A longitudinal model for the spread of bovine tuberculosis

Lead Research Organisation: University of Cambridge
Department Name: Veterinary Medicine

Abstract

Bovine tuberculosis (bTB) is an important disease of cattle and badgers with substantial socio-economic impact in the UK, currently costing the exchequer over £100 million per year in surveillance and compensation and also resulting in costly movement and trade restrictions for farmers. Despite intensive controls, disease incidence is still increasing. Currently herds are monitored for the disease through slaughterhouse surveillance and through regular skin testing. The frequency of routine testing for an individual herd is based on localised incidence of the disease, which acts as a proxy for risk of infection, but does not account for individual herd-level characteristics or cattle movements. Recent bTB research has focussed on examining potential underlying causes for this, including environmental contamination (e.g. re-infection from local wildlife reservoirs), insensitivity to the surveillance test and the impact of large-scale cattle movements. It is the purpose of this proposal to extend our recent work identifying markers for the persistence of infection in individual herds into a dynamic longitudinal framework in order to quantify the mechanisms of transmission in the GB national herd and to test the utility of our results as an aid to risk-based surveillance. The dynamics of transmission of bTB infection can be represented by a model with transmission driven by chance processes, with an observation process that is governed by an imperfect test procedure (or slaughterhouse identification of visible lesions), leading to partially hidden infection. Herds that contain one or more reactors are classified as breakdowns, which then have movement restrictions and more rigorous testing imposed until the herd tests clear. Testing and cattle movement information is available through several large national datasets. Recent mathematical modelling approaches have been developed using these data and, while these will provide useful information on population-level parameters, they average out some detailed information available at the individual herd level. Also, they were not designed to predict disease recurrence at the individual-herd level. Here we propose to build a dynamic, statistical, individual-herd level model, based on continuous surveillance data, which we will fit to the data using a likelihood-based approach. The main methodological challenge will be to deal with the hidden states (infection) and the movement of animals between the herds. Recent advances in statistical methodology, such as 'data-augmented' and 'reversible-jump' Markov chain Monte Carlo allow the joint distribution of the observed and hidden states to be estimated simultaneously along with key infection related parameters. We will explore an exciting alternative called 'sequential filtering'. The main challenge is that these statistical techniques are computationally intensive, especially given the large scale (approx. 130,000 premises) and long time frame (6+ years) of the datasets. However, advances in computer processing technology, such as architectures for running algorithms in parallel on graphics cards, provide an exciting and cost-effective way to approach this problem. The focus here is on bTB, but these sorts of models and the estimation issues that we will address are relevant to a wide range of infectious disease systems, and the methodology developed in this project would be applicable to a range of disease systems. It is the aim of this project to elicit information about the hidden states of the system from the test observations using robust statistical methodology, in a way that allows us to identify high-risk herds based on the past history of infection, as well as on localised incidence and connectedness to other premises. This information would have a practical use in terms of targeting specific herds with more stringent or more regular testing.

Technical Summary

Currently the incidence of bovine tuberculosis in Great Britain is increasing, despite over £100 million per year being spent in surveillance and compensation. Herds are monitored for the disease through slaughterhouse surveillance and through imperfect routine skin testing. Our proposal is to build a longitudinal statistical model to account for within- and between-herd transmission of bTB using information from the VetNet, Vebus and Cattle Tracing System data sets. The aim is to understand better the mechanisms of transmission and ultimately to be able to identify herds with a high-risk of harbouring undetected infection, based on the past history of infection as well as localised incidence and connectedness to other premises, and also to characterise areas or premises of high- or low-risk relative to that due to the cattle network. The system is characterised by a hidden dynamic non-linear infection process overlaid with a probabilistic observation mechanism. We intend to explore the development of data-augmented Markov chain Monte Carlo or sequential filtering methods to fit the model. Localised (environmental) transmission will be dealt with using spatial and non-spatial random effects. A surveillance model of this nature would have a practical use in terms of targeting specific herds with more stringent or more regular testing, and if successful could have substantial impacts on slowing the spread of the disease. However, these methods are computationally intensive, given the large scale and long time frame of the datasets. Recent advances in parallel processing technology, both on multi-core CPUs as well as on graphics processing units (e.g. CUDA/OpenCL) provide an exciting and cost-effective way of approaching this problem. Although the focus here is on bTB, the computational problems of fitting models to large-scale data sets in a robust manner are relevant to a wide range of infectious disease systems.

