SESSION 1
Evolving inspection regimes

An FAO perspective

Grimur Valdimarsson
FAO Fisheries Department, Rome, Italy

ABSTRACT

Food safety is receiving more attention than ever before, accompanied by a wide public discussion on how best to a secure safe and wholesome food supply. The fish industry has been in the forefront of implementing HACCP based food quality and safety regulations and programs since the late 1980s, long before other food sectors did. Food production from fishery resources constitutes a very complicated scenario, not only due to the number of different species harvested or grown, but the vast number of different food poisoning agents that can be carried with fishery products. Yet, fishery products have become the most international of all food products with some 37 percent of all fish products for human consumption being traded across national borders.

As identified in various fora, there is a need to enhance the availability and harmonization of scientific information related to the complex fields of safety and quality of fishery products. Information from various disciplines, such as chemistry, microbiology and toxicology, need to be pulled together to support hazard analysis, risk assessments and epidemiological studies. Such collated information would become a valuable tool for industry and government agencies involved in fish processing. For the last four years, the Fish Utilization and Marketing Service (FIIU) of the Fishery Industries Division of FAO has been examining the feasibility of launching an Internet based system that is capable of assisting in this area. We have now launched the Aquatic Food Programme using a system termed FishPort. FishPort is based on the EcoPort platform that is under the auspices of FAO in collaboration with the University of Florida and the Smithsonian Museum. The International Association of Fish Inspectors has agreed to be co-founders of the system with FAO. Using the EcoPort technology and the multi-species base of information, we believe that FishPort will provide the necessary harmonized information platform for practical guidance on HACCP based systems.

Microbial risk assessment work of special interest to fisheries has been ongoing for several years. The assessment on Listeria monocytogenes in ready-to-eat foods is now finished and one on Vibrio spp. in seafoods is in its final phase. These are already giving new insights in controlling risks associated with these organisms.

Also, FIIU is publishing a document which gives a "broad brush approach" to assess and rank chemical and microbiological hazards in fishery products. This will be particularly useful for training in the emerging field of risk analysis.

Even if new and interesting approaches are being discussed and promoted, such as the development of Food Safety Objectives, it is necessary to further elaborate on the use of the existing tools, particularly HACCP, and how their performance can be measured.

INTRODUCTION

The total food fish supply has been growing at an annual rate of 3.6% over the last 40 years while the world population has been expanding at 1.8% annually over the same period. This means that the per caput intake has increased from 9 to 16 kg. That is a remarkable achievement despite worries concerning inefficient fisheries management in many countries which has led to over fishing of some 25% of the fish stocks that FAO monitors. This increase in fisheries activities has led to very significant increases in international fish trade with some 37% of all products now being traded across national borders. Developing countries contribute close to 50% of this trade (by value) and thus it has become a very important source of foreign currency for many of these countries (FAO, 2002). There has been little change over the last few years, in that two-thirds of global fish exports go to the European Union, the USA and Japan, but fish trade between developing countries remains low. The growing importance of fish trade for developing countries makes it particularly important for the FAO to promote the necessary harmonization of fish quality and safety approaches to ensure fair trade practices.

It is important to be aware of the fact that trade liberalization has been at the heart of the development agenda for decades and Agenda 21 (Rio conference in 1992) states that the international economy should "promote sustainable development through trade liberalization". Thus, there is a strong commitment by the international community towards harmonization and equivalence of food control systems to facilitate world trade. Through the GATT, and subsequently the WTO trade negotiations, import duties are being lowered to stimulate world trade. As a result, import duties for fish and fishery products have come down to around 4.5%, whereas meat products commonly carry some 70-80% duties. As a result of lowered duties, there is an increased incentive for countries to fend off foreign competition through technical barriers. It is also important that "tariff escalation" is reduced - fishery products were specifically mentioned in that respect during the current Doha round of negotiations. Lower import duties for value-added products could create significant new opportunities for fish processors, not the least in developing countries.

It should also be borne in mind that the fish industry is increasingly expected to meet new requirements in environmental monitoring and traceability of products, in addition to ever stiffer competition on the markets. An example is the growing number of environmental certifications undertaken through the International Standards Organization (ISO). Almost 50 000 certificates of the ISO 14000 series, which deals with environmental management, have been issued. This represents a 35% increase from 2001 (ISO, 2003). A number of fishing enterprises are among these.

With increasing demands facing the food processing industry, there is a growing need for special programmes to help developing countries implement the new international instruments.

In this paper, we will attempt to address the major issues on fish safety and quality and discuss FAO initiatives to assist member countries in developing national capacity to deal with the fast changing framework in this area.

