Australian Quarantine Review Secretariat Australian Quarantine a shared responsibility
Download 1.49 Mb.
|
- Navigate this page:
- 7.5.1 Trends in Risk Assessment
- 7.5.2 Qualitative, Semi-quantitative and Quantitative Approaches
7.5 METHODS FOR RISK ASSESSMENTThe Review Committee was asked specifically to make recommendations on 'revisions to the quarantine risk assessment process, including the potential for greater use of quantitative methods of assessment'. Comment was sought on risk assessment methods through submissions and public hearings. In addition, the Review Committee paid particular attention to current practice and trends in risk assessment methods used overseas, especially in Canada, New Zealand and the United States. As pointed out in a number of submissions, risk assessment is a relatively new discipline — and this is especially true for quantitative approaches, many of which have become feasible only in the past few years as advances in computer technology have enabled more complex systems such as import pathways to be simulated. The submission from New South Wales Agriculture acknowledged this and noted that there is no recognised forum in Australia for discussing risk assessment methods. It also recommended that a national network be developed 'to ensure that a wide range of scientists use this technology consistently'. 7.5.1 Trends in Risk AssessmentAt present, methods used in risk assessment range from qualitative approaches (as used by quarantine authorities worldwide for many years) to quantitative approaches (in which numbers are used for all stages or steps in an import chain) — with many actual assessments using both approaches (i.e. using quantitative data where these are available and qualitative assessment where quantitative data are not readily available). Internationally, there appears to be a gradual trend towards increasing use of quantitative risk assessment methods in import risk analysis — fully for the relatively small number of instances where they can be applied, and partially in semi-quantitative approaches. This trend was evident in Japan, Canada, New Zealand and the United States. The Review Committee noted that the latter two countries in particular were devoting significant resources to risk assessment, especially to more quantitative approaches. As noted by the Lindsay Review 'the approach is being pursued by overseas countries, despite the difficulties: Australia cannot afford to ignore this trend' (DPIE 1988, p. 45). This is discussed further in Section 7.8. Comments made by the Lindsay Review on 'import risk assessment' are still appropriate, particularly for quantitative methods. The Lindsay Review noted that risk assessment:
The Review Committee believes that import risk assessment should use the method most appropriate to the import access request being considered — whether qualitative, semi-quantitative or quantitative — and that each IRAT or RAP should determine which method is most appropriate for each import access request. The perception held in some quarters that quantitative approaches are inherently 'better' or 'more scientific' than qualitative approaches is misguided — a poor quantitative risk assessment (e.g. one using poor data or using inappropriate quantitative techniques) can be quite misleading and far less scientific than a good semi-quantitative or qualitative assessment. Methods used in import risk assessment are reviewed in a range of papers, including, for example, Kellar (1993), MacDiarmid (1993), and OIE (1994) for animal health, and IPPC (1995) and McNamara (1995) for plant health. Reviews are available on trends in both animal and plant import risk assessment in the United States (Chang et al. 1994, APHRAN 1994, Bossé et al. 1996) and Canada (APHD 1994). Reviews are also available on the incorporation of economic analysis into import risk assessment (Dijkhuizen et al. 1995, 1996). However, most information on methods used in import risk assessment is gleaned by examining examples of specific assessments, whether qualitative (e.g. Cassidy et al. 1996 on risks associated with private quarantine facilities for horses), semi-quantitative (e.g. APHIS 1996 on the risks of bovine spongiform encephalopathy in the United States) or quantitative (e.g. Beckett et al. 1996, on risks associated with imports of porcine semen). There is also a significant and expanding literature on modelling the probable spread and effect of incursions of pests and diseases, based on climatic factors (e.g. the use of the CLIMEX program to predict the likely range of introduced insect pests) or epidemiological spread (e.g. Markov chain or state-transition models of the spread of pathogens in susceptible populations). An initial step in import risk analysis is to determine which pests or diseases in the country of origin of a proposed import do not occur in Australia and are of sufficient concern to warrant exclusion. Such pests and diseases may be either exotic to Australia or may occur in Australia but be considered undesirable in a particular import. Import risk analysis basically establishes a scenario tree or outline of the pathway or pathways of entry and establishment that might be associated with a proposed import. In qualitative approaches, emphasis focuses on the key points in the pathway where risk management factors can be applied to eliminate (e.g. by heat treatment of a product) or reduce (e.g. by vaccinating or testing live animals) the risk of importing pests or diseases of concern. In semi-quantitative approaches, numerical values (e.g. the prevalence of the pest or disease of concern) are applied