Development of risk-based trading farm scoring system to assist with the control of bovine tuberculosis in cattle in England and Wales

Is badger culling associated with risk compensation behaviour among cattle farmers?

BACKGROUND

Risk compensation theory suggests that behaviours are modified in response to interventions that remove risks by substituting them with other risky behaviours to maintain a 'risk equilibrium'. Alternatively, risk reduction interventions may result in spill-over behaviours that seek to minimise risks further. This paper assessed evidence for these behavioural risk responses among farmers in response to badger culling that seeks to remove the risk of bovine tuberculosis in cattle.

METHODS

Data from the UK's randomised badger culling trial were re-analysed, comparing farmers' cattle movement practices in proactive and reactive culling areas and control areas. Analysis compared cattle movements during and after the trial using zero-inflated negative binomial regression.

RESULTS

The analysis found no strong evidence of risk compensation behaviours among farmers who experienced proactive culling. However, strong evidence for a reduction in cattle movements in reactive culling areas was found. The results indicate high levels of inertia within farming systems in relation to cattle purchasing.

LIMITATIONS

Data do not account for the risk of cattle purchases and reflect previous policy regimens. Evidence from recent badger culling interventions should be analysed.

CONCLUSION

Proactive badger culling was not associated with risk compensation behaviours, while reactive badger culling was associated with decreased risk taking among farmers.

Presence of antibodies against Mycobacterium avium subspecies paratuberculosis in Brazilian high-producing dairy herds

This study aimed to determine the presence of antibodies against Mycobacterium avium subspecies paratuberculosis (MAP) in high-producing dairy cows, the presence of the pathogen in the feces, and the risk factors associated with the disease. Blood and fecal samples were collected from 708 dairy cows over 2 years from 54 herds located in five municipalities of Paraná, Brazil. The serum samples were evaluated for the presence of antibodies against MAP using enzyme-linked immunosorbent assay (ELISA). Fecal samples from 100 cows (69 seropositive and 31 seronegative) were assessed using real-time PCR (qPCR) for IS900 of MAP. The herd prevalence of antibodies against MAP was 61.1% (33/54; 95% CI 46.88-74.08), ranging from 12.5 to 80% across the municipalities, and the prevalence in the animals was 9.8% (69/708; 95% CI 7.77-12.15); it ranged from 0 to 87.5% per herd. Only one of the 69 (1.45%) fecal samples from the seropositive cows was positive for the qPCR. The factors associated with the occurrence of paratuberculosis in herds were the use of compost barn system and the type of bed, whereas only the type of bed was associated with the infection of cows. The only risk factor (OR = 2.45; 95% CI 1.03-5.85) associated with the occurrence of paratuberculosis was the introduction of animals purchased from other dairy farms. The prevalence of active infection was low; however, our results demonstrate the presence of MAP in high-producing dairy herds in Paraná state, Brazil.

Machine learning classification methods informing the management of inconclusive reactors at bovine tuberculosis surveillance tests in England

Bovine tuberculosis (bTB) remains one of the most complex, challenging, and costly animal health problems in England. Identifying and promptly removing all infected cattle from affected herds is key to its eradication strategy; the imperfect sensitivity of the diagnostic testing regime remaining a serious obstacle. The main diagnostic test for bTB in cattle in England, the Single Intradermal Comparative Cervical Tuberculin Test (SICCT: also known as the skin test), can produce inconclusive results below the reactor threshold. The immediate isolation of inconclusive reactor (IR) animals followed by a 60-day retest may not prevent lateral spread within the herd (if it is substandard, allowing transmission) or transmission to wildlife. Over half of IR-only herds that went on to have a positive skin test result (a bTB herd 'incident') in 2020, had it triggered by at least one IR not clearing their 60-day retest, instead of by another test within the previous 15 months. Machine learning classification algorithms (classification tree analysis and random forest), applied to England's 2012-2020 IR-only surveillance herd tests, identified at-risk tests for an incident at the IRs' 60-day retest. In this period, 4 739 out of 22 946 (21 %) IR-only surveillance tests disclosing 6 296 out of 42 685 total IRs, had an incident at retest (2 716 IRs became reactors and 3 580 IRs became two-time IRs). Both models showed an AUC above 80 % in the 2012-2019 dataset. Classification tree analysis was preferred due to its easy-to-interpret outputs, 70 % sensitivity, and 93 % specificity in the 20 % of 2019-2020 testing dataset. The paper aimed to identify IR-only surveillance tests at-risk of an incident at the 60-day retest to target them with appropriate measures to mitigate the IRs' risk. Sixteen percent (341 out of 2 177) of IR-only herd tests were identified as high-risk in the 2020 dataset, with 265 (78 %) of these having at least one reactor or IR at retest. Severe-level reinterpretation of the high-risk IR-only disclosing tests identified in this dataset would turn 68 out of the 590 (12 %) IRs into reactors, generating 23 incidents, the majority (19 or 83 %) part of the 265 incidents that would have been declared at the retest. Classification tree analysis used to identify IR-only high-risk tests in herds eligible for severe interpretation would enhance the sensitivity of the test-and-slaughter regime, cornerstone of the bTB eradication programme in England, further mitigating the risk of disease spread posed by IRs.

