale. The majority of the ICVI sample consisted of small shipments (<20 head) moved for feeding and beef production. Geographically, the central plains states had the most connections, correlated to feeding infrastructure. The entire nation was closely connected when interstate movements were summarized at the state level. At the county-level, the U.S. is still well connected geographically, but significant heterogeneities in the location and identity of counties central to the network emerge. Overall, the network of interstate movements is described by a hub structure, with a few counties sending or receiving extremely large numbers of shipments and many counties sending and receiving few shipments. The county-level network also has a very low proportion of reciprocal movements, indicating that high-order network properties may be better at describing a county's importance than simple summaries of the number of shipments or animals sent and received. We suggest that summarizing cattle movements at the state level homogenizes the network and a county level approach is most appropriate for examining processes influenced by cattle shipments, such as economic analyses and disease outbreaks.

Assessment of paper interstate certificates of veterinary inspection used to support disease tracing in cattle.

OBJECTIVE: To evaluate the differences among each state's Interstate Certificate of Veterinary Inspection (ICVI) form and the legibility of data on paper ICVIs used to support disease tracing in cattle. DESIGN: Descriptive retrospective cross-sectional study. SAMPLE: Examples of ICVIs from 50 states and 7,630 randomly sampled completed paper ICVIs for cattle from 48 states. PROCEDURES: Differences among paper ICVI forms from all 50 states were determined. Sixteen data elements were selected for further evaluation of their value in tracing cattle. Completed paper ICVIs for interstate cattle exports in 2009 were collected from 48 states. Each of the 16 data elements was recorded as legible, absent, or illegible on forms completed by accredited veterinarians, and results were summarized by state. Mean values for legibility at the state level were used to estimate legibility of data at the national level. RESULTS: ICVIs were inconsistent among states in regard to data elements requested and availability of legible records. A mean ± SD of 70.0 ± 22.1% of ICVIs in each state had legible origin address information. Legible destination address information was less common, with 55.0 ± 21.4% of records complete. Incomplete address information was most often a result of the field having been left blank. Official animal identification was present on 33.1% of ICVIs. CONCLUSIONS AND CLINICAL RELEVANCE: The inconsistency among state ICVI forms and quality of information provided on paper ICVIs could lead to delays and the need for additional resources to trace cattle, which could result in continued spread of disease. Standardized ICVIs among states and more thorough recording of information by accredited veterinarians or expanded usage of electronic ICVIs could enhance traceability of cattle during an outbreak.

A bayesian approach for modeling cattle movements in the United States: scaling up a partially observed network.

Networks are rarely completely observed and prediction of unobserved edges is an important problem, especially in disease spread modeling where networks are used to represent the pattern of contacts. We focus on a partially observed cattle movement network in the U.S. and present a method for scaling up to a full network based on bayesian inference, with the aim of informing epidemic disease spread models in the United States. The observed network is a 10% state stratified sample of Interstate Certificates of Veterinary Inspection that are required for interstate movement; describing approximately 20,000 movements from 47 of the contiguous states, with origins and destinations aggregated at the county level. We address how to scale up the 10% sample and predict unobserved intrastate movements based on observed movement distances. Edge prediction based on a distance kernel is not straightforward because the probability of movement does not always decline monotonically with distance due to underlying industry infrastructure. Hence, we propose a spatially explicit model where the probability of movement depends on distance, number of premises per county and historical imports of animals. Our model performs well in recapturing overall metrics of the observed network at the node level (U.S. counties), including degree centrality and betweenness; and performs better compared to randomized networks. Kernel generated movement networks also recapture observed global network metrics, including network size, transitivity, reciprocity, and assortativity better than randomized networks. In addition, predicted movements are similar to observed when aggregated at the state level (a broader geographic level relevant for policy) and are concentrated around states where key infrastructures, such as feedlots, are common. We conclude that the method generally performs well in predicting both coarse geographical patterns and network structure and is a promising method to generate full networks that incorporate the uncertainty of sampled and unobserved contacts.