Assessment and quantification of post-weaning multi-systemic wasting syndrome severity at farm level

Post-weaning multi-systemic wasting syndrome (PMWS) causes major economic losses for the English pig industry and severity of clinical signs and economic impact vary considerably between affected farms. We present here a novel approach to quantify severity of PMWS based on morbidity and mortality data and presence of porcine circovirus type 2 (PCV2). In 2008-2009, 147 pig farms across England, non-vaccinating for PCV2, were enrolled in a cross-sectional study. Factor analysis was used to generate variables representing biologically meaningful aspects of variation among qualitative and quantitative morbidity variables. Together with other known variables linked to PMWS, the resulting factors were included in a principal component analysis (PCA) to derive an algorithm for PMWS severity. Factor analysis resulted in two factors: Morbidity Factor 1 (MF1) representing mainly weaner and grower morbidity, and Morbidity Factor 2 (MF2) which mainly reflects variation in finisher morbidity. This indicates that farms either had high morbidity mainly in weaners/growers or mainly in finishers. Subsequent PCA resulted in the extraction of one component representing variation in MF1, post-weaning mortality and percentage of PCV2 PCR positive animals. Component scores were normalised to a value range from 0 to 10 and farms classified into: non or slightly affected farms with a score <4, moderately affected farms with scores 4-6.5 and highly affected farms with a score >6.5. The identified farm level PMWS severities will be used to identify risk factors related to these, to assess the efficacy of PCV2 vaccination and investigating the economic impact of potential control measures.

Evaluating the control of HPAIV H5N1 in Vietnam: virus transmission within infected flocks reported before and after vaccination

Background

Currently, the highly pathogenic avian influenza virus (HPAIV) of the subtype H5N1 is believed to have reached an endemic cycle in Vietnam. We used routine surveillance data on HPAIV H5N1 poultry outbreaks in Vietnam to estimate and compare the within-flock reproductive number of infection (R₀) for periods before (second epidemic wave, 2004-5; depopulation-based disease control) and during (fourth epidemic wave, beginning 2007; vaccination-based disease control) vaccination.

Results

Our results show that infected premises (IPs) in the initial (exponential) phases of outbreak periods have the highest R₀ estimates. The IPs reported during the outbreak period when depopulation-based disease control was implemented had higher R₀ estimates than IPs reported during the outbreak period when vaccination-based disease control was used. In the latter period, in some flocks of a defined size and species composition, within-flock transmission estimates were not significantly below the threshold for transmission (R₀ < 1).

Conclusions

Our results indicate that the current control policy based on depopulation plus vaccination has protected the majority of poultry flocks against infection. However, in some flocks the determinants associated with suboptimal protection need to be further investigated as these may explain the current pattern of infection in animal and human populations.

Risk mapping of Rinderpest sero-prevalence in Central and Southern Somalia based on spatial and network risk factors

Background

In contrast to most pastoral systems, the Somali livestock production system is oriented towards domestic trade and export with seasonal movement patterns of herds/flocks in search of water and pasture and towards export points. Data from a rinderpest survey and other data sources have been integrated to explore the topology of a contact network of cattle herds based on a spatial proximity criterion and other attributes related to cattle herd dynamics. The objective of the study is to integrate spatial mobility and other attributes with GIS and network approaches in order to develop a predictive spatial model of presence of rinderpest.

Results

A spatial logistic regression model was fitted using data for 562 point locations. It includes three statistically significant continuous-scale variables that increase the risk of rinderpest: home range radius, herd density and clustering coefficient of the node of the network whose link was established if the sum of the home ranges of every pair of nodes was equal or greater than the shortest distance between the points. The sensitivity of the model is 85.1% and the specificity 84.6%, correctly classifying 84.7% of the observations. The spatial autocorrelation not accounted for by the model is negligible and visual assessment of a semivariogram of the residuals indicated that there was no undue amount of spatial autocorrelation. The predictive model was applied to a set of 6176 point locations covering the study area. Areas at high risk of having serological evidence of rinderpest are located mainly in the coastal districts of Lower and Middle Juba, the coastal area of Lower Shabele and in the regions of Middle Shabele and Bay. There are also isolated spots of high risk along the border with Kenya and the southern area of the border with Ethiopia.

