Australia - Measures Affecting Importation of Salmon Report of the Panel (Continued) 5.30 Second, according to Australia's own Final Report, an essential link in the chain of events necessary for imported salmon to introduce a disease was that "the pathogens must be present in biologically significant numbers to initiate infection in the host". In other words, Australia claimed that, for each remaining disease, the level of any disease agent that would be present in the salmon at issue was large enough to be an infective dose. Again, there was no scientific evidence to support this claim with respect to each of these disease agents and, in many instances, there was scientific evidence that this claim was not valid. Australia had not provided scientific evidence that all of the disease agents of concern even occurred in the edible muscle, or other tissues which would be imported, at a level which would constitute a risk of infection. 5.31 Third, Australia's claims were based on the assertion that each of the disease agents was exotic to Australia. Australia had not conducted an adequate survey of its waters or domestic stocks of salmonids and native fishes to determine if these diseases were present. Instead, it relied on the fact that the disease had not been reported and that it would turn up in routine testing. However, this approach stood in stark contrast to the approach used to determine if the disease was present in the salmon at issue. For instance, in the United States, the states of Alaska, Oregon and Washington examined salmon submitted to their fisheries laboratories by commercial fishers and processors. The testing by these states was much more extensive than that done by Australia. In fact, only a few of the disease agents listed by Australia had ever been recorded in wild, ocean-caught Pacific salmon in any life stage or in any tissue and might well have not been found had the United States used Australia's testing system. According to Australia's approach, none of the other disease agents existed in the salmon at issue. 5.32 As a result, in the view of the United States, Australia's approach to the evidence was inconsistent. Australia had to either accept that the evidence was sufficient to establish that these other diseases were not present in the salmon at issue (in which case there was no reason to ban their importation in fresh or frozen from) or else that the evidence was no sufficient to establish that these diseases were exotic to Australia. 5.33 The United States noted that many of the disease agents about which Australia claimed to be concerned also occurred in other finfish, and Australia imported large quantities of other marine finfish in a fresh or frozen condition, and imported live ornamental fish, without any sanitary controls. Although Australia claimed that it had commenced, and would commence, risk assessments on some of these other fish, it was doubtful that Australia was really concerned about the introduction of these diseases or that the risk of disease was the true basis for its ban. For over 20 years Australia banned the importation of salmon without taking any steps to assess risks from, or protect against the introduction of disease from, these other finfish. 5.34 Australia had stated that its appropriate level of protection was that "fish have been subject to such treatment as in the opinion of the Director of Quarantine is likely to prevent the introduction of any infectious or contagious disease." "Likely to prevent" would mean that it was likely that the fish would not introduce an exotic disease. Australia's own characterization of the risks from the salmon at issue made it clear that it was likely that the salmon would not introduce the disease. Accordingly, it was not clear why Australia's ban was necessary. Instead, Australia's ban appeared to be designed to ensure that there was no likelihood of any disease agent being found in imported salmon. "No likelihood of disease" was a different test altogether from "likely to prevent disease," and "no likelihood" was a much higher level of protection than "likely to prevent." The United States did not see how Australia could reconcile its ban with its stated level of protection. 5.35 Finally, the United States noted that it could not endorse all of the arguments put forward by Canada. In particular, the United States did not agree that the SPS Agreement required a "quantitative" risk assessment. Nothing in that agreement specified that the risk assessment had to be quantitative. The term "likelihood" in the definition of "risk assessment" in Annex A to the SPS Agreement did not mean a numerical value of probability. This simply recognized the constraints of scientific knowledge. Ideally the objective of a risk assessment would be to establish, through scientific evidence, a quantitative level of risk whenever this was possible. However, in many if not most cases, this would not be possible due to a lack of evidence that could link an activity quantitatively to any effect on the population covered. VI. PANEL'S CONSULTATION WITH SCIENTIFIC EXPERTS Panel procedures with regard to scientific expertise 6.1 The Panel recalled that paragraph 2 of Article 11 of the SPS Agreement provided that: "In a dispute under this Agreement involving scientific or technical