|........||Practical Illustrations of risk
assessment and risk management in the Poultry Industry.
Paul McMullin MVB DPMP MRCVS
British Veterinary Poultry Association
Poultry production systems are complex, multi-layered and deal with perishable or short shelf-life products. It is common for feed-mills, hatcheries, farms and processing plants to be owned and/or managed by the same company. Effective recognition of hazards and informal assessment of risk has long been the basis of progress in the industry. One early example was the eradication of the poultry specific pathogen Salmonella pullorum from commercial breeding stock. Many other hazards are dealt with on a daily basis. Many developments are implemented through a "quality circle" approach involving monitoring, analysis of data, planning interventions and applying the appropriate intervention. HACCP systems essentially apply this principle at a series of control points in the food production chain. The industry has always been good at collecting production data, but increasingly is also collecting microbiological and other data relevant to the control of poultry disease and potential food-borne infections. Quantitative risk analysis is a method applied to the data to calculate the relative chance of specific undesired outcomes. In addition to the data it may incorporate expert opinion and information on the severity of outcome in the model. Examples dealing with control of Salmonella infections in meat poultry, and the use of digestive enhancers, are used to explore the benefits and limitations of risk analysis.
Risk is an inherent feature of modern poultry production. Production systems are complex and both the intermediate products (hatching eggs, day-old chicks etc.) and end products (meat and eggs) are perishable. Poultry production is based on a pyramid such as is illustrated in figure 1.
At the apex of the pyramid is a very small population of elite breeding birds. Successive generations both within the primary breeding company and at the level of commercial farms means that 1 male selected by the primary breeder could theoretically contribute genes to up to 20 million broiler chickens. In this way genetic progress can be rapidly distributed among the commercial poultry population, but, on the other hand, genetic and infectious problems can be too. For all of these reasons the practices of hazard identification, followed by informal risk assessment and implementation of risk management measures have for many years been the normal way of doing things in the poultry industry. The sorts of routine hazards which must be considered are:
Physical, - incubation or brooding conditions, weather
Most of these hazards are of no direct relevance to food safety, and formal risk assessment is often not required because the magnitude of risk is patently obvious (such as, for instance, the effects of a power failure in a hatchery or controlled environment farm). Probably the first example of a risk assessment and management exercise in modern poultry production was the Pullorum Disease eradication programme initiated in the 1930's. This disease is caused by a host-specific strain of Salmonella which causes severe disease. There is a high rate of maternal transmission (from hen to chick), yet low transmission among mature hens. The development of a simple on-farm blood test allowed the implementation of a test-and-remove-reactors programme to develop breeding flocks free from the disease. Similar strategies (with some modifications) have been applied to eradication programmes for a number of other diseases, such as Mycoplasmosis, Avian Leucosis and so on.
The above example demonstrates that we have been practising risk management in the poultry industry for some time, even if we did not always call it this. Other practices aimed at controlling and improving product quality have, however, a more obvious and direct relationship to risk analysis. One such system has been developed by and promoted by the American management consultant W. Edwards Deming from the 50's to the present. His main innovation was to use a statistical approach to solve practical problems, but he also promoted a concept of "continuous improvement" and "Total Quality Management". One tool used to good effect by Deming is the "Quality Circle", one version of which is illustrated in Figure 2 below. This is characterised by a data-collection phase, analysis of the data, planning an intervention, and implementing the intervention. Then the process starts all over again. It could be argued that, in one way or another (with some insight, luck and serendipity thrown in), most technological developments are based on this approach. The strength of this approach is that is essentially experimental, at least in the sense that it has a specific goal. Any results obtained, however, are immediately applicable because they are generated under practical conditions. On the debit side it could be said that it is reactive rather than proactive, that it may not take into account unintended side-effects, and that it may actually introduce cyclical variation. All of these risks associated with this model can, once they are recognised, be taken account of.
The HACCP (Hazard analysis and critical control points) system (Cross, 1996) could be viewed as a specific application of the Deming circle. What it adds is the recognition that in any production process there will be a whole series of possible quality circles. It, in effect, states that such activities should be concentrated on those points of the production chain which are known to most affect end product quality - the critical control points. It is also intended to be proactive in that it aims to identify and control hazards before they cause an adverse effect.
Quantitative Risk Analysis.