Conclusions

The identification of point locations and areas with high risk of presence of rinderpest and their spatial visualization as a risk map will be useful for informing the prioritization of disease surveillance and control activities for rinderpest in Somalia. The methodology applied here, involving spatial and network parameters, could also be applied to other diseases and/or species as part of a standardized approach for the design of risk-based surveillance activities in nomadic pastoral settings.

Associations between attributes of live poultry trade and HPAI H5N1 outbreaks: a descriptive and network analysis study in northern Vietnam

Background

The structure of contact between individuals plays an important role in the incursion and spread of contagious diseases in both human and animal populations. In the case of avian influenza, the movement of live birds is a well known risk factor for the geographic dissemination of the virus among poultry flocks. Live bird markets (LBM's) contribute to the epidemiology of avian influenza due to their demographic characteristics and the presence of HPAI H5N1 virus lineages. The relationship between poultry producers and live poultry traders (LPT's) that operate in LBM's has not been adequately documented in HPAI H5N1-affected SE Asian countries. The aims of this study were to document and study the flow of live poultry in a poultry trade network in northern Vietnam, and explore its potential role in the risk for HPAI H5N1 during 2003 to 2006.

Results

Our results indicate that LPT's trading for less than a year and operating at retail markets are more likely to source poultry from flocks located in communes with a past history of HPAI H5N1 outbreaks during 2003 to 2006 than LPT's trading longer than a year and operating at wholesale markets. The results of the network analysis indicate that LPT's tend to link communes of similar infection status.

Conclusions

Our study provides evidence which can be used for informing policies aimed at encouraging more biosecure practices of LPT's operating at authorised LBM's. The results suggest that LPT's play a role in HPAI H5N1 transmission and may contribute to perpetuating HPAI H5N1 virus circulation amongst certain groups of communes. The impact of current disease prevention and control interventions could be enhanced by disseminating information about outbreak risk and the implementation of a formal data recording scheme at LBM's for all incoming and outgoing LPT's.

Meticillin-resistant Staphylococcus aureus carriage in UK veterinary staff and owners of infected pets: new risk groups

Meticillin-resistant Staphylococcus aureus (MRSA) nasal carriage on admission to hospital remains one of the most important risk factors for subsequent infection. Identification of high risk groups for MRSA carriage is vital for the success of infection control programmes. Veterinary staff may be one such risk group but little is known about pet owners and the role of contact with infected pets. As part of a UK-wide case-control study investigating risk factors for MRSA infection in dogs and cats between 2005 and 2008, 608 veterinary staff and pet owners in contact with 106 MRSA and 91 meticillin-susceptible S. aureus (MSSA)-infected pets were screened for S. aureus nasal carriage. Laboratory isolation and characterisation included salt broth enrichment, standard and automated microbiological tests, demonstration of the S. aureus-specific thermonuclease gene (nuc) and of mecA, and polymerase chain reaction-based lineage characterisation. MRSA carriage was 12.3% in veterinarians attending MRSA-infected animals and 7.5% in their owners. In the MSSA control group, MRSA carriage was 4.8% in veterinary staff and 0% in owners. Veterinary staff carried MRSA more frequently than owners (odds ratio: 2.33; 95% confidence interval: 1.10-4.93). All MRSA from humans and all but one animal MRSA were CC22 or CC30, typical for hospital MRSA in the UK. This study indicates for the first time an occupational risk for MRSA carriage in small animal general practitioners. Veterinary staff and owners of MRSA-infected pets are high risk groups for MRSA carriage despite not having direct hospital links. Strategies to break the cycle of MRSA infection must take these potential new reservoirs into account.