THE FOOD SAFETY PROBLEM

The WHO states that food borne diseases are a widespread and a growing public health problem, both in developed and developing countries. The global incidence is difficult to estimate but it has been reported that in 1998 alone 2.2 million people, including 1.8 million children, died from diarrhoeal diseases. A great proportion of these cases can be attributed to contamination of food and drinking water. In industrialized countries, the percentage of people suffering from food borne diseases has been reported to be up to 30% (WHO, 2001). In the USA, for example, around 76 million cases of food borne diseases, resulting in 325 000 hospitalizations and 5 000 deaths, are estimated to occur each year. As for seafood, the US Centers for Disease Control and Prevention estimate that seafood is one of the leading causes of food borne outbreaks in the USA as seafood constituted 15% of all outbreaks with a confirmed source - a level greater than that associated with meat and poultry products, which are consumed at, respectively, 8 and 6 times the rate of seafoods. However, the number of individual cases associated with seafood were lower than for meat and poultry apparently due to the much lower consumption rate (US Accounting Office, 2001). It should be noted, however, that more than 50% of seafood related outbreaks in the USA are due to virus infections caused by eating molluscan shellfish.

More recent data from the USA shows significant increases in food borne outbreaks (CDC, 2002a). This increase, however, has largely been attributed to improvements in surveillance programmes. However, there are also reports from the US of sustained decreases in the reported incidences of Campylobacter, Yersinia, Listeria and Salmonella infections from 1996 to 2001, i.e. by 15-49%. These reductions are attributed to implementation of food safety measures, not the least HACCP (CDC, 2002b).

While less well documented, developing countries bear the brunt of the problem due to the presence of a wide range of food borne diseases, including those caused by parasites. A study made by the WHO in 1995 estimated that 39 million people worldwide, mostly in Asia, are infected with parasites transmitted by the ingestion of raw or improperly cooked freshwater fish and crustaceans (WHO, 1999).

APPROACHES TO CONTROLLING FOOD HAZARDS

Earlier approaches to controlling food hazards were end product inspection regimes, accompanied by sampling and laboratory analysis. Scientifically based sampling plans were devised as well as standards or guidelines to determine compliance of product lots. It soon became evident that this approach did not halt the trend of increased outbreaks. By the end of the 1980 it had become clear to public health authorities in developed countries that a new approach that focused on prevention was necessary. This led to the Hazard Analysis Critical Control Point (HACCP) approaches.

The fish sector is now at the forefront of HACCP based food quality and safety regulations, which have been around since the late 1980s for seafood and long before having been applied to other food categories. This comes as no surprise as food production from fishery resources constitutes a very complicated scenario, not only due to the number of different species harvested or grown but the vast number of different food poisoning agents that can be carried with fishery products. As the approach has been changing from end product inspection regimes to HACCP based systems, there has been a growing need for cooperation and consultation.

The all important Sanitary and Phytosanitary (SPS) Agreement of the WTO in 1995 obliges countries to ensure that the measures they take to protect animal and plant life and health are based on an assessment of the risks involved and that measures should provide Appropriate Level of Protection (ALP). Measures should be based on international standards, guidelines and recommendations of the Codex Alimentarius. However, a country may adopt stricter measures if these can be scientifically justified. The TBT Agreement of the WTO is aimed at preventing the use of unjustified technical barriers to trade and it includes measures to protect the consumer against deception and economic fraud.

The spirit of the new global arrangements are noble, i.e. to ensure a level playing field for all parties by applying scientific principles, promoting transparency of the process, emphasizing the preventive approach rather than end product inspection and taking a risk-based approach. Thus, it is acknowledged that we must distinguish between legitimate food safety measures and protectionism. However, there is no denying that this is a bumpy road and there is, for example, as yet no agreement on what constitutes an Appropriate Level of Protection (ALP). After the European Union and the USA legislated for the HACCP based approach, many developed and developing countries have followed suit. In 1997, the HACCP system was incorporated into the Codex Alimentarius in the form of a general guideline which it is now the basic reference for international trade disputes under the WTO's SPS Agreement.

Now there are new initiatives being suggested, expanding on the HACCP approach, such as Food Safety Objectives (FSOs). These are defined as "the maximum frequency and/or concentration of a microbiological hazard in a food at the time of consumption that provides the Appropriate Level of Protection" (FAO/WHO, 2002). However, this concept is still being debated with regard to how to make it practically operational. However, setting such objectives is clearly seen as the responsibility of national authorities, leaving the industry to live up to them.