Planned Impact

The main aim of this project is to develop a longitudinal statistical risk-based surveillance model for bovine tuberculosis (bTB) in the UK, which will aid better understanding of the mechanisms of bTB spread in the GB national herd, and could be used to help better optimise targeting of available control resources. Given the huge socio-economic costs of the disease and the increasing incidence, this information may help to increase the efficacy and cost-effectiveness of control, and hopefully might help slow the spread of infection in the national herd. As such it would be of immediate potential benefit to both Defra and Animal Health (AH), and in the longer term to farmers and all those affected by the disease. Animal movements will be modelled explicitly and localised environmental transmission will be modelled through the use of spatio-temporal random effects. While this does not model explicit mechanisms for localised spread, for which there is little or no data available nationally, it will instead highlight areas or herds for which there is a higher or lower risk of infection or persistence than can be explained by cattle movements alone. This in itself would be useful information that would inform further studies in specifically identified areas in order to ascertain the reasons behind these observed differences. The efficacy of the model as a predictive tool will be assessed using both simulated and real data. If these results show potential practical worth, then regular dialogues with Defra and AH will help to hone the optimal way in which the model could be used in practice. A user-friendly interface will be developed (probably web-based, using open-source software) in conjunction with Defra, VLA and AH. A requirement would be that new information could be entered into the model as data becomes available, with the results output in a manner that is straightforward to implement in the field. This technology and the science behind it would be disseminated at various different workshops such as those run by AH. This work involves development of cutting-edge statistical methodology for fitting complex non-linear dynamic models to large scale data sets. With the impacts of infectious diseases becoming more widely realised and technology for data collection improving rapidly, this research has potentially important consequences for developing and fitting statistical models to a wide range of dangerous infectious disease systems, and importantly could help to add to the suite of real-time modelling tools available for modelling epidemic outbreaks.

Publications

10 25 50

publication icon
Brooks-Pollock E (2015) Eliminating bovine tuberculosis in cattle and badgers: insight from a dynamic model. in Proceedings. Biological sciences

publication icon
Brooks-Pollock E (2013) Age-dependent patterns of bovine tuberculosis in cattle in Veterinary Research

publication icon
Conlan A (2012) Estimating the Hidden Burden of Bovine Tuberculosis in Great Britain in PLoS Computational Biology

publication icon
McKinley T (2014) Simulation-based Bayesian inference for epidemic models in Computational Statistics & Data Analysis

 
Description Bovine tuberculosis (bTB) is an important disease of cattle and badgers with substantial socio-economic impact in the UK, currently costing the exchequer over £100 million per year in surveillance and compensation and resulting in costly movement and trade restrictions for farmers. Despite intensive controls, disease incidence is still increasing. Currently herds are monitored for the disease through slaughterhouse surveillance and through routine surveillance through skin testing. The frequency of routine testing for an individual herd is based on regional incidence of disease, which acts as a proxy for risk of infection, but does not account for individual herd-level characteristics or cattle movements. Some recent bTB research has focussed on examining potential underlying causes for this, including environmental contamination (e.g. re-infection from local wildlife reservoirs), insensitivity to the surveillance test leading to persistence in cattle herds and the impact of cattle movements. The purpose of this project was to extend recent work identifying markers for the persistence of infection in individual herds into a dynamic longitudinal framework and so to quantify the mechanisms of transmission in the GB national herd, testing the utility of our results as an aid to risk-based surveillance.

The dynamics of transmission of bTB infection can be represented by a model with transmission driven by chance processes, with an imperfect observation process based on the testing used, leading to partially hidden infections (and slaughterhouse identification of visible lesions in routinely slaughtered animals). Herds that contain one or more reactors are classified as breakdowns, which have movement restrictions and more rigorous testing imposed until the herd tests 'clear'. Testing and cattle movement information is available in national datasets. Recent mathematical modelling approaches based on these data provide some useful information on population-level parameters, but they average out some of the detailed information available at the individual herd level.

We built a dynamic, statistical, individual-animal level model, fitted to the surveillance data using a likelihood-based approach. The main methodological challenge was to deal with the hidden infection states and the movement of animals between the herds. Recent advances in statistical methodology, such as "data-augmented" and "reversible-jump" Markov chain Monte Carlo allow the joint distribution of the observed and hidden states to be estimated simultaneously along with key infection-related parameters. The main challenge is that these statistical techniques are very computationally intensive, especially given the large scale (approx. 130,000 premises) and long time frame (6+ years) of the datasets. However, recent advances in computer processing technology, such as architectures for parallelising algorithms, provide exciting and cost-effective potential approaches for this problem. The focus of this project is on bTB, but these models and the related parameter estimation issues are relevant to a broad range of infectious disease systems, particularly as mathematical model complexity and the size of available datasets continue to increase.