Survey of emergency response plans for managing the movement of cattle during a foot-and-mouth disease outbreak in North America

OBJECTIVE

To collect information from US state animal health officials (SAHOs) and beef feedlot managers and veterinarians regarding emergency response plans for movement of cattle in the event of a foot-and-mouth disease (FMD) outbreak in North America.

SAMPLE

36 SAHOs, 26 feedlot veterinarians, and 7 feedlot managers.

PROCEDURES

3 versions of an electronic questionnaire were created and distributed to SAHOs and US feedlot veterinarians and managers to gather information about planned or expected responses to an FMD outbreak that originated at 1 of 3 geographic locations (Mexico or Canada, a bordering state, or a nonbordering state). Descriptive data were reported.

RESULTS

All respondents recognized that the risk of FMD transmission to livestock in their area or care increased as the outbreak got closer in proximity to their location. Most SAHOs indicated that they would immediately close their state's borders to livestock movement at the beginning of an FMD outbreak, particularly if the disease was identified in a bordering state. During an extended FMD outbreak, 29 of 36 (80.6%) SAHOs reported they would resume interstate movement of cattle under some conditions, including enhanced permitting, whereas feedlot veterinarians and managers commonly reported they would be willing to receive cattle from states where no FMD-infected animals were identified, regardless of permit requirements.

CONCLUSIONS AND CLINICAL RELEVANCE

Information gained from this survey can be used to inform disease modeling and preparedness efforts to facilitate business continuity of US beef feedlots in the event of an FMD outbreak in North America.

A comparison of the value of two machine learning predictive models to support bovine tuberculosis disease control in England

Nearly a decade into Defra's current eradication strategy, bovine tuberculosis (bTB) remains a serious animal health problem in England, with c.30,000 cattle slaughtered annually in the fight against this insidious disease. There is an urgent need to improve our understanding of bTB risk in order to enhance the current disease control policy. Machine learning approaches applied to big datasets offer a potential way to do this. Regularized regression and random forest machine learning methodologies were implemented using 2016 herd-level data to generate the best possible predictive models for a bTB incident in England and its three surveillance risk areas (High-risk area [HRA], Edge area [EA] and Low-risk area [LRA]). Their predictive performance was compared and the best models in each area were used to characterize herds according to risk. While all models provided excellent discrimination, random forest models achieved the highest balanced accuracy (i.e. average of sensitivity and specificity) in England, HRA and LRA, whereas the regularized regression LASSO model did so in the EA. The time since the last confirmed incident was resolved was the only variable in the top-ten ranking in all areas according to both types of models, which highlights the importance of bTB history as a predictor of a new incident. Risk categorisation based on Receiver Operating Characteristic (ROC) analysis was carried out using the best predictive models in each area setting a 99 % threshold value for sensitivity and specificity (97 % in the LRA). Thirteen percent of herds in the whole of England as well as in its HRA, 14 % in its EA and 31 % in its LRA were classified as high-risk. These could be selected for the deployment of additional disease control measures at national or area level. In this way, low-risk herds within the area considered would not be penalised unnecessarily by blanket control measures and limited resources be used more efficiently. The methodology presented in this paper demonstrates a way to accurately identify high-risk farms to inform a targeted disease control and prevention strategy in England that supplements existing population strategies.