Biochemical markers of bone metabolism and risk of dorsal metacarpal disease in 2-year-old Thoroughbreds

REASONS FOR PERFORMING STUDY

Dorsal metacarpal disease (DMD) is a common problem in 2-year-old racehorses and results in loss of a significant number of days from training. Biochemical markers of bone cell activity measured early in the training season could have value for identifying 2-year-old Thoroughbred racehorses that develop DMD.

OBJECTIVES

To determine the association between serum concentrations of osteocalcin, the carboxyterminal propeptide of type I collagen (PICP) and the carboxyterminal cross-linked telopeptide of type I collagen (ICTP) measured early in the training season and the risk of DMD.

METHODS

Blood samples were collected from 165 two-year-old Thoroughbreds during late November/early December. Osteocalcin and PICP were measured as markers of bone formation, and ICTP as a marker of bone resorption. Training and veterinary records for each horse were monitored over the following training/racing season (10 months). Cases were defined as an episode where signs of DMD were sufficiently severe for a horse to miss at least 5 consecutive days of training. Classification tree and logistic regression analysis were used to identify the most important factors suitable for prediction of DMD risk.

RESULTS

There were 24 cases of DMD during the season (14.6% cumulative incidence), with an average time to recognition of approximately 6 months (May). The earliest recognised case was in February and the latest in September. Osteocalcin and ICTP concentrations in the early stages of the training season were significantly higher in horses that subsequently developed DMD (P = 0.017 and 0.019, respectively). DMD cases were also significantly older compared to noncases (21.04 vs. 20.44 months, P = 0.023). Using a multivariable logistic regression model, it was possible to postulate a set of diagnostic rules to predict the likelihood of DMD injury during the season. This suggested that horses with ICTP concentrations above 12365 ug/l and older than 20.5 months are 2.6 times more likely to develop DMD.

CONCLUSIONS

The measurement of the bone resorption marker ICTP could be useful for identification of 2-year-olds at increased risk of developing DMD.

POTENTIAL RELEVANCE

These findings, together with other strategies such as modification of training regimens, e.g. early introduction of short distances of high-speed exercise into the training programme, could help reduce the days lost to training as a result of DMD.

Recent advances in the analysis of behavioural organization and interpretation as indicators of animal welfare

While the incorporation of mathematical and engineering methods has greatly advanced in other areas of the life sciences, they have been under-utilized in the field of animal welfare. Exceptions are beginning to emerge and share a common motivation to quantify 'hidden' aspects in the structure of the behaviour of an individual, or group of animals. Such analyses have the potential to quantify behavioural markers of pain and stress and quantify abnormal behaviour objectively. This review seeks to explore the scope of such analytical methods as behavioural indicators of welfare. We outline four classes of analyses that can be used to quantify aspects of behavioural organization. The underlying principles, possible applications and limitations are described for: fractal analysis, temporal methods, social network analysis, and agent-based modelling and simulation. We hope to encourage further application of analyses of behavioural organization by highlighting potential applications in the assessment of animal welfare, and increasing awareness of the scope for the development of new mathematical methods in this area.

African swine fever: how can global spread be prevented?

African swine fever (ASF) is a devastating haemorrhagic fever of pigs with mortality rates approaching 100 per cent. It causes major economic losses, threatens food security and limits pig production in affected countries. ASF is caused by a large DNA virus, African swine fever virus (ASFV). There is no vaccine against ASFV and this limits the options for disease control. ASF has been confined mainly to sub-Saharan Africa, where it is maintained in a sylvatic cycle and/or among domestic pigs. Wildlife hosts include wild suids and arthropod vectors. The relatively small numbers of incursions to other continents have proven to be very difficult to eradicate. Thus, ASF remained endemic in the Iberian peninsula until the mid-1990s following its introductions in 1957 and 1960 and the disease has remained endemic in Sardinia since its introduction in 1982. ASF has continued to spread within Africa to previously uninfected countries, including recently the Indian Ocean islands of Madagascar and Mauritius. Given the continued occurrence of ASF in sub-Saharan Africa and increasing global movements of people and products, it is not surprising that further transcontinental transmission has occurred. The introduction of ASF to Georgia in the Caucasus in 2007 and dissemination to neighbouring countries emphasizes the global threat posed by ASF and further increases the risks to other countries. We review the mechanisms by which ASFV is maintained within wildlife and domestic pig populations and how it can be transmitted. We then consider the risks for global spread of ASFV and discuss possibilities of how disease can be prevented.