The approach works by simulating epidemics across the system in a controlled manner, and then linking these simulated epidemics with the observed data through a series of probabilistic functions. The likelihood of any given simulation being consistent with the observed data can then be evaluated and weighted accordingly. The algorithm then proceeds by repeating this process, such that the uncertainty due to the hidden infection process can be averaged over. This presents us with a means of producing uncertainty estimates for the parameters of the system that reflects the degree of missing information present in the data.

We constructed simulations at the individual-animal level, since information from the Cattle Tracing System (CTS) can be used to build a detailed network of movements between herds. This was then linked with testing, slaughterhouse surveillance, and breakdown data from the VetNet surveillance system, meaning that the only process we are simulating is the infection process, thus reducing unnecessary uncertainty as much as possible by utilising as much information as we can from the data.

This approach has several useful properties. Since we simulate the hidden infection states, we can pose different questions regarding the mapping between the hidden and observed processes. For example:

1. We can simulate how many reactors we might expect at any given test, which can then be compared to the observed data as a goodness-of-fit measure.
2. We can simulate how many hidden infections remain in a herd after any given test (as well as associated measures like the probability of clearing infection).
3. We can explore in more detail the likely sources of infection (e.g. for each infection that occurs, we can look at the relative risk that the infection was seeded from another cow, or through environmental contamination). In addition we can look at sources of re-infection into clean herds (i.e. in clean herds infected animals are either moved into the herd-which we can map, or otherwise it must be from the environment).

Since all of the above can be linked to explicit spatial locations and times, we can average over space and time in order to look at spatio-temporal heterogeneity in these measures (as well as assessing the uncertainty in the estimates due to the stochastic nature of the hidden processes). The major challenge lies in developing ways in which we can produce simulations of the hidden processes in an efficient manner.

In this project we have developed a framework which allows us to fit various models to this complex data set. We have utilised parallel processing, and developed novel ways to parallelise parts of the algorithm that are non-trivial to compute. This approach is scalable, but improvements in efficiency are linked to the amount of processing power available. Hence, these improvements notwithstanding, the code still takes a long time to run on available servers. As such the final model runs have not yet completed and the work is still ongoing. However, a publication is being written and preliminary work looks promising to give us useful insights into bTB transmission in GB. An important output from this project is the generation of a collated and clean data set that links individual animal-level movements to the testing and surveillance data. Furthermore, we have developed a set of modular code functions that enable us to manipulate and alter these data sets, which we can use as foundations for exploring alternative methodologies for model fitting and prediction on these complex data. As such, future work aims to utilise this framework to explore and compare different models, and also as a basis for developing other inference methods that may be more efficient and scalable.
Exploitation Route The framework we have developed in this project shows great potential as a means for monitoring and understanding the spread of a disease such as bovine tuberculosis. Many of the initial reasons for developing a framework in this way have proved well founded, however the key hurdle left to overcome is to be able to obtain an MCMC chain with sufficiently robust convergence diagnostics to give us a reasonable degree of confidence in the estimates obtained from the model. Once this is done then the framework could be very useful to Defra.
Sectors Agriculture, Food and Drink

URL http://randd.defra.gov.uk
 
Description CattleBCG Vaccine Transmission Working Group
Geographic Reach National 
Policy Influence Type Participation in a advisory committee
Impact Briefly describe the impacts of this change in policy or practice. This should include (if applicable) the reach and significance of the impact, such as quantitative information regarding the benefits (increases in survival, quality of life, decreases in incidence, improvements in clinical service delivery, economic impacts etc.) * On the basis of our research on the potential use of cattle vaccinatino we were written into the specification of an open competition for a consortium to design field trials for the evaluation of BCG vaccination in cattle in 2014. As members of the winning consortium bid led by the Veterinary consultancy firm Triveritas we carried out modelling studies to inform the design and size of these proposed field trials. On the basis of our advice, government policy shifted away from field evaluation of BCG and towards development and validation of a replacement DIVA (Diagnose Infected from Vaccinated Animals) test on which the success of field trials would hinge. In a written statement to the House of Commons by Elizabeth Truss, The Secretary of State for Environment, Food and Rural Affairs. 18th December 2014 said: "An independent report on the design of field trials of cattle vaccine and a test to detect infected cattle among vaccinated cattle (DIVA) shows that before cattle vaccination field trials can be contemplated, we need to develop a better DIVA test. This research is likely to take a further two years." As part of the following policy review I was invited to join a working group (CattleBCG Vaccine Transmission Working Group) to design experimental studies as an alternative to field evaluation of vaccination. Our new designs suggested that a three-year study with 200 animals could be sufficient to demonstrate both the efficacy and mode of action of BCG. Although Defra have chosen not to go ahead with such trials, we have recently been awarded funding from the Bill and Melinda Gates Foundation, working with government agencies in India and Ethiopia to carry out these experiments to both increase our understanding of the action of BCG vaccination in cattle and to accelerate the use of vaccination in developing countries endemically infected with bovine TB.
 
