Defra - Department for Environment, Food and Rural Affairs.

Science Search

Science and Research Projects

Return to Science Search homepage   Return to Search List

The problem TB herd- characterisation, prediction and resolution. - SE3230

Since the foot and mouth epidemic of 2001 both the distribution and intensity of bovine tuberculosis in the national herd of Great Britain have increased, such that until recently the national trend showed a yearly increase of 18% in confirmed incidents. The current control programme is designed to detect infection (by routine skin testing, slaughterhouse inspection and movement tracing), to remove infection as thoroughly as possible (by short interval testing) and to prevent spread to other herds (by pre-movement testing and movement restrictions).
By its nature, much of the programme is reactive. Parish testing intervals, for example, tend to be modified only after a changed risk of infection is detected, usually several months after the risk has actually increased. Pre-movement testing is a welcome recent step in preventing new cases occurring in areas of previous high or low risk, and the present proposal would yield needed methods for estimating its effectiveness.
However, this programme, apart from modifying parish testing intervals and allowing discretion to treat some ‘high risk’ categories of herd more severely (such as dealer herds), treats all herds in a similar fashion. Yet some herds are known to be ‘problem’ herds, i.e. experience has shown that these herds are more likely to have new incidents (sometimes repeatedly), tend to have more affected animals when they do break down, repeated short interval testing may take a long time to clear infection, or total or partial herd slaughter is deemed necessary to eliminate infection... and so on, depending on the definition of 'problem herd' adopted.
The current research proposal first seeks to define and characterise the ‘problem herd’, then to predict outcomes of possible control strategies, and finally to select appropriate actions. For example, we can describe herds according to their disease history and then focus on characteristics of those herds that may predispose to the defined 'problem’. VLA is now in an excellent position to define and characterise different types of ‘problem’ herd, using detailed and relevant datasets. They include VetNet, the British Cattle Movement System database, the TB99 and CCS05 herd questionnaires and VLA’s molecular typing datasets. Both in size of database and in integrating them, VLA has more experience at using these datasets for investigating TB epidemiology than any other institute. We would focus on problem herds in conjunction with the experience of specialists in TB in both Defra and the State Veterinary Service to help focus on the key characteristics for investigation.
Currently available datasets provide useful data on all types of ‘problem’ herd except for the thousands of herds that have positive avian tuberculin responses. These can delay identification of any bovine TB infection that has occurred. There are several aspects to be considered. For example, herds may not be detected as infected at a routine herd test, introducing a dangerous delay; individuals may not be detected during a confirmed breakdown, resulting in undue extension of the incident; or, then again, herds may clear restrictions only to find further infection at the six month check test because infection was not detected at previous tests. This issue will be treated as a separate objective within the study.
By using the BCMS data in particular, we will be able to quantify with some level of precision how dangerous these ‘problem’ herds are in terms of risk of spread to other herds. By using the full range of data we can attempt to quantify the risks they pose to themselves (in terms of further incidents and degree of within-herd spread). We would then be able to quantify how costly these herds are in the context of the whole control programme, utilising cost data recently gathered in other Defra-funded projects. These outputs would inform the best approaches to controlling the bTB risk to, and presented by, these herds.
An important element of this study will be to develop tools for predicting which herds would fit into ‘problem’ categories, and tools for recommending cost-effective control actions (decision trees or in some cases Expert Systems). These could then help the SVS decide on what action to take to control the risk presented. The decision trees could use information accrued from suspected ’problem’ herds, either -
· Information accumulated before and during a TB incident, leading to recommendations for eliminating infection, or
· Information that is available before infection has been detected, leading to recommendations for reducing the risk of the herd becoming infected or spreading infection.
Information of significance might include:
· Herd size, type, and other husbandry details (source: VetNet, CCS2005 and TB99 data);
· Biosecurity details, e.g. efforts to exclude wildlife from buildings, to prevent cattle coming into contact with wildlife, or coming into contact with other herds (CCS2005 and TB99 data).
· Movement history: numbers of animals moved into the herd, and the time they had spent in herds with various levels of bTB prevalence (source: CTS and VetNet data);
· Distribution of the risk of infection with bovine TB in GB (geographical (GIS) analysis of VetNet data);
· Information gaps – and needs for additional investigations – as identified in this process.
The project will develop decision trees that can enable the SVS to be more proactive in the elimination of infection in ‘problem’ herds, and to prevent the emergence of new problem herds.
1. Defining and describing potential ’problem’ herds with characterisation of principal risk factors;
• Milestones 1 and 3 (months 3 and 12) – A preliminary definition of potential problem herds, and epidemiological and economic characterization of unfavourable outcomes
2. Defining and describing the avian reactor ‘problem’ herd;
• Milestones 2 and 4 (months 6 and 15) – Definition of different scenarios where reaction to avian tuberculin in the skin test may interfere with bovine TB control, collection of data and assessing the actual impact
3. Developing decision trees for the SVS so that appropriate control actions can be based on data accumulated in (potential) problem herds during a TB incident, through analysis of cases, actions and outcomes;
• Milestones 5 and 6 (months 15 and 18) – Ability to predict the outcomes of bTB incidents from observations early in the incident and analysis of effects of local prevalence and other epidemiological factors
4. Developing decision trees for the SVS so that appropriate control actions can be based on data on potential problem herds and their surroundings, before a TB incident has occurred, through analysis of cases, actions and outcomes;
• Milestones 7, 8 and 9 (months 18, 21 and 24) – Ability to predict risk of bovine TB from herd attributes, movement patterns and bTB history (in herds and their localities) and from other epidemiological factors; analysis of benefits and costs; contribution of ancillary tests.
5. Disseminating the above findings to Defra, the State Veterinary Service and the broader scientific community
• Milestones 10 and 11 (months 27 and 30) – Testing the decision trees in collaboration with the SVS; final report drafted and issued, peer-reviewed paper drafted.
Project Documents
• FRP - Final Report : Final Report   (1388k)
Time-Scale and Cost
From: 2007

To: 2011

Cost: £459,875
Contractor / Funded Organisations
Veterinary Laboratories Agency
Animal Diseases              
Animal Health              
Plants and Animals              
Fields of Study
Animal Health