PrP genotype: a flock-level risk factor for scrapie?

Previous epidemiological studies of risk factors for classical scrapie at flock level have identified a variety of management and purchase related variables, along with increased flock size and, in some cases, breed effects. Although known as a risk factor at the individual animal level, PrP genotype frequencies at flock level have not yet been studied. In an unmatched case-control study, three measures of flock-level prion protein (PrP) frequency estimates were investigated with respect to the scrapie status of the flock in 293 British sheep flocks (195 control flocks and 98 case flocks). Flocks with positive frequencies (more than 0 per cent) of two genotypes (VRQ/VRQ and AHQ/VRQ), large frequencies (more than 10 per cent) of the ARR/VRQ genotype, and large frequencies (more than 5.2 per cent) of the VRQ allele were at increased odds of being affected with clinical classical scrapie. When adjusted for flock size, breed and sampling strategy the genotype and allele effects remained, except that for flocks with positive frequencies of VRQ/VRQ. The known effect of increased risk with increased flock size was confirmed. A measure of the flock PrP genotype frequency profile should thus be included in studies of risk factors for scrapie. It could also be integrated into risk-based surveillance strategies for identification of "at-risk-of scrapie" flocks.

Multivariate analysis of management and biosecurity practices in smallholder pig farms in Madagascar

A cross-sectional study was carried out in 2005 and 2006 in three geographical areas of Madagascar to investigate and differentiate swine farm management and biosecurity practices in smallholder farming communities. Questionnaire data from a total of 709 pig farms were analysed using multiple factor analysis (MFA) and hierarchical cluster analysis (HCA). Variables describing management and biosecurity practices were organised into five groups: structure of the farm, animal-contacts, person- and vehicle-contacts, feeding, and sanitary aspects. In general, few biosecurity measures were implemented in the pig farms included in the study. Regional differences in management and biosecurity practices emerged from the MFA and were mainly due to, in order of decreasing importance: structure of the farm, sanitary aspects, feeding and animal-contacts and, to a lesser extent, person- and vehicle-contacts. HCA resulted in the differentiation of four distinct types of farms in each of two study areas, Arivonimamo and Marovoay, while no grouping could be identified amongst farms in Ambatondrazaka area. The characterisation of the different types of smallholder pig farms will allow adapting recommendations on husbandry practices and control measures in pig farms of these regions of Madagascar. The development of tailored recommendations is essential for Malagasy smallholders who have limited resources and need to make evidence-based management changes to reduce the risk of contagious diseases in their herds.

Lack of transmission of methicillin-resistant Staphylococcus aureus (MRSA) between apparently healthy dogs in a rescue kennel

Although it is widely accepted that methicillin-resistant Staphylococcus aureus (MRSA) can be transmitted between humans and animals in both directions, little is known about the dynamics of animal-to-animal transfer. This study aimed to investigate aspects of dog-to-dog MRSA transfer in a rescue facility in the South-East of England during an MRSA outbreak. One hundred and twenty-nine apparently healthy dogs, mostly housed in pairs, were swabbed at nasal, oral, axillary and perianal sites. Swabs were enriched in selective broth and staphylococci identified using standard biological methods. MRSA isolates were confirmed by demonstration of the thermonuclease gene (nuc) and mecA. After initial swabbing, a dog excluded from the study design but housed at the same facility was discovered to have a wound infection due to MRSA. MRSA carriage was identified in 10/129 dogs (7.8%) and all isolates were of the same lineage as the one isolated from the infected dog. All carrier dogs lived in shared kennels and their 16 kennel partners sampled negative on two occasions. Concurrently with successful antimicrobial treatment of the infected patient, MRSA carriage resolved spontaneously in all dogs within two weeks. In conclusion, MRSA did not transmit readily between apparently healthy dogs, MRSA carriage was not supported for long periods in a regularly cleaned environment and exposure alone may not lead to MRSA acquisition by dogs without the presence of additional risk factors.