issues, a panel should seek advice from experts chosen by the panel in consultation with the parties to the dispute. To this end, the panel may, when it deems it appropriate, establish an advisory technical experts group, or consult the relevant international organizations, at the request of either party to the dispute or on its own initiative." Noting that this dispute involved scientific or technical issues, the Panel consulted with the parties regarding the need for expert advice. Neither party requested the Panel to seek such advice, however both parties indicated that they had no objection to the Panel seeking advice. The Panel decided to seek scientific and technical advice as foreseen in paragraphs 1 and 2, first sentence, of Article 13 of the DSU, and pursuant to paragraph 2, first sentence, of Article 11 of the SPS Agreement. 6.2 The parties were invited to submit names of individuals expert in the subject matter before the Panel. At the same time, names of individuals were also sought from the Office international des epizooties (OIE). Brief curricula vitae were solicited from all experts who were prepared to assist the Panel. 6.3 The parties were provided the opportunity to comment on these potential experts on the basis of the curricula vitae, and in particular to state any compelling objections they might have with regard to any individual. The Panel than selected four individuals from the list taking into account the comments of the parties and the need for expertise in a number of areas (eg., risk assessment, transmission of diseases of fish, and the procedures of the OIE). These experts were requested to serve, in their personal capacities, as individual advisers to the Panel. 6.4 The Panel, in consultation with the parties, prepared specific questions which it submitted to each expert individually. The experts were requested to provide their responses, in writing, to those questions they felt qualified to address. The parties agreed that their written submissions to the Panel, including the written versions of their oral statements, be provided to each of the selected experts. The written responses of the experts were provided to the parties, and the parties were given the opportunity to comment on these. 6.5 The experts were invited to meet with the Panel and the parties to discuss their written responses to the questions and to provide further information. A summary of the written responses provided by the experts is presented below. 191 6.6 The experts selected to advise the Panel were: Dr. David E. Burmaster, Alceon Corporation, United States Dr. Christopher J. Rodgers, fish disease consultant, Spain Dr. James Winton, National Fisheries and Research Center, US Fish and Wildlife Service, United States Dr. Marion Wooldridge, Department of Risk Research, Central Veterinary Laboratory, United Kingdom Questions to the experts - Compiled Responses 6.7 By way of general introductory comments, Dr. Wooldridge noted that a number of different terminology systems were still in use even within the veterinary and animal health sphere (including fish). In the terminology system now generally advised for use within this field, risk analysis and risk assessment had different, distinct meanings although in the past they were often used interchangeably. The term risk analysis as now generally used comprised four components: hazard identification, risk assessment, risk management and risk communication. This terminology was taught in the WTO/OIE sponsored Risk Analysis Training Workshops. 6.8 In the SPS Agreement definitions, hazards were defined as additives, contaminants, toxins or disease-causing organisms which may potentially have an adverse effect. A particular risk analysis might look only at a particular range of hazards, for example disease-causing organisms; and within the scope of that particular risk analysis, hazard identification (HI) was then the process of identifying all potential hazards in the commodity being considered. 6.9 Risk assessment was the process of estimating the risk presented, by the identified hazard or hazards, for the outcomes of interest. The relevant outcomes of interest for import risk assessments were specified in the SPS Agreement. In theory, the assessment could be done either qualitatively or quantitatively. In a qualitative assessment one estimated the risk (that is, the probability) of a particular outcome in words; thus the probability of a particular outcome might be estimated as, for example, high, or low, or negligible. In a quantitative risk assessment the probability was assessed numerically and the resultant probability (P) was given a numerical value; thus, for example, P= 0.99, or P=1x10-6, or "there is a 95 per cent probability that P<0.1". In practice a quantitative assessment could only be undertaken where quantitative data allowed. 6.10 Risk management was the process of deciding whether a particular risk estimated under the pertinent circumstances was acceptable (which was a matter of judgment dependent upon various factors including the results of a risk assessment) and, if not, selecting and implementing measures that could be applied to reduce the level of risk. Risk communication included all the communication relevant to the risk analysis, including communicating results of risk assessments and obtaining information, opinions, and feedback from all interested parties on any relevant factor including scientific data and method, and acceptable risk levels. 