Northern Ireland farm-level management factors for prolonged bovine tuberculosis herd breakdowns

This study determined farm management factors associated with long-duration bovine tuberculosis (bTB) breakdowns disclosed in the period 23 May 2016 to 21 May 2018; a study area not previously subject to investigation in Northern Ireland. A farm-level epidemiological investigation (n = 2935) was completed when one or more Single Intradermal Comparative Cervical Test (SICCT) reactors or when one or more confirmed (positive histological and/or bacteriological result) lesion at routine slaughter were disclosed. A case-control study design was used to construct an explanatory set of management factors associated with long-duration bTB herd breakdowns; with a case (n = 191) defined as an investigation into a breakdown of 365 days or longer. Purchase of infected animal(s) had the strongest association as the most likely source of infection for long-duration bTB herd breakdowns followed by badgers and then cattle-to-cattle contiguous herd spread. However, 73.5% (95% CI 61.1-85.9%) of the herd type contributing to the purchase of infection source were defined as beef fattening herds. This result demonstrates two subpopulations of prolonged bTB breakdowns, the first being beef fattening herds with main source continuous purchase of infected animals and a second group of primary production herds (dairy, beef cows and mixed) with risk from multiple sources.

Characterization of potential superspreader farms for bovine tuberculosis: A review

Background

Variation in host attributes that influence their contact rates and infectiousness can lead some individuals to make disproportionate contributions to the spread of infections. Understanding the roles of such ‘superspreaders’ can be crucial in deciding where to direct disease surveillance and controls to greatest effect. In the epidemiology of bovine tuberculosis (bTB) in Great Britain, it has been suggested that a minority of cattle farms or herds might make disproportionate contributions to the spread of Mycobacterium bovis, and hence might be considered ‘superspreader farms’.

Objectives and Methods

We review the literature to identify the characteristics of farms that have the potential to contribute to exceptional values in the three main components of the farm reproductive number ‐ R_f: contact rate, infectiousness and duration of infectiousness, and therefore might characterize potential superspreader farms for bovine tuberculosis in Great Britain.

Results

Farms exhibit marked heterogeneity in contact rates arising from between‐farm trading of cattle. A minority of farms act as trading hubs that greatly augment connections within cattle trading networks. Herd infectiousness might be increased by high within‐herd transmission or the presence of supershedding individuals, or infectiousness might be prolonged due to undetected infections or by repeated local transmission, via wildlife or fomites.

Conclusions

Targeting control methods on putative superspreader farms might yield disproportionate benefits in controlling endemic bovine tuberculosis in Great Britain. However, real‐time identification of any such farms, and integration of controls with industry practices, present analytical, operational and policy challenges.

Effects of trading networks on the risk of bovine tuberculosis incidents on cattle farms in Great Britain

Trading animals between farms and via markets can provide a conduit for spread of infections. By studying trading networks, we might better understand the dynamics of livestock diseases. We constructed ingoing contact chains of cattle farms in Great Britain that were linked by trading, to elucidate potential pathways for the transmission of infection and to evaluate their effect on the risk of a farm experiencing a bovine tuberculosis (bTB) incident. Our findings are consistent with variation in bTB risk associated with region, herd size, disease risk area and history of previous bTB incidents on the root farm and nearby farms. However, we also identified effects of both direct and indirect trading patterns, such that connections to more farms in the England High-Risk Area up to three movements away from the root farm increased the odds of a bTB incident, while connections with more farms in the England Low-Risk Area up to eight movements away decreased the odds. Relative to other risk factors for bTB, trading behaviours are arguably more amenable to change, and consideration of risks associated with indirect trading, as well direct trading, might therefore offer an additional approach to bTB control in Great Britain.

Exploring the Risk Posed by Animals with an Inconclusive Reaction to the Bovine Tuberculosis Skin Test in England and Wales

The single intradermal comparative cervical tuberculin (SICCT) test is the primary test for ante-mortem diagnosis of bovine tuberculosis (TB) in England and Wales. When an animal is first classified as an inconclusive reactor (IR) using this test, it is not subject to compulsory slaughter, but it must be isolated from the rest of the herd. To understand the risk posed by these animals, a case-control study was conducted to measure the association between IR status of animals and the odds of them becoming a reactor to the SICCT at a subsequent test. The study included all animals from herds in which only IR animals were found at the first whole herd test in 2012 and used data from subsequent tests up until the end of 2016. Separate mixed-effects logistic regression models were developed to examine the relationship between IR status and subsequent reactor status for each risk area of England and for Wales, adjusting for other explanatory variables. The odds of an animal becoming a subsequent reactor during the study period were greater for IR animals than for negative animals in the high-risk area (odds ratio (OR): 6.85 (5.98-7.86)) and edge area (OR: 8.79 (5.92-13.04)) of England and in Wales (OR: 6.87 (5.75-8.22)). In the low-risk area of England, the odds were 23 times greater, although the confidence interval around this estimate was larger due to the smaller sample size (11-48, p < 0.001). These findings support the need to explore differential controls for IR animals to reduce the spread of TB, and they highlight the importance of area-specific policies.