Bone biomarkers and risk of fracture in two- and three-year-old Thoroughbreds

The aim of this study was to determine whether bone biomarkers (osteocalcin, PICP, ICTP and CTX-I) could be used to identify 2- and 3-year-olds at increased risk of fracture in the subsequent flat racing season. It was concluded that these bone biomarkers cannot be used to identify 2- and 3-year-olds that sustain a fracture. Whether bone biomarkers have better predictive value in older horses or when measured serially in the same animal remains to be determined.

Industrial food animal production and global health risks: exploring the ecosystems and economics of avian influenza

Many emerging infectious diseases in human populations are associated with zoonotic origins. Attention has often focused on wild animal reservoirs, but most zoonotic pathogens of recent concern to human health either originate in, or are transferred to, human populations from domesticated animals raised for human consumption. Thus, the ecological context of emerging infectious disease comprises two overlapping ecosystems: the natural habitats and populations of wild animals, and the anthropogenically controlled habitats and populations of domesticated species. Intensive food animal production systems and their associated value chains dominate in developed countries and are increasingly important in developing countries. These systems are characterized by large numbers of animals being raised in confinement with high throughput and rapid turnover. Although not typically recognized as such, industrial food animal production generates unique ecosystems -- environments that may facilitate the evolution of zoonotic pathogens and their transmission to human populations. It is often assumed that confined food animal production reduces risks of emerging zoonotic diseases. This article provides evidence suggesting that these industrial systems may increase animal and public health risks unless there is recognition of the specific biosecurity and biocontainment challenges of the industrial model. Moreover, the economic drivers and constraints faced by the industry and its participants must be fully understood in order to inform preventative policy. In order to more effectively reduce zoonotic disease risk from industrial food animal production, private incentives for the implementation of biosecurity must align with public health interests.

Days lost from training by two- and three-year-old Thoroughbred horses: a survey of seven UK training yards

REASONS FOR PERFORMING STUDY

The first major epidemiological study of injury incidence in the UK flat racing Thoroughbred (TB), published in 1985, found lameness to be the single largest reason for days when horses failed to train. It was considered advisable to ascertain if progress has been made in reducing the problem of musculoskeletal injuries in the intervening period.

OBJECTIVE

To quantify injury incidence and days lost from training by 2- and 3-year-old TBs in UK training yards during 2002 and 2003.

METHODS

One-hundred-and-eighty-two yearling TBs were recruited at the end of 2001 and daily training and injury records maintained over the following 2 training and racing seasons. Days were defined as lost from training when a horse failed to train at a slow canter speed or faster, and could be assigned to one of 4 categories: lameness, medical, traumatic and unknown. The incidence and number of days lost due to specific injuries and medical conditions was determined by further subdividing the lameness and medical categories.

RESULTS

The study period provided a total of 52,601 2-year-old and 29,369 3-year-old days available for training, with 2-year-olds failing to train on a significantly greater proportion of days available than 3-year-olds. Lameness was the most important condition causing horses to miss training, with stress fractures being the most important cause of lameness. Medical conditions were a relatively minor cause of days lost from training, accounting for approximately 5% of untrained days in the 2 age groups.

CONCLUSIONS

In UK flat racehorses there has been little change in the proportion of days lost from training due to lameness over the last 20 years.

POTENTIAL RELEVANCE

This study highlights the need for further efforts to reduce the problem of lameness in the racing TB.