The analysis component of both the "Deming Circle" and the HACCP system will commonly involve collating available information and calculating the effects of various "what-if" scenarios. Electronic spread-sheet programmes are an extremely convenient, quick, simple and transparent way of carrying out this analysis for simple models or systems. The technique of quantitative risk analysis (QRA) has recently been developed to more accurately represent risk in systems composed of a web of interacting factors. The key difference in quantitative risk assessment is that it attempts to "take into account every possible value for each variable and weights each possible scenario by the probability of its occurrence" (Vose, 1996). It has been proposed that techniques of quantitative risk assessment should be incorporated into HACCP systems as soon as there is sufficient information to allow this (Notermans & Mead, 1996). A number of mathematical techniques have been developed for the purpose of QRA (for example, exact algebraic solution and Monte Carlo Simulation). These techniques allow incorporation of both distributions derived from data and from expert opinion in the model. For details of the mathematical techniques, and the software tools available for this purpose the reader is referred to Vose (1996).
In order to use these techniques there are 2 basic requirements:
1. Enough must be known about the nature of the
problem, and the relationship between its parts so that
the structure of the model can be created.
Risk Analysis and Management in Practical examples relevant to poultry
1. Salmonella infection on meat
The general principles of risk assessment as applied to the food chain have been reviewed recently (Ahl and Buntain, 1997). The depiction of a food chain in Figure 3 was adapted from a figure in that paper. The concept of a food chain is particularly relevant to the control of chemical and microbiological hazards in food. Risk assessment has been proposed as a possible valuable tool in this area (Kindred, 1996) though there are still very few references in the literature to practical application of these techniques to poultry production.
These concepts of risk assessment have considerable promise for the control of food-borne infections , such as Salmonella sp.(Notermans & Teunis 1996). In fact work is already ongoing at the UK Central Veterinary Laboratory to develop QRA models for this purpose (Kelly et al 1998). The UK has one of the most sophisticated regimes in the world for the testing and reporting of Salmonella infections in food producing animals and their environment. In addition to official tests required by legislation a large number of tests are carried out in accordance with agreements between producers and retailers and under voluntary monitoring exercises. The pattern of occurrence of different serotypes of Salmonella sp. in chickens is actually quite different from that which is reported to occur in the human population, however some serotypes of considerable human health importance (e.g. S.enteritidis, and S.typhimurium ) are capable of establishing themselves in poultry production systems. Unfortunately Salmonella sp. are widely dispersed in nature - Figure 4 attempts to illustrate some of the complexity this introduces in any model of this infection. The thicker arrows indicate what are, in the authors opinion, the greater risks of transmission of these infections. Note that other infected chickens (either parents, or other commercial flocks) are only one of a large number of possibilities.
There is a voluminous literature on the epidemiology of Salmonella infections. Recently the strategies to control both Salmonella and Campylobacter on raw poultry products have been reviewed (White et al. 1997). This body of published information in conjunction with expert opinion already forms the basis of control programmes in most countries. In spite of the considerable knowledge accumulated there is still much uncertainty about what will happen in any given combination of circumstances. One possible risk analysis pathway is shown in Figure 5, but it needs to be kept in mind that there are multiple other pathways by which infection may occur. A recent paper (Vose, 1997) illustrates 2 mathematical approaches to the calculation of risk in an example similar to this. It is hoped that this approach will allow the industry to better allocate resources among the broad range of control measures now available.
2. Digestive Enhancers
Recently considerable attention has been focused on the use of antimicrobial digestive enhancers in animal agriculture. It has been suggested that there is a risk that such use has a significant impact on the ability of doctors to control certain rare but serious infections, by transfer of infection or resistance genes from the animal to the human population. To people unfamiliar with agricultural production systems digestive enhancement may seem a trivial use for anti-bacterials. The effects of such use are, however, far from trivial. In addition to their direct economic effects, they can have a significant benefit for animal welfare. Sometimes they act directly by helping control a specific disease such as necrotic enteritis in poultry, but they may also improve welfare by improving the utilisation of nutrients and reducing the volume or moisture of undigested material deposited in the animals environment. Wider environmental issues are also significant. It is currently estimated that the use of digestive enhancers in pigs and broiler chickens in the UK saves 290500 tonnes of feed, 11620 lorry journeys, 714000 cubic metres of water and avoids the need for 25538 hectares of arable land planted with cereals. In addition 532000 cubic metres of pig slurry do not need to be spread. These savings are made every year we continue to use these products. There are also benefits for society and the consumer through the maintenance of animal agriculture and food processing in the Europe, both of which provide employment and revenue. Under the current, more liberal, international trade agreements it is inevitable that unilateral bans on these production aids will result in production moving to other countries. To address medical and public concerns about the use of these compounds FEFANA (the European federation of feed additive manufacturers) plans to carry out a detailed survey on resistance patterns in intestinal bacteria from the major food species in a number of European countries. It is to be hoped that the results of this information and various other research initiatives in this area will provide sufficient information to develop an adequate risk assessment of this issue.