6.11 A risk assessment was therefore only part of a risk analysis, but in answering the Panel's questions on risk assessment, it might be necessary to consider the other elements of the risk analysis within the Australian document. In addition, the two Australian documents did not use the two terms in exactly this way. They appeared to call both the whole process, and the risk assessment itself risk analysis, therefore those parts of the document relating specifically to the risk assessment have to be identified by the reader. Also, the qualitative assessment of risk often leads very directly into a risk management recommendation, making separation more difficult. 6.12 Dr. Wooldridge further noted that her main area of competence in this matter was on risk analysis and risk assessment. Risk assessment procedures Question 1. When comparing Australia's Draft 1995 risk assessment with the 1996 Final Report, what are the main differences between the two risk assessments? Would you characterize one or the other of these risk assessments as technically or scientifically more sound than the other? If so, on what basis? What, in your view, are the minimum requirements of a risk assessment? Would requirements vary depending on the product and/or diseases addressed? With regard to both the 1995 Draft Report and the 1996 Final Report, do these reports, from a technical/scientific point of view, meet the minimum requirements of a risk assessment generally accepted in the specific area of aquatic animal health? 6.13 Dr. Burmaster replied that the two reports used similar methods but came to opposite policy conclusions. He did not consider that one or the other of these risk assessments was technically or scientifically more sound than the other, but that the two reports had similar strengths and limitations. In terms of the minimum requirements for such a risk assessment, Dr. Burmaster indicated that, with some exceptions in some situations (but not in the case of this dispute), a risk assessment must use quantitative methods to estimate the probability and the magnitude of desired and adverse consequences. Neither the draft risk assessment (May 1995) nor the final risk assessment (December 1996) used quantitative methods, so neither met the minimum requirements for this situation. Dr. Burmaster noted that the requirements might possibly vary depending on the product and/or diseases addressed. For example, the minimum requirements for a risk assessment concerning insect pests in fruit orchards might differ in some ways from those for a risk assessment concerning bacterial diseases in fish. However, he believed that each such risk assessment must, at a minimum, use probabilistic methods to distinguish and to quantify the variability and the uncertainty inherent in the problem. [References to the current thinking in the risk assessment community about distinguishing and quantifying variability and uncertainty could be found in (i) Burmaster & Thompson, 1997; and (ii) Burmaster & Wilson, 1996, via the Internet from http://www.Alceon.com.] 6.14 Dr. Rodgers observed that the 1996 Final Report had been restructured and divided into 4 main sections, which were linked by an overall summary, reflecting the two previous draft reports. The 1995 Draft Report was designed as a draft import risk analysis (IRA) for public comment, which was a valuable additional step for this type of exercise. Unfortunately, this led to the subsequent 1996 reports being less specific, although the sections on disease considerations presented the background information in a more clearly structured format by using recommended risk assessment factor groupings (e.g. country, commodity, agent, etc.). The 1996 Draft Report adopted a more cautious approach, since the conclusion used expressions such as "... if the importation of product were to be permitted ...", rather than statements such as "... AQIS recommends that importation be permitted under the draft conditions described ...", as occurred in the 1995 Draft Report. On the other hand, the 1996 Final Report went one step further and categorically stated that "... it is recommended that uncooked salmon products should not be permitted entry ...". This could only be as a result of the public consultation exercise and the uncertainty arising from missing or difficult to interpret data. This approach gave the impression that the 1996 Draft Report considered the available options as a technical review but fell short of commitment. The 1996 Final Report fully committed itself. 