The animal-human interface and infectious disease in industrial food animal production: rethinking biosecurity and biocontainment

Understanding interactions between animals and humans is critical in preventing outbreaks of zoonotic disease. This is particularly important for avian influenza. Food animal production has been transformed since the 1918 influenza pandemic. Poultry and swine production have changed from small-scale methods to industrial-scale operations. There is substantial evidence of pathogen movement between and among these industrial facilities, release to the external environment, and exposure to farm workers, which challenges the assumption that modern poultry production is more biosecure and biocontained as compared with backyard or small holder operations in preventing introduction and release of pathogens. An analysis of data from the Thai government investigation in 2004 indicates that the odds of H5N1 outbreaks and infections were significantly higher in large-scale commercial poultry operations as compared with backyard flocks. These data suggest that successful strategies to prevent or mitigate the emergence of pandemic avian influenza must consider risk factors specific to modern industrialized food animal production.

Risk factors and characteristics of H5N1 Highly Pathogenic Avian Influenza (HPAI) post-vaccination outbreaks

Highly pathogenic avian influenza (HPAI) virus H5N1 is now endemic in South-East Asia but HPAI control methods differ between countries. A widespread HPAI vaccination campaign that started at the end of 2005 in Viet Nam resulted in the cessation of poultry and human cases, but in 2006/2007 severe HPAI outbreaks re-emerged. In this study we investigated the pattern of this first post-vaccination epidemic in southern Viet Nam identifying a spatio-temporal cluster of outbreak occurrence and estimating spatially smoothed incidence rates of HPAI. Spatial risk factors associated with HPAI occurrence were identified. Medium-level poultry density resulted in an increased outbreak risk (Odds ratio (OR) = 5.4, 95% confidence interval (CI): 1.6–18.9) but also climate-vegetation factors played an important role: medium-level normalised difference vegetation indices during the rainy season from May to October were associated with higher risk of HPAI outbreaks (OR = 3.7, 95% CI: 1.7–8.1), probably because temporal flooding might have provided suitable conditions for the re-emergence of HPAI by expanding the virus distribution in the environment and by enlarging areas of possible contacts between domestic waterfowl and wild birds. On the other hand, several agricultural production factors, such as sweet potatoes yield, increased buffalo density, as well as increased electricity supply were associated with decreased risk of HPAI outbreaks. This illustrates that preventive control measures for HPAI should include a promotion of low-risk agricultural management practices as well as improvement of the infrastructure in village households. Improved HPAI vaccination efforts and coverage should focus on medium poultry density areas and on the pre-monsoon time period.

Peste des petits ruminants (PPR) in Ethiopia: analysis of a national serological survey

Background

Peste des petits ruminants (PPR) is a contagious viral disease of small ruminants in Africa and Asia. In 1999, probably the largest survey on PPR ever conducted in Africa was initiated in Ethiopia where 13 651 serum samples from 7 out of the 11 regions were collected and analyzed by competitive enzyme-linked immunosorbent assay (cELISA). The objective of this paper is to present the results of this survey and discuss their practical implications for PPR-endemic regions.

Methods

We explored the spatial distribution of PPR in Ethiopia and we investigated risk factors for positive serological status. Intracluster correlation coefficients (ρ), were calculated for 43 wereda (administrative units).

Results

Seroprevalence was very heterogeneous across regions and even more across wereda, with prevalence estimates ranging from 0% to 52.5%. Two groups of weredas could be distinguished on the basis of the estimated ρ: a group with very low ρ (ρ < 0.12) and a group with very high ρ (ρ > 0.37).

Conclusion

The results indicate that PPRV circulation has been very heterogeneous, the values for the ρ may reflect the endemic or epidemic presence of the virus or the various degrees of mixing of animals in the different areas and production systems. Age appears as a risk factor for seropositive status, the linear effect seeming to confirm in the field that PPRV is highly immunogenic. Our estimates of intracluster correlation may prove useful in the design of serosurveys in other countries where PPR is of importance.

