title / A systems analysis methodology to elucidate and evaluate
critical control points for E.coli O157:H7 in cattle and sheep ... / DEFRA
project code / OZ0708
Department for Environment, Food and Rural Affairs CSG 15
Research and Development
Final Project Report
(Not to be used for LINK projects)
Two hard copies of this form should be returned to:Research Policy and International Division, Final Reports Unit
DEFRA, Area 301
Cromwell House, Dean Stanley Street, London, SW1P 3JH.
An electronic version should be e-mailed to
Project title / A systems analysis methodology to elucidate and evaluate
critical control points for E.coli O157:H7 in cattle and sheep ...
DEFRA project code / OZ0708
Contractor organisation and location / Silsoe Research Institute
Wrest Park, Silsoe
Bedford, MK45 4HS
Total DEFRA project costs / £ £151,494.00
Project start date / 01/01/00 / Project end date / 30/06/02
Executive summary (maximum 2 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.
CSG 15 (9/01) 3
Projecttitle / A systems analysis methodology to elucidate and evaluate
critical control points for E.coli O157:H7 in cattle and sheep ... / DEFRA
project code / OZ0708
A model-based study was conducted to assess means of controlling the risk of verocytotoxin-producing Escherichia coli O157 (VTEC O157) reaching abattoirs from cattle and sheep. An analysis of British cattle and sheep calves was conducted to define the systems to be considered, and a literature review collected information on the epidemiology of VTEC O157, with emphasis on quantative data suitable for model development.
A simulation model was developed with two main aspects
· a general, deterministic simulation of the interations between animals, feed, enclosures and waste on cattle and sheep farms to study the impact of different factors;
· a stochastic simulation of the epidemiology of VTEC O157 on such farms to study the likelihood of harmful outcomes.
This was used to predict the effect of interventions on the farm on the prevalence and shedding levels of animals leaving the farm for the abattoir.
The model is capable of simulating almost any type of beef or sheep enterprise. Five representative farms were simulated for this study:
· Grass based beef breeder-finisher
· Grass based beef fattener-finisher
· Intensive beef fattener-finisher
· Sheep (grass based breeder-fattener)
· Mixed beef and sheep
Batches of 100 runs of the model were carried out for each intervention so that statistics could be collected. The outcomes were assessed by considering the mean animal prevalence and mean shedding rate in animals for the abattoir and also the complete distribution of results, through the method of (first-order) stochastic dominance. The interventions for each farm were then ranked by their effectiveness in controlling VTEC O157 in animals being sent to the abattoir.
The main conclusions from the study on the basis of the available data and information (using the shorthand that “risk” means the risk of carrying VTEC O157 to the abattoir) were:
1. Sheep have a lower risk than cattle.
2. Mixing cattle and sheep increases the risk in both groups.
3. Merging groups of animals of the same species into larger groups increases the risk substantially.
4. Bringing in cattle with a prevalence of 5% increases the risk substantially.
5. Increasing farm/group size increases the risk, independently of stocking density.
6. Increasing stocking density increases the risk, independently of group size.
7. A very high level of barn hygiene reduces the risk; there is little difference between intermediate and poor levels.
8. Reducing the interval between slurry spreading and grazing from 4 to 2 weeks increases the risk; increasing the interval only produces a small reduction in risk.
9. Background sources of VTEC O157 in the model have some effect on the prevalence and shedding; these could be influenced in practice by the presence of wild animals, etc., carrying the organism.
10. There may be a higher risk among intensively housed cattle, but this is related to assumptions made in the parameterisation of the model and should be tested against independent evidence.
11. The parameters to which the model is most sensitive (and hence those that it is most important to quantify) are the ones related to
· transmission from grass and enclosures to animals
· contact between animals
· pathogen survival on grass, in slurry and in barns
12. The model could be expanded and refined as new data become available. Indeed it is intended to develop further the waste handling aspects within OZ0709. Other areas where additional data could be used to refine the model are:
· Infection of animals by farm strains of VTEC O157 under farm conditions (as opposed to high doses of laboratory cultures) to test and calibrate the dose-response model.
· Quantified monitoring of shedding in cohorts of animals to calibrate the transition probability matrix.
· Long-term monitoring with quantification of ‘simple’ farming systems to provide validation data sets for this and other models.
CSG 15 (9/01) 3
Projecttitle / A systems analysis methodology to elucidate and evaluate
critical control points for E.coli O157:H7 in cattle and sheep ... / DEFRA
project code / OZ0708
Scientific report (maximum 20 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.
CSG 15 (9/01) 3
Projecttitle / A systems analysis methodology to elucidate and evaluate
critical control points for E.coli O157:H7 in cattle and sheep ... / MAFF
project code / OZ0708
Introduction
Food poisoning is a major cause of public concern over food safety. Notifications in England and Wales have risen from just over 14,000 in 1982 to over 83,000 in 1996, and anxiety has been further heightened by well-publicised incidents, such as the outbreak caused by Escherichia coli O157 in Scotland. The Pennington report (Pennington Group, 1997) stated ‘…it is generally accepted that the main reservoir of E. coli O157 exists in cattle, and possibly sheep, farms…’ and recommended further investigation and research, and an education/awareness programme for farm workers. A comprehensive risk analysis is required to ensure that both of these activities are directed towards the most critical parts of the system.
This project aimed to carry out a risk assessment for verocytotoxin-producing Escherichia coli O157 (hereafter VTEC O157) in cattle and sheep from farm to abattoir based on a systems analysis methodology to ensure that the assessment was complete and comprehensive. Using this methodology, a quantitative framework would be constructed within which risks can be calculated and the consequences of actions explored. The framework was to have been based on causal probability networks (CPNs), whose structure would be generated automatically by a farm systems model, with probabilities supplied by an epidemiological model. It became clear that CPNs were not able to represent some of the essential features of the systems, so a simulation approach to the risk modelling was adopted.