6.15 In contrast, the 1995 Draft Report, although appearing as a straightforward literature review in some sections, appeared to use the available information to attempt a qualitative risk assessment, which lead to subjective assumptions about the various possibilities for disease introduction. As such, the 1995 Draft Report was a more useful document, in the sense of an internal risk assessment exercise, since it evaluated the data to conclude that a negligible risk existed, while at the same time recognising that the overall risk of disease introduction could not be quantified. The 1996 Final Report gave more weight to the unknown elements of the assessment and as such was more cautious, which resulted in an outcome closer to the "unacceptable" rather than the "negligible but acceptable" end of the scale. Dr. Rodgers indicated that the lack of actual data and disagreement between scientists on likely outcomes probably led to this anomaly, since subjective opinion is usually argued as unrepresentative of the true situation. Unfortunately, qualitative risk assessment contains an element of assumption which is a necessity as a substitute for gaps in the scientific data base. In fact, a range of "scientific expert opinion" could actually be very useful in these circumstances and might even be adapted to create probability distributions for a more accurate quantitative risk assessment. This would be equally true for both the disease agent factors and the unique factors related to Australia, such as the susceptibility of Australian species to exotic salmonid diseases. Only the 1995 Draft Report made an attempt in this direction. 6.16 Dr. Rodgers noted that the reports had been compiled from an Australian perspective and in this respect met the methodology requirements laid down by the OIE, which stated that countries may design their own process. The reports were supported by references to the scientific literature and other sources, as required. The question of the potential volume of trade was addressed in the 1995 Draft Report by listing the Canadian export statistics for salmon to other countries and simply estimating that about 3000 tons would be imported annually. The 1996 Final Report only referred to "commercial quantities of product", even though the number of aquatic animal units being imported significantly influences the risk assessment. The other basic factors required for a risk assessment, namely country factors, commodity factors, exposure factors and reduction factors, were considered on a disease-by-disease basis from a textual technical/scientific point of view. The potential adverse consequences of disease introduction were also outlined in the same way. 6.17 One of the more important necessities for a risk assessment, particularly a quantitative risk assessment, was good scientific data concerning the reported prevalence of infection in the exporting country. For fish or shellfish (and other animals) this generally occurred in one of three ways: data from the OIE annual returns, scientific reports and published peer-reviewed scientific literature. In practice, each of these varied in detail from basic information (OIE returns) to detailed data (scientific literature). However, even the most detailed data varied widely, in both content and usefulness. 6.18 Prior to a consideration of the factors involved in a risk assessment exercise it was necessary to address the basic question that the analysis would be designed to answer. The question needed to be as specific as necessary in order to provide useful information in support of a subsequent decision concerning the importation of a batch of fish or their products. A question usually specified the disease, or diseases, of concern, the fish species to be imported and possibly the time-frame of interest. However, the basic question might need refining and could be divided into more precise questions. Sometimes it was not clear, until the problem was considered in detail, exactly which risks needed to be estimated. As a result, the original question might need to be re-framed. Once the hazard had been identified, all the events likely to occur during importation needed to be listed in order to represent the expected pathway which would prevent the unwanted event from occurring. 6.19 The potential risk factors could be generally divided into prioritized categories, such as: - the diseases of concern and their prevalence in the species for export and any closely related species in the exporting country and surrounding areas; - the epizootiology of the diseases of concern; - the effectiveness of the disease surveillance and monitoring systems in the exporting zone/country and the powers of the government administration over the movement of fish and fish diseases; - the sensitivity and specificity of diagnostic tests for detection and identification of the pathogens of concern. The identified risk factors could also be further broken down into more specific topics as follows: (a) Country factors concerned with the diseases present in both the exporting and importing country or region and would include factors such as disease prevalence, disease characteristics (methods of spread, clinical signs, carrier state and reservoir species) and the prevalence of residual infection; the establishment/facility factor (the presence of known vectors, presence of anadromous fish to water supplies, means of processing); assessment of monitoring services (import/export policies and legislation) and transportation. Destination factors were also a consideration but would need to include host susceptibility and the end use of the import. These factors would be considered for each disease of concern. (b) Commodity factors, mainly applicable to the diseases present in the fish or products rather than live animals. These were generally concerns related to whether any infected material was removed after slaughter, the pathogen survival rate after processing (storage and transit temperatures, processing plant waste and disposal practices, elimination during storage) and the volume of any permitted trade (disease introduction is directly proportional to amount imported). Specific exposure factors related to the use of the product could also be considered in this category, or form a separate topic dealing specifically with the probability of a susceptible species being infected and transmission of a disease. (c) Risk reduction factors could be considered to include options available to reduce the probability of introducing a disease agent. These might include factors such as the origin of the product, selective destination establishments and specified product processing or treatments. 