Description Input to development of bovine TB control in Great Britain
Geographic Reach National 
Policy Influence Type Membership of a guideline committee
Impact I am the independent scientist member of the Bovine TB Eradication Advisory Committee in Defra, which advises Defra Ministers on matters relating to bovine TB control in England. Working with Defra policy teams, it was responsible for the 25 year Government Strategy for the Eradication of Bovine TB, published 4 years ago
URL https://www.gov.uk/government/publications/a-strategy-for-achieving-officially-bovine-tuberculosis-f...
 
Description Prof James Wood membership to The Science Advisory Council (SAC)
Geographic Reach National 
Policy Influence Type Membership of a guideline committee
Impact The Science Advisory Council (SAC) provides independent advice on science policy and strategy to the Department for Environment Food and Rural Affairs. Sub groups are set-up to react to sudden global threats and other challenges (one of which is exotic diseases).
URL https://www.gov.uk/government/organisations/science-advisory-council/about/our-governance
 
Description Round table TB event hosted by Michael Gove
Geographic Reach National 
Policy Influence Type Participation in a national consultation
Impact Professor James Wood inputted into a round table event chaired by The Secretary of State for Environment, Food and Rural Affairs (Micheal Gove) with invited experts in the field of TB, in order to discuss plans for eradicating TB from England by 2038. Since then a small working group has been set-up to develop and drive the necessary strategies.
URL https://www.gov.uk/government/news/bovine-tb-strategy-review
 
Description A Study to identify factors associated with the detection of TB breakdowns via abattoir surveillance in GB
Amount £323,597 (GBP)
Funding ID SE3133 
Organisation Department For Environment, Food And Rural Affairs (DEFRA) 
Sector Public
Country United Kingdom
Start 10/2013 
End 03/2016
 
Description ETHICOBOTS (Ethiopia Control of Bovine Tuberculosis Strategies)
Amount £2,903,108 (GBP)
Funding ID BB/L018977/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 09/2014 
End 08/2019
 
Description ETHICOBOTS 2 - One Health Research for Impact
Amount £449,990 (GBP)
Funding ID BB/S013806/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 02/2019 
End 01/2021
 
Description Exploring the richness of Mycobacterium bovis strain diversity to decipher the epidemiology of bovine tuberculosis ecology
Amount £530,175 (GBP)
Funding ID BB/N00468X/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 07/2016 
End 08/2019
 
Description The Design of a Field Trial to test and validate the performance of cattle BCG vaccine
Amount £51,002 (GBP)
Funding ID SE3827 
Organisation Department For Environment, Food And Rural Affairs (DEFRA) 
Sector Public
Country United Kingdom
Start 03/2014 
End 10/2014
 
Description Bovine TB Science Conference 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact I was invited to give a presentation and participate in a panel debate on the science of bovine Tuberculosis organised by the NFU for it's members. Other speakers included leading experts in the field, the Chief Scientific Advisor to Defra, the Chief Veterinary officer for Wales and senior civil servants responsible for bTB policy. My presentation highlighting the importance of cattle in the transmission of bovine Tuberculosis sparked robust discussion with farmers more likely to focus on the, still poorly quantified and undrstood, role of badgers.
Year(s) Of Engagement Activity 2014
 
Description Bovine TB policy engagement 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact Regular meetings, as part of ongoing work, with the Defra bovine TB policy team, using research data to inform policy development. The discussions are broad and do not relate to specific individual findings.
Year(s) Of Engagement Activity 2015,2016,2017
 
Description TB Meeting organised by Eurobadger 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Third sector organisations
Results and Impact I was invited to give a presentation on the "Known Knowns and Known Unknowns of transmission of bovine TB in cattle." to a group of wildlife conservation activists at an Event organised by the Eurobadger group. My presentation emphasised the necessity of both cattle and badger transmission to explain the pattern of expansion of bovine Tuberculosis and stimulated lively discussion on the interpretation of historical data from the Randomised Badger Culling Trials (RBCT).
Year(s) Of Engagement Activity 2017
 
Description TB or not TB? 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact A presentation was made of the overall funded study at work as an invited talk at the GEOMED 2017 International Conference on Spatial Statistics, Spatial Epidemiology and Spatial Aspects of Public Health. This international conference was held in Porto, Portugal in September 2017.
Year(s) Of Engagement Activity 2017