Decision tree machine learning applied to bovine tuberculosis risk factors to aid disease control decision making

Identifying and understanding the risk factors for endemic bovine tuberculosis (TB) in cattle herds is critical for the control of this disease. Exploratory machine learning techniques can uncover complex non-linear relationships and interactions within disease causation webs, and enhance our knowledge of TB risk factors and how they are interrelated. Classification tree analysis was used to reveal associations between predictors of TB in England and each of the three surveillance risk areas (High Risk, Edge, and Low Risk) in 2016, identifying the highest risk herds. The main classifying predictor for farms in England overall related to the TB prevalence in the 100 nearest cattle herds. In the High Risk and Edge areas it was the number of slaughterhouse destinations and in the Low Risk area it was the number of cattle tested in surveillance tests. How long ago the last confirmed incident was resolved was the most frequent classifier in trees; if within two years, leading to the highest risk group of herds in the High Risk and Low Risk areas. At least two different slaughterhouse destinations led to the highest risk group of herds in England, whereas in the Edge area it was a combination of no contiguous low-risk neighbours (i.e. in a 1 km radius) and a minimum proportion of 6-23 month-old cattle in November. A threshold value of prevalence in 100 nearest neighbours increased the risk in all areas, although the value was specific to each area. Having low-risk contiguous neighbours reduced the risk in the Edge and High Risk areas, whereas high-risk ones increased the risk in England overall and in the Edge area specifically. The best classification tree models informed multivariable binomial logistic regression models in each area, adding statistical inference outputs. These two approaches showed similar predictive performance although there were some disparities regarding what constituted high-risk predictors. Decision tree machine learning approaches can identify risk factors from webs of causation: information which may then be used to inform decision making for disease control purposes.

Negotiated Management Strategies for Bovine Tuberculosis: Enhancing Risk Mitigation in Michigan and the UK

Bovine tuberculosis (bTB) is an epidemiologically, politically, and socially complex disease. Across multiple international contexts, policy makers have struggled to balance the competing demands of wildlife and agricultural interests in their efforts to create workable and effective disease management strategies. This paper draws comparative lessons between the cases of Michigan in the USA and the UK to exemplify some of the challenges of developing an effective strategy for the long-term control of endemic disease, particularly reflecting on efforts to “responsibilise” cattle producers and engage them in proactive activities to mitigate transmission risks on their own farms. Using qualitative data derived from 22 stakeholder interviews, it is argued that the management of bTB in Michigan has important lessons for the UK on the role of human dimensions in influencing the direction of disease control. The management of endemic bTB relies on the actions of individuals to minimise risk and, in contrast to the predominantly voluntary approach pursued in the UK, Michigan has shifted the emphasis towards obtaining producer support for wildlife risk mitigation and biosecurity via a mix of regulatory, fiscal, and social interventions. Whilst the scale of the bTB challenge differs between these two contexts, analysis of the different ideological bases for selecting management approaches offers interesting insights on the role of negotiated outcomes in attempts to adaptively manage a disease that is characterised by complexity and uncertainty.

Mapping Disease Data: A Usability Test of an Internet-Based System of Disease Status Disclosure

Disease maps are important tools in the management of disease. By communicating risk, disease maps can help raise awareness of disease and encourage farmers and veterinarians to employ best practice to eliminate the spread of disease. However, despite the importance of disease maps in communicating risk and the existence of various online disease maps, there are few studies that explicitly examine their usability. Where disease maps are complicated to use, it seems that they are unlikely to be used effectively. The paper outlines an attempt to create an open access, online, searchable map of incidents of bovine tuberculosis in England and Wales, and analyzes its usability among veterinarians. The paper describes the process of creating the map before describing the results of a series of usability trials. Results show the map to score highly on different measures of usability. However, the trials also revealed a number of social and technical limitations and challenges facing the use of online disease maps, including reputational dangers, role confusion, data accuracy, and data representation. The paper considers the challenges facing disease maps and their potential role in designing new methodologies to evaluate the effectiveness of disease prevention initiatives.