6.20 The risks were identified by drawing up a list of potential diseases of concern that would be associated with the importation of the fish product, followed by an examination of the consequences of their entry and establishment. In addition, it was prudent to identify management options to further reduce the likelihood of disease entry and to consider the socio-economic and environmental impact of disease establishment. 6.21 Dr. Rodgers indicated that actual historical data was required for a quantitative risk assessment, since it was necessary to define the distribution of the probability of a disease introduction. For disease prevalence, this information would come from surveillance and monitoring programmes. In the absence of reliable data, a qualitative risk assessment might need to be performed particularly for factors related to the potential pathogens themselves, such as pathology, epizootiology, control, survival and inactivation. A certain amount of assumption might also be necessary in these cases, although a worst case scenario could be adopted, even if expert opinion had been included as a short term substitute for actual documented data. 6.22 Dr. Rodgers noted that the current OIE guidelines for undertaking risk assessment were being completely rewritten and a draft should be available shortly. Whilst not specifically referring to fish, the principles would be applicable to aquatic animals and their products in general terms, as in the current version. Consequently, this would probably entail a change in the OIE guidelines for importation of aquatic animals and their products, as detailed in the International Aquatic Animal Health Code. 6.23 Dr. Wooldridge identified the following minimum requirements of a risk assessment, providing an explanation for each. Requirement A. A risk assessment must be transparent, that is it must be clearly set out, and fully referenced in the risk assessment report produced. Explanation B. The risk which the risk assessment evaluates must be defined and clearly set out. For an import risk assessment this includes the definition of the commodities to be included in the assessment and the definition or description of the outcomes of interest. For an import risk assessment these outcomes will usually be the various identified unwanted consequences. C. The hazard or hazards to be addressed must be defined and clearly set out. If a particular hazard has not been specified in the request for a risk assessment, this will require a hazard identification of appropriate breadth. A decision must be made as to which are the hazards of interest. If the decision is that all disease-causing organisms are of interest, then a hazard identification to identify all disease-causing organisms potentially in the commodities of interest must be undertaken. D. The potential pathways from the hazards of interest to the outcomes of interest (that is the sequence of events necessary) must be elucidated, and clearly set out. Details of any processes incorporated in this pathway (for example testing for infection) must be fully referenced where appropriate. These pathways will be based upon the (mainly biological) requirements necessary to arrive at the defined outcomes and are most clearly illustrated as a series of steps in a diagram. Full referencing allows for transparency. For most import risk assessments, there will be (either implicitly or explicitly) a risk release assessment stage, an exposure assessment stage, and a consequences assessment stage, and outcomes will generally be consequences. E. For each identified step in the pathway, information (data) must be gathered to evaluate the probability of that step occurring. This may vary from hazard to hazard, and therefore must be hazard specific where necessary. This information must be clearly set out, and the source of this information should be fully referenced. The information (data) will be either qualitative or quantitative, depending upon availability and the type of assessment being undertaken. The assessors should undertake a search which is as thorough as is practicable to find the most appropriate information for the assessment in hand. Full referencing allows for transparency. F. The overall probability of the pathway of events from hazard to defined outcome actually occurring is evaluated either qualitatively or quantitatively for each defined hazard and outcome, using the information obtained. A quantitative assessment also uses appropriate mathematical manipulations. Reasoning, mathematical manipulations (where used) and conclusions must be fully set out. A quantitative assessment must include a time-frame. This will be a probability given in either qualitative terms (for example high, low, negligible risk), or quantitative terms. A time-frame may be actual time (e.g. risk per year) or quantity based (e.g. risk per batch). Dr. Wooldridge indicated that these basic requirements did not vary with the product or disease addressed. The specific variations which would occur in each stage of the table, depending upon the assessment required, were allowed for within the defined minimum requirements. 