Poxviral disease in red squirrels Sciurus vulgaris in the UK: spatial and temporal trends of an emerging threat

The squirrel poxvirus (SQPV) is the probable mediator of apparent competition between the introduced invading gray squirrel (Sciurus carolinensis) and the red squirrel (Sciurus vulgaris) in the UK, and modeling studies have shown that this viral disease has had a significant impact on the decline of the red squirrel in the UK. However, given our limited understanding of the epidemiology of the disease, and more generally the effects of invasive species on parasite ecology, there is a need to investigate the transmission dynamics and the relative pathogenicity of the virus between species. We aimed to increase our knowledge of these processes through an empirical study in which we: (i) used pathological signs and transmission electron microscopy (TEM) to diagnose SQPV disease in red squirrels found dead during scanning surveillance between 1993 and 2005; (ii) detected antibody to SQPV using an enzyme-linked immunosorbent assay (ELISA) in the same animals; and (iii) mapped cases of the disease, and the gray squirrel distribution, using a geographical information system. We analyzed the distribution of cases of SQPV disease according to woodland type, a measure of squirrel density. SQPV disease occurred only in areas of England also inhabited by seropositive gray squirrels, and as the geographical range of gray squirrels expanded, SQPV disease occurred in these new gray squirrel habitats, supporting a role for the gray squirrel as a reservoir host of the virus. There was a delay between the establishment of invading gray squirrels and cases of the disease in red squirrels which implies gray squirrels must reach a threshold number or density before the virus is transmitted to red squirrels. The spatial and temporal trend in SQPV disease outbreaks suggested that SQPV disease will have a significant effect on Scottish populations of red squirrels within 25 years. The even spread of cases of disease across months suggested a direct rather than vector-borne transmission route is more likely. Eight juvenile and sub-adult free-living red squirrels apparently survived exposure to SQPV by mounting an immune response, the first evidence of immunity to SQPV in free-living red squirrels, which possibly suggests a changing host-parasite relationship and that the use of a vaccine may be an effective management tool to protect remnant red squirrel populations.

Use of data mining techniques to investigate disease risk classification as a proxy for compromised biosecurity of cattle herds in Wales

Background

Biosecurity is at the forefront of the fight against infectious diseases in animal populations. Few research studies have attempted to identify and quantify the effectiveness of biosecurity against disease introduction or presence in cattle farms and, when done, they have relied on the collection of on-farm data. Data on environmental, animal movement, demographic/husbandry systems and density disease determinants can be collated without requiring additional specific on-farm data collection activities, since they have already been collected for some other purposes. The aim of this study was to classify cattle herds according to their risk of disease presence as a proxy for compromised biosecurity in the cattle population of Wales in 2004 for risk-based surveillance purposes.

Results

Three data mining methods have been applied: logistic regression, classification trees and factor analysis. Using the cattle holding population in Wales, a holding was considered positive if at least bovine TB or one of the ten most frequently diagnosed infectious or transmissible non-notifiable diseases in England and Wales, according to the Veterinary Investigation Surveillance Report (VIDA) had been diagnosed in 2004. High-risk holdings can be described as open large cattle herds located in high-density cattle areas with frequent movements off to many locations within Wales. Additional risks are associated with the holding being a dairy enterprise and with a large farming area.

Conclusion

This work has demonstrated the potential of mining various livestock-relevant databases to obtain generic criteria for individual cattle herd biosecurity risk classification. Despite the data and analytical constraints the described risk profiles are highly specific and present variable sensitivity depending on the model specifications. Risk profiling of farms provides a tool for designing targeted surveillance activities for endemic or emerging diseases, regardless of the prior amount of information available on biosecurity at farm level. As the delivery of practical evidence-based information and advice is one of the priorities of Defra's new Animal Health and Welfare Strategy (AHWS), data-driven models, derived from existing databases, need to be developed that can then be used to inform activities during outbreaks of endemic diseases and to help design surveillance activities.