The simulation model now contains all three aspects:
· the farm system
· the epidemiology of VTEC O157
· the risk that animals for the abattoir are carrying VTEC O157
This approach has the benefit of integrating all aspects into a single model. Each run of the simulation results in one possible outcome of the system with values for prevalence and distributions of shedding of VTEC O157 throughout the year for each group of animals. By performing batches of runs, distributions of outcomes are obtained, which represent the range of possibilities arising from the assumptions in the model. The impact of changes to the management of the farm or to the underlying assumptions can be examined and the most significant ones identified. The model is designed to be flexible – for example it can be changed from cattle to sheep through a few settings in the input file – and easy to extend or modify if required in future. (It is anticipated that project OZ0709 will update and extend the slurry handling aspects.) The main advantage of the CPN approach is its ability to use inductive reasoning from consequences to causes, for example by entering evidence on observations made on a particular farm to update the internal variables. This is not a major requirement for the present purpose and is outweighed by the disadvantages.
This report summarises all the main aspects of the project: the review of the epidemiology, the systems analysis, the farm model, the epidemiological/risk model and the results from the model, then discusses the results and draws conclusions. Given the limitations of space, these sections have to be brief. The results of the literature review given here focus on the ecology of VTEC O157 and the full review is given Annex 1. The model is described in non-mathematical form; papers that will give more detail are in preparation.
Objectives
The overall objective is to develop a systems analysis/network flow QRA methodology for VTEC O157 in cattle and sheep production, and apply it to a range of systems to identify critical control points and make recommendations for research priorities.
01. Develop a systems analysis methodology for sheep and cattle production systems. Document the methodology.
02. Develop a model framework of the life cycle of VTEC O157 in cattle and sheep and identify and extract the data required to quantify infection and contamination risks (qualitative and quantitative hazard identification). Document the methodology.
03. Develop a general probability network model by quantifying the probabilities associated with those hazards and quantifying them for the general case. Apply the model to identify critical control points and make recommendations.
Review of epidemiology
A literature review of the epidemiology of VTEC O157 in cattle and sheep was carried out to obtain information to construct a quantitative risk assessment model of VTEC O157 in cattle and sheep from farm to abattoir. To develop the model framework, a good understanding of the life cycle of VTEC O157 on cattle and sheep farms was required. To construct a quantitative model that could be used to assess the impact of control strategies on the farm, data were needed on the critical steps that might govern the prevalence of VTEC O157-positive cattle and sheep entering the slaughterhouse. Therefore, the review was structured to document the data available on the prevalence, epidemiology, survival and control of VTEC O157 on cattle and sheep farms. These data were obtained largely from the published scientific literature on VTEC O157 as was knowledge of the life cycle of VTEC O157 on farms. An understanding of the ecology of VTEC O157 on cattle and sheep farms was enhanced greatly by personal contact with other researchers working in the field, both formally through the VTEC O157 workshop and in informal discussions.
The ecology of VTEC O157 on farms has been reviewed by several authors (Hancock et al., 1998; Wallace, 1999; Synge, 2000; Hancock et al., 2001; Rasmussen & Casey, 2001). The following key points formed the working hypothesis for the risk assessment model.
1. VTEC O157 is present on virtually all cattle and sheep farms in Great Britain although intermittently. However, animal prevalence, at between 2-16%, is low. Prevalence is highest in cattle between 2-24 months of age and lower in adult cattle and calves less than two months of age.
2. Faecal shedding of VTEC O157 by cattle is transient and rarely lasts for more than eight weeks.
3. The number of VTEC O157 shed and the duration of shedding by individual cattle are highly variable but are generally greater in calves than in adult cattle.
4. Cattle can have more than one episode of VTEC O157 colonisation and shedding.
5. Faecal shedding of VTEC O157 by cattle tends to be seasonal with animal prevalence being higher between late spring and early autumn than in winter.
6. The pathogenesis of VTEC O157 colonisation and shedding in sheep is similar to that in cattle.
7. Species other than cattle and sheep can serve as transient hosts for VTEC O157, e.g., goats, deer, horses, pigs, dogs, wild birds, rats and flies.
8. VTEC O157 may survive in faeces, manure, soil and water for many weeks.
9. Because VTEC O157 is common on cattle and sheep farms, is not host specific and can survive in the environment for long periods, it is not technically feasible to eliminate the pathogen from livestock. Therefore, effort should be directed at strategies to reduce the load of VTEC O157 in the environment, the number of cattle and sheep on a farm that shed the organism and the magnitude and duration of shedding particularly in slaughter cattle and sheep.
10. Potential risk factors that might be amenable to manipulation to reduce the load of VTEC O157 in livestock and the environment include diet, feeding regime, manure and slurry management, feed and water sources and delivery systems, stocking rate, grazing and housing systems, and vermin control.
Data for the risk assessment model were obtained on the infectious dose of VTEC O157 and the magnitude and duration of faecal shedding of VTEC O157 over time in cattle and sheep. The few longitudinal studies on VTEC O157 shedding patterns in individual animals over long periods of time were particularly useful in assessing the transmission of the organism within groups of cattle. Prevalence rates of VTEC O157 shedding in cattle and sheep in Great Britain were obtained from recent DEFRA surveys. There were data available on the survival of VTEC O157 in cattle effluents, soil, pasture, water, water trough sediments, feed and silage. Most data came from laboratory studies and some studies investigated factors affecting the survival of VTEC O157 in the environment, usually temperature. Data on risk factors associated with VTEC O157 shedding in cattle were sparse and often conflicting.