6.24 With regard to the 1995 Draft Report, Dr. Wooldridge observed that the major part of this risk analysis document comprised a very clearly set out and transparent qualitative risk assessment. Included within the document were: - A short, clear executive summary which summarized all major points. - An introduction (Section 1) which included a definition of the risk analysis, and thus the risk assessment, being undertaken. - A background (Section 2) which included a brief overview outline of the risk assessment methodology, including the sequence of events necessary for unwanted consequences to occur (Section 2.2) and an assessment of the current disease status of Australian salmonids, necessary for consequence assessment given a disease incursion by import (Section 2.9). - A section on risk factors (Section 3) which described the potential pathways from hazard to risk in detail and illustrated diagrammatically with an assessment, where appropriate, of the effect of each of these stages on the probability of viable disease-causing organisms successfully passing through that step. - A section on diseases (Section 4) which initially gave the results of the hazard identification, which comprises a list of the defined hazards, or diseases, being considered in this assessment. This was followed by a fully referenced section on each disease hazard including, for each, a consideration of the factors involved in the evaluation of the probability of a disease outbreak due to that hazard. - A discussion section (Section 5) whose risk assessment part included summaries of the disease considerations, product conclusions, environmental concerns (including the current recorded Australian disease situation and the regulations applied to other imported fish species) and consequence considerations. Also considered was evidence from the experience of fish importation into other countries, and that of Australia before the current regulations were put into force. The discussion initially considered each potential disease hazard and then, taking into account all the relevant factors, qualitatively evaluated the probability of an unwanted outcome from import of the commodities under consideration. In addition, risk management issues were discussed, including considerations in setting the level of risk. Dr. Wooldridge specifically noted the following points made on page 217; that the mere possibility of an imported product containing exotic pathogens was not sufficient to prohibit importation; that Australia had rejected the impractical ‘zero-risk’ approach; and that quarantine decisions must be made based on the best information available and the most likely outcomes given that information. There was therefore emphasis throughout on attempting to estimate probability rather than merely possibility. In conclusion, in this analysis risk management recommendations were given resultant upon the probabilities estimated in the qualitative risk assessment. The 1995 Draft Report also included a glossary and appendices on draft import conditions, translating the findings into a recommended policy given in Appendix 6. 6.25 Dr. Wooldridge summarized that the 1995 Draft Report was technically and scientifically a clearly and well executed qualitative risk assessment with all essential requirements present, particularly with respect to the clarity of the rationale for the conclusions reached. She noted that she was not competent to comment upon the validity of the fish-related information given. 6.26 With regard to the 1996 final risk analysis report, Dr. Wooldridge noted that this document resulted from the process of risk communication entered into after the publication of the draft document. It was divided into an executive summary and four sections. The first section was described as a risk analysis, the second section contained the data used for the diseases identified as hazards. The first two sections therefore included the final risk assessment for the risk of disease introduction, as well as the risk management conclusions and decisions. These sections had counterparts within the draft document. She expressed the opinion that, for clarity of comparison, the risk assessment layout within related documents should be broadly similar but, noted that here this was not the case. 6.27 The third section of the 1996 Final Report was an economic assessment of the effect of the introduction of two of the diseases, and the final section contained the bulk of the responses to the draft risk assessment; that is, the results of the risk communication process. These sections had no counterparts in the draft document. 