Bovine tuberculosis in Northern Ireland: quantification of the population disease-level effect from cattle leaving herds detected as a source of infection

Determination of the proportion of bovine tuberculosis (bTB) breakdowns attributed to a herd purchasing infected animals has not been previously quantified using data from the Animal and Public Health Information System (APHIS) database in Northern Ireland. We used a case-control study design to account for the infection process occurring in the disclosing bTB breakdown herds. Cases (N = 6926) were cattle moving to a future confirmed bTB breakdown where they would disclose as a confirmed bTB reactor or a Lesion at Routine Slaughter (LRS). Controls (N = 303 499) were cattle moving to a future confirmed bTB breakdown where they did not become a bTB reactor or LRS. Our study showed that the cattle leaving herds which disclosed bTB within 450 days had an increased odds of becoming a confirmed bTB reactor or LRS compared with the cattle which left herds that remained free for 450 days (odds ratio (OR) = 2·09: 95% CI 1·96-2·22). Of the 12 060 confirmed bTB breakdowns included in our study (2007-2015 inclusive), 31% (95% CI 29·8-31·5) contained a confirmed bTB reactor(s) or LRS(s) at the disclosing test which entered the herd within the previous 450 days. After controlling for the infection process occurring in the disclosing bTB breakdown herd, our study showed that 6·4% (95% CI 5·9-6·8) of bTB breakdowns in Northern Ireland were directly attributable to the movement of infected animals.

Developing a risk-based trading scheme for cattle in England: farmer perspectives on managing trading risk for bovine tuberculosis

This paper examines farmer attitudes towards the development of a voluntary risk-based trading scheme for cattle in England as a risk mitigation measure for bovine tuberculosis (bTB). The research reported here was commissioned to gather evidence on the type of scheme that would have a good chance of success in improving the information farmers receive about the bTB risk of cattle they buy. Telephone interviews were conducted with a stratified random sample of 203 cattle farmers in England, splitting the interviews equally between respondents in the high-risk area and low-risk area for bTB. Supplementary interviews and focus groups with farmers were also carried out across the risk areas. Results suggest a greater enthusiasm for a risk-based trading scheme in low-risk areas compared with high-risk areas and among members of breed societies and cattle health schemes. Third-party certification of herds by private vets or the Animal and Plant Health Agency were regarded as the most credible source, with farmer self-certification being favoured by sellers, but being regarded as least credible by buyers. Understanding farmers' attitudes towards voluntary risk-based trading is important to gauge likely uptake, understand preferences for information provision and to assist in monitoring, evaluating and refining the scheme once established.

Assessing the impact of tailored biosecurity advice on farmer behaviour and pathogen presence in beef herds in England and Wales

The term 'biosecurity' encompasses many measures farmers can take to reduce the risk of pathogen incursion or spread. As the best strategy will vary between settings, veterinarians play an important role in assessing risk and providing advice, but effectiveness requires farmer acceptance and implementation. The aim of this study was to assess the effectiveness of specifically-tailored biosecurity advice packages in reducing endemic pathogen presence on UK beef suckler farms. One hundred and sixteen farms recruited by 10 veterinary practices were followed for three years. Farms were randomly allocated to intervention (receiving specifically-tailored advice, with veterinarians and farmers collaborating to develop an improved biosecurity strategy) or control (receiving general advice) groups. A spreadsheet-based tool was used annually to attribute a score to each farm reflecting risk of entry or spread of bovine viral diarrhoea virus (BVDV), bovine herpesvirus-1 (BHV1), Mycobacterium avium subsp. paratuberculosis (MAP), Leptospira interrogans serovar hardjo (L. hardjo) and Mycobacterium bovis (M. bovis). Objectives of these analyses were to identify evidence of reduction in risk behaviours during the study, as well as evidence of reductions in pathogen presence, as indications of effectiveness. Risk behaviours and pathogen prevalences were examined across study years, and on intervention compared with control farms, using descriptive statistics and multilevel regression. There were significant reductions in risk scores for all five pathogens, regardless of intervention status, in every study year compared with the outset. Animals on intervention farms were significantly less likely than those on control farms to be seropositive for BVDV in years 2 and 3 and for L. hardjo in year 3 of the study. Variations by study year in animal-level odds of seropositivity to BHV1 or MAP were not associated with farm intervention status. All farms had significantly reduced odds of BHV1 seropositivity in year 2 than at the outset. Variations in farm-level MAP seropositivity were not associated with intervention status. There were increased odds of M. bovis on intervention farms compared with control farms at the end of the study. Results suggest a structured annual risk assessment process, conducted as a collaboration between veterinarian and farmer, is valuable in encouraging improved biosecurity practices. There were some indications, but not conclusive evidence, that tailored biosecurity advice packages have potential to reduce pathogen presence. These findings will inform development of a collaborative approach to biosecurity between veterinarians and farmers, including adoption of cost-effective strategies effective across pathogens.