FAQ: Risk assessment

The time interval between first infection -in a country or region - and first detection.

Risk analysis according to the World Organisation for Animal health (OIE) framework is the process of hazard identification, risk assessment, risk management and risk communication (Figure 1). Where hazard identification is the process of identifying the pathogenic agents which could potentially cause harm; risk management is the process of identifying, selecting and implementing measures that can be applied to reduce the level of risk; risk communication is the interactive exchange of information on risk among risk assessors, risk managers and other interested parties. Risk assessment is described in more detail in question 2. Risk is the likelihood of a specific occurrence and the likely magnitude of the resultant consequences whilst hazard refers to any condition or agent – biological, chemical or physical - with the potential to cause an adverse effect.

Risk assessment (RA) is one component of risk analysis and refers to the evaluation of the likelihood and the biological and economic consequences of a hazard spreading. Risk assessments can inform the animal and public health decision making process. In conducting a risk assessment information is collected and analysed in a transparent and systematic manner. Prior to formal animal health risk assessments, human and veterinary epidemiological information was consulted. However, there was a lack of consistency and transparency in the decision making process. The formal risk assessment process increases confidence in international animal trade, with the World Trade Organisation (WTO) recognising RAs’ role in members’ regulations. Stringent and formal safeguards are now required with the increased industrialisation of the food industry, risk assessments commissioned by the European Food Safety Authority (EFSA) have also assisted in this process.

The main steps taken to complete a risk assessment are: 1. Define a risk question, which forms the scope of the Risk Assessment; 2. Construct the risk framework outlining the possible pathways required for the unwanted outcome to occur; 3. Data collection; 4. Model construction and parameterisation if assessment is quantitative; 5. Risk estimation; 6. Sensitivity analysis if risk assessment is quantitative; 7. and lastly a transparent report outlining all model assumptions, data, references and conclusions.

The World Organisation for Animal Health (OIE) framework is useful, particularly in trade disputes. The system defines risk assessment as release, exposure and consequence assessment • release assessment – describing the routes a pathogen could be potentially released into a particular environment; • exposure assessment – describing the biological pathways within the importing country leading to exposure of either animals or humans, given release; • consequence assessment – describing the relationship between specified exposures and the potential consequences either to health or the environment. The release, exposure and consequence assessments are ultimately integrated to produce a risk estimate which gives an overall measure of the risk associated with the specific hazard.

Uncertainty refers to the lack of precise knowledge about the input values, which may be due to measurement error or to lack of knowledge. Uncertainty can be reduced with increased measurement and knowledge. Variability refers to real-world diversity in which the value varies e.g. height of people or bat forearm length. Variability can not be reduced, unlike uncertainty, by increased measurement as it is an inherent characteristic of the system.

Risk assessment and risk management are two separate components of risk analysis. Risk assessment is the procedural evaluation of the likelihood and the biological and economic consequences of a hazard spreading whilst risk management is the process of identifying, selecting and implementing measures that can be applied to reduce the level of risk. An acceptable risk means a risk level judged by the relevant competent authorities to be compatible with health protection; whether that is animal, human or environmental health will be dependent upon the scope of the risk assessment and the remit of the competent authority.

A quantitative risk assessment gives a numerical estimate of the risk by combining numerical inputs for the parameters identified in the risk pathway. Risk assessments that use single point quantitative estimates of risk (typically the arithmetic mean or worst-case value) are called deterministic risk assessments. In probabilistic risk assessment, inputs are described by probability distributions resulting in a probability distribution for the output. The output of a probabilistic risk assessment might be, for example, that we are 95% certain there will be between 10 and 35 cases of infection in a given country annually. In qualitative risk assessment, the probability of an event’s happening is described using qualitative terms such as negligible, low, medium or high. The qualitative estimates of risk for each step in the risk pathway may be combined multiplicatively to give an overall estimate of risk.

Models do not have to accurately model every single biological process in question, but need to provide a useful answer. According to Prof. George E. P. Box 1987 “Essentially, all models are wrong, but some are useful”. Determining if the model is useful requires an understanding of the underlying scientific assumptions, and the uncertainty in the model parameters and framework, so that the final estimates of the uncertainty in the predicted risk can be clearly understood. Importantly, risk assessment has to provide plausible estimates and be fit for purpose. It may be possible to validate a risk assessment model by comparing its predictions with observed values from, for example, epidemiological data or surveys. However, a different data set to that used to parameterise the model is required as a comparator. In comparing the model outputs, account would need to be taken of systematic differences between the data sets used in validation and model parameterisation.

Pathogenic agents may emerge through mutations, or through taking advantage of changes in their environment or both. Assessing the risk to human or animal health from these agents is challenging as historical epidemiological data may be limited or absent. The first step in developing a risk assessment for an emerging agent is to review the current state of scientific knowledge associated with the agent and environment. Secondly, theories should be formulated on the nature of the changes, perhaps using comparisons with similar pathogens where data are available. Where knowledge is limited, worst case assumptions may be used, and a greater reliance placed on the use of multi-disciplinary teams of experts to inform the risk assessment process. It is important for these pathogens to bear in mind that risk assessment is an iterative process, and as more information and data are obtained, the risk assessment can be revised and updated.

• Chappell, C.L., Okhuysen, P.C., Sterling, C.R., Wang, C., Jakubowski, W. and DuPont, H.L. (1999) Infectivity of Cryptosporidium parvum in healthy adults with pre-existing anti-C. parvum serum immonoglobulin G. American Journal of Tropical Medicine and Hygiene 60(1), 157-164. European Food Safety Authority (EFSA) Good risk assessment practice http://www.efsa.europa.eu/EFSA/AboutEfsa/HowWeWork/efsa_locale-1178620753812_GoodRiskAssessmentPractice.htm Accessed on 11th November 2008 at 11.30 GMT. 
• Gallagher, E., Ryan, J., Kelly, L., Leforban, Y. & Wooldridge, M (2002). Estimating the risk of importation of foot-and-mouth disease into Europe. Veterinary Record 150, 25, 769-772.  
• Lasmézas, C.I., Comoy, E., Hawkins, S., Herzog, C., Mouthon, F., Konold, T., Auvré, F., Correia, E., Lescoutra-Etchegaray, N., Salès, N., Wells, G., Brown, P. and Deslys, J.-P. (2005) Risk of oral infection with bovine spongiform encephalopathy agent in primates. Lancet 365, 781-783. Legal texts: the World Trade Organisation (WTO) agreements http://www.wto.org/english/docs_e/legal_e/ursum_e.htm #Agreement Accessed on 19th November 2008 at 11.04 GMT. Terrestrial Animal Health Code 2008 http://www.oie.int/eng/normes/mcode/en_sommaire.htm Accessed on 19th November 2008 at 11.21 GMT.
• Vose, D (2000). Risk Analysis: A Quantitative Guide. John Wiley and Sons, Chichester, UK 2000.