6.28 A risk communication process served two major functions; it might identify inaccuracies in the risk assessment process; and it would provide information regarding the acceptable level of risk amongst interested sections of the community. These were two separate types of feedback, with different implications. In Dr. Wooldridge's view, each point in the documentation of such a consultation process should refer specifically, by cross-reference to the place in the initial risk analysis and risk assessment to clearly demonstrate whether, how and exactly where and why the communication process had modified and altered the initial assessment, conclusions and resultant policy recommendations. 6.29 New or altered ‘factual’ evidence or methodology incorporated into the hazard identification or risk assessment itself should be clearly indicated, and it should also be immediately possible to see how this related to any altered conclusions reached in the assessment. This allowed the risk assessment, that is the estimation of the risk involved, to be agreed as an issue separate from any considerations of acceptable risk. Any changes in recommended risk management policy due solely to the differing conclusions of the modified risk assessment should also be made clear. 6.30 Any changes in the level of risk considered to be an acceptable risk by the risk manager after consultation during the communication process should be separately discussed and described and any resultant effect on risk management policy should be clearly differentiated from the effects of an altered risk assessment. 6.31 Dr. Wooldridge noted that this was not the case with this document as, although the risk communication responses were documented in the final section, the format of the risk analysis and risk assessment were very different from the draft. This made it more difficult than it otherwise would be to identify differences, particularly of detail, between the two risk assessments and exactly how they came about. Nevertheless some highly significant differences were present. 6.32 In comparing the 1995 Draft Report and the 1996 Final Report, Dr. Wooldridge noted that there were differences in format, clarity and transparency. The first major difference was that the processes involved and conclusions reached in the actual risk assessment were much less clearly set out than in the draft. The order had been altered and the initial contents page (v) was particularly unhelpful in tracing the information sought; data was often difficult to find. 6.33 Partly due to the methodology employed and partly due to the layout, it was more difficult (if not impossible) to disentangle the sections relevant to the assessment of the risk from the sections relevant to a consideration of risk management issues, in particular that of acceptable risk; in her opinion this was a serious flaw. In a risk analysis care should always be taken to distinguish between assessed risk and acceptable risk. 6.34 She further observed that the order of presentation was confusing. The major part of the information on the diseases themselves (Section 2) now came after the assessment in which that information was used (Section 1), as did an outline of "risk analysis" methodology (Appendix 1.7.7). The potential risk pathways considered were much less clearly laid out and less detailed than in the draft. In particular Sections 1.3.2 (Risk Analysis: Considerations) and 1.4 (Risk Analysis Considerations) were, in her opinion, poorly set out (and in addition the headings were confusing). This made the reaching of relevant assessment conclusions more difficult. 6.35 The responses from the consultation process, on which the changes in both risk assessment and risk acceptance were based, were documented in the final two sections (Sections 3 and 4), and were not easily related to the changes made in the final risk assessment as compared with the draft version. A clear separation of the responses by category (in particular whether concerning fact, or conclusion, or policy recommendations) and a system of cross-referencing these with their effect on the initial risk assessment data, risk estimate conclusions and risk management issues would enhance clarity and comparison of the risk analysis. 6.36 There was no obvious explanation as to why the finalised report should be less clear than the draft report. In her opinion it was far less transparent. 6.37 Dr. Wooldridge indicated that the next major difference was that the underlying basic methodology of the risk assessment had been altered. Both draft and final document were qualitative assessments. In the final document there was a discussion of the impossibility, given information available, of undertaking any quantitative risk assessment (page 24) and specifically, on page 25, the final document stated that "This paper discusses potential routes of entry to the aquatic environment but does not calculate the likelihood that exposure will occur.". However, not only did the final document explicitly rule out the possibility of a quantitative assessment, but the qualitative assessment then proceeded implicitly on the basis that without quantification, the assessment could only be concerned with the possibility of agent entry, rather than making any attempt to qualitatively ascribe probabilities based on the information available. 