at each point for which data are available. In fully quantitative approaches, such data are applied at all points of the pathway of entry and establishment. The Lindsay Review noted that 'as with many biological issues, it is difficult to make quantifiable judgements about probabilities in relation to quarantine' (DPIE 1988, p. 41). It is only in relatively simple cases that reliable quantitative data are available for all steps in the pathway of entry and establishment. True quantitative import risk analyses are thus the exception rather than the norm. However, simple scenario trees or pathways can be analysed in a semi-quantitative or quantitative manner even where there are gaps in data. For example, one may include an extreme value (e.g. assume that the prevalence of infection in the population of origin is 100%) for missing data points and run the simulation. One can also use expert opinion to provide a 'best guess' of the value for a particular data point (e.g. using the Delphi technique). Such approaches enable the analyst to conduct sensitivity analyses to determine whether or not the particular parameter for which data are not available has a major impact on the overall probability of an exotic agent entering and establishing. Such analysis often shows that there are only a few critical points in the pathway that have a significant effect on the overall probability of entry or establishment. If good data are available on these points, the analyst can be confident that his or her assessment is robust. However, if good data are not available on these critical points, the analyst can report that robust quantitative risk analysis is not possible until information is available to fill these gaps. A RAP reaching this conclusion might encourage applicants or research providers to commission or conduct appropriate research to fill the gaps identified (as discussed in Section 7.4.6), or decide to use a semi-quantitative or qualitative approach. Semi-quantitative or quantitative approaches can be either deterministic or stochastic. The deterministic approach assigns a single number (e.g. an amount or a probability) to each point in a scenario tree or pathway. Simulation over the whole pathway thus leads to a single value, ignoring the fact that variation is an integral component in all biological systems. The stochastic approach assigns each point a value that takes account of variation — it uses a parameter defined as a probability distribution for each point in the pathway. For example, a deterministic approach might assign a value of 10% for the prevalence of a particular disease in the population of origin. The stochastic approach would assign this a value determined by a normal distribution with a mean of 10% and a standard deviation of perhaps 0.5%, thus approximating the real range of values encountered in the population. Stochastic analysis, using computer simulation, leads not to a single value for the overall pathway but to a range of values defined as a probability density distribution. For example, a deterministic analysis might conclude that the risk of importing a particular disease is 1 in 15 000 000 per tonne per annum. A stochastic analysis of the same pathway might lead to a result of a 95% confidence that the risk is between 1 in 14 000 000 and 1 in 17 000 000 per tonne per annum. Stochastic analysis provides a more realistic estimate than does deterministic analysis because it takes account of biological variation. In most cases, there are only one or a few critical points in the scenario tree or pathway that are the primary determinants of the overall risk of entry or establishment of pests or diseases of concern. In addition, in complex situations with multiple possible pathways that each have only an extremely small probability of occurrence, the mathematics of fully quantitative assessment is problematic and not yet well defined. Such situations are assessable only by qualitative or semi-quantitative approaches even if good data are available for all points in the pathway. For most import risk analyses there are — and are likely to continue to be — data gaps that preclude a fully quantitative approach. In addition, from a practical perspective, it should also be appreciated that quantitative assessments tend to be extremely resource-intensive, requiring skilled staff, large amounts of data, sophisticated computer software and a large investment of time. Thus the Review Committee concludes that although quantitative approaches to risk analysis have some application in evaluating selected import access requests, semi-quantitative and qualitative approaches are most appropriate for the vast majority of import risk analyses. Directory: SiteCollectionDocuments SiteCollectionDocuments -> Emerging Transport Technologies SiteCollectionDocuments -> Lesson Plan What are smart goals? SiteCollectionDocuments -> Melbourne Library Service Policy Public Access Internet and Computer Use Policy SiteCollectionDocuments -> Navy Drug Screening Laboratory Jacksonville SiteCollectionDocuments -> Building Management Systems (bms) Seminar 2 Advanced Management and Improvement Opportunities SiteCollectionDocuments -> Commitments and Pledges for Training and Capacity Building 2014-15 SiteCollectionDocuments -> Galileo® and Apollo® Systems – Airline Participants SiteCollectionDocuments -> Northern Australia Quarantine Strategy 25 years of protecting Australia SiteCollectionDocuments -> Final pest risk analysis report for Drosophila suzukii April 2013 SiteCollectionDocuments -> Permitted Seeds List – 16 June 2016 Download 1.49 Mb. Share with your friends:
|