The Precautionary Principle
There seems to be a growing tendency for society to demand zero risk, while accepting zero responsibility, indeed this seems to be the central message of many "consumerist" organisations. The political response to this (if the pressure groups are sufficiently vocal) is to apply "the precautionary principle" - i.e. risk, no matter how small, is unacceptable. Although this may appear at first glance to be reasonable, it fails to take into account the complex web of interacting factors which make up real-life food production systems. Most importantly, it assumes that the precaution itself introduces no risk. Take, for example, the Swedish proposal to extend their ban on digestive enhancers throughout Europe. There is no evidence that this will actually improve human health, but it will certainly reduce the already meagre profitability of European poultry companies, making it even more difficult to justify the investments required to further reduce the level of Salmonella infections. It is also likely to increase the proportion of poultry meat sourced from third countries, many of which have limited control programmes for food-borne infections and no or very lax controls on the use of antibiotics in food animals. Thus implementation of a ban on a European basis may actually be counter-productive in the things it is seeking to achieve! A full risk assessment of these issues should be able to take account of all such factors.
The poultry industry is a large and well-organised system for the efficient production of animal protein foods. It has a long history of pragmatic measures for the control of animal health and other risks. More recently a structured approach of cyclic data collection, analysis, planning, and implementation has become the norm. This process underpins most technological developments. Quantitative risk assessment techniques hold considerable promise for analysing and allocating resources in complex production systems. However they are complex and cannot produce zero-risk. Whether they can produce information in a form appropriate for communication of risk to the general public has yet to be seen. It is, in any case, vital that all discussions of the results of risk analysis make the assumptions on which the analysis is based totally clear.
1. Ahl A.S. and Buntain (1997). Risk and the food safety chain: animal health, public health and the environment. Rev. Sci. tech. Off. int. Epiz., 16 (1), 322-330.
2. Cross, H.R. (1996) International meat and poultry HACCP alliance. J Am Vet Med Assoc 209
3. Deming, W.E. Out of Crisis Cambridge University Press ISBN 0-521-30553-5
4. Kelly, L., Newell, D.G. and Wooldridge, M. (1998) Quantitative risk assess ment in food safety research: Potentially contaminated poultry meat as an example application. Annual Conference Association of Veterinary Teachers and Research Workers, Scarborough.
5. Kindred, T.P. Risk assessment and its role in the safety of foods of animal origin. J Am Vet Med Assoc 209:2055-2056, 1996.
6. Notermans, S. and Mead, G.C. Incorporation of elements of quantitative risk analysis in the HACCP system. Int J Food Microbiol 30:157-173, 1996.
7. Notermans, S. and Teunis, P. (1996) Quantitative risk analysis and the production of microbiologically safe food: an introduction. Int J Food Microbiol 30:3-7
8. Vose, D. (1996) Quantitative Risk Analysis: A guide to Monte Carlo simulation. John Wiley & Sons ISBN 0-471-95803-4
9. Vose, D. (1997) Risk analysis in relation to the importation and exportation of animal products. Rev. Sci. tech. Off. int. Epiz., 16 (1), 17-29
10.White, P.L., Baker, A.R. and James W.O. Strategies to control Salmonella and Campylobacter in raw poultry products. Rev. Sci. tech. Off. int. Epiz., 16 (2), 525-541.
On the Intenet see: http://www.nal.usda.gov/fnic/foodborne/risk.htm
The author gratefully acknowledges the help of Louise Kelly in providing material and helping point me in the direction of some useful published information. David Vose also helped suggest sources of information.