6.38 This resulted in conclusions such as the following: "There will be an unknown chance - with some degree of uncertainty about the confidence that can be had in any estimation - that imported product will be infected." (page 50). But we can work this out before we undertake a risk assessment. 6.39 In Dr. Wooldridge's opinion, looking only at the possibility of particular consequences, rather than the probability, was neither an appropriate technical method nor adequate scientific outcome from a risk assessment. This was the most important difference between the draft and the final report. 6.40 Dr. Wooldridge further observed that there were differences in data used in the assessment. A number of specific additional risk factors were considered in the final document, and this was an appropriate result of the communication process. Those identified included, for example, biofilms on processing equipment (page 25), potential additional diseases (page 26) and a potentially incorrect assumption about the amount of salmon which would be cooked (page 30). 6.41 Some data was present in the initial risk assessment that did not appear to be present in the final assessment. Examples included the tables of data given in the section on risk factors in the draft document (e.g. page 17), although with the very different format, it was difficult to be completely certain that it had not been included elsewhere. She could find no explanation for the data having been left out. 6.42 In summary, Dr. Wooldridge stated that in her opinion the risk assessment methodology in the 1995 Draft Report was superior to the 1996 Final Report, and technically acceptable, whereas that in the 1996 Final Report was not. For the reasons given above, she did not consider that the 1996 Final Report met the explicit minimum methodological requirements of a risk assessment. She noted that she was not competent to comment upon the specific information regarding the fish diseases identified, nor the completeness of the hazard identification, nor on the technical information on the processes which the fish undergo. The accuracy and completeness of this information would of course affect the final conclusions from either assessment. Question 2. The Panel notes that the government of New Zealand conducted a risk assessment with regard to the importation of Canadian Pacific salmon in 1994 (submitted by both parties). In what manner and to what extent do you consider the risk assessment undertaken by New Zealand to be technically/scientifically relevant to the dispute at hand? What is the relevance of the scientific evidence considered in the New Zealand risk assessment to the Australian Final Report? 6.43 Dr. Burmaster replied that the report completed by New Zealand in 1994 met the minimum requirements for quantification in a risk assessment of this nature. In terms of the relevance of the scientific evidence, he noted that based on conversations with a biologist familiar with New Zealand and Australia, he believed that there were many similarities between the conditions in the two countries. From his perspective as a probabilistic risk assessor (but not as a fisheries biologist), he thought that New Zealand's Report resolved many - and perhaps all - of the issues between Australia and Canada. If Australia did not agree that the New Zealand's Report resolved all the issues between Australia and Canada, he thought Australia needed to prepare a new risk assessment using probabilistic methods to quantify the risks associated with different policies. 6.44 Dr. Rodgers indicated that the 1994 New Zealand government risk assessment, referring to the risk of introducing exotic diseases of fish through the importation of ocean-caught Pacific salmon from Canada, contained some factual information which was relevant to the dispute at hand. As such it was useful background information. Data limitations at the time of the report led to a combination of both qualitative and quantitative risk assessments being carried out, although the latter was only concerned with one disease (furunculosis). The assessment analyzed the available data in a competent scientific way and used the opinion of experts, where necessary, to supplement published information and provide a substitute for gaps in the data base. Unfortunately, the lack of data in certain important areas meant that assumed outcomes were qualitative and therefore some were not relevant to the current dispute, since non-validated opinion can be open to different conclusions. In addition, the conditions applicable to New Zealand might not be of equal importance to another country and additional risk factors might need to be considered in another assessment. For instance, the environment, the fish species being farmed and the presence of susceptible host species would all have to be reassessed. Consequently, a similar risk assessment undertaken by Australia, using the same information in support of additional risk factors, might not arrive at the same conclusions. To continue with Australia - Measures Affecting Importation of Salmon: Section 6.45 191 A transcript of the meeting with the experts is attached as Annex 2 of this document.

Agreements  |  Disciplines | Trade Policy Developments | Countries | Trade & SMEs Home | What's New | SiteMap | Resources | Search | Disclaimer