It is important to note that the older inspection regimes of food safety systems mentioned earlier are of a "command and control" nature i.e. where governments send in regulators with a "rulebook" written by that same government. The HACCP approach is significantly different as it places clear responsibilities on the companies to set in place a system that minimizes the risk associated with their products. By not being too prescriptive, it allows the companies flexibility and stimulates innovation to achieve the objective of producing safe products. Even if there is, as yet, no perfect match between HACCP and an official regulatory system, despite considerable work in the area of HACCP auditing and verification, there is no doubt that solutions to that problem will emerge. However, it is very important that it is now widely recognized that food safety is best achieved by an integrated multidisciplinary approach considering the entire food chain. That means that those involved in production, processing and trade and, last but not least, those who consume the food all have clear responsibilities. Thus, farmers, fishermen, food handlers and processors, transport operators, distributors, consumers as well as government officials all become stakeholders in protecting public health. This approach requires not only an enabling policy and regulatory environment, but also clearly defined rules and standards as well as the establishment of appropriate food control systems.

THE RISK APPROACH AND CONTRIBUTION OF FAO FISHERIES TO CODEX

The SPS agreement states that sanitary measures must be based on an assessment of the risk to human, plant and animal life, using internationally accepted risk assessment techniques.

The agreement further states that:

• Sanitary and phytosanitary measures should minimize negative trade effects.

• Protection levels shall not be more trade restrictive than required to provide an "appropriate level of protection".

This means that there should be practically achievable norms developed even though there is no agreement on what constitutes an "appropriate level of protection". The methodology to perform microbiological risk assessment has been developed by the Codex Committee of Food Hygiene.

In the area of risk analysis, risk assessment of microbiological hazards in foods is a priority area and the Food Hygiene Committee (CCFH) identified a list of 21 pathogen-commodity combinations that require expert risk assessment advice. Two risk assessments are of current interest with regard fish and fishery products, i.e. for Listeria monocytogenes in ready-to-eat foods and Vibrio spp. in seafoods. Whereas the different risk assessments yield different results, as one might expect, they already have become valuable tools in identifying new research areas and a number of the findings are being used to support risk management decisions. For example, water salinity and temperatures can be used as an indicator for Vibrio parahaemolyticus in shellfish. The expert panel also concluded that the current routine testing of seafood for Vibrio spp. did not distinguish well enough between pathogenic and non-pathogenic strains (FAO, 2002).

The panel also concluded, based on US data for Vibrio cholera, that four cases of Cholera resulting from consumption of shrimps could be expected per 100 years.

It is clear that such qualitative risk assessments are very demanding in time and money. Therefore, a "broad brush approach" to assess and rank chemical and microbiological hazards in foods could be a good way to start systematizing such approaches.

FAO has prepared a document to fill this need (Sumner et al., 2003). The document works through examples of how risk analysis can be used for different reasons and provides back-up resources for situations without extensive library or on-line facilities. The document also includes a spreadsheet tool, Risk Ranger, which helps people do semi-quantitative risk analysis and risk profiles. Risk Ranger assigns a ranking figure to the products from 0 to 100. Low risk ranking is < 35, medium risk ranking is 39-59 and high risk ranking is > 60. It is hoped that this document can become an important tool in training risk assessors and risk managers by systematically going through the steps of risk assessment.

The approach is particularly useful for determining where to place emphasis on control measures such as in determining critical control points in a HACCP plan. It is also excellent for training purposes, as it makes clear what data are required (e.g. how is food prepared prior to eating, sizes of portions, epidemiological data, immunoprofiles of the population, etc.) and thus "demystifies" the whole process. It will undoubtedly take considerable time before results of full quantitative risk assessment are expressed in food safety legislation, standards and guidelines. In the meantime the Sumner-Ross method could become a valuable tool.

It has not yet been decided if the Codex Committee on Food Hygiene will be requested to develop quantitative microbiological standards for food in international trade, but if it does, a risk management decision will have to be reached regarding the appropriate level of protection. It is important to note that this scenario for microbial hazards is different from that for chemical hazards in that the standard setting process for chemical hazards is pre-set to a "nominal zero risk".

The setting of quantitative microbiological standards would have to take into account what is technologically possible or reasonable to achieve in terms of microbial content. Pathogens like Salmonella, Campylobacter and Listeria are frequently encountered in the food supply. We know now that almost none of our presently used processes render the final product completely free of potential pathogens or toxins and that includes ready-to-eat products.

Studies show, for example, that even under best processing conditions a high proportion of products, such as ready-to-eat smoked salmon and cooked crabmeat, contain cells of Listeria monocytogenes. The technology currently available for making these products simply cannot render these products free of this potentially pathogenic bacterium.

It is important to make clear that if such standards were developed that it does not mean going from the preventive approach back to "end product inspection philosophy". They would represent the absolute maximum values to be tolerated in foods and they would be decided on through the risk analysis process and based on what is technologically achievable.

The growing international trade is highlighting the need for harmonization and this is being stressed in various fora. That, along with the decision in 1995 of making Codex standards the tools for arbitration in WTO disputes, has underlined the importance of this work. The importance of this activity is seen by the following list of Codex achievements as of July 2003.

Codex has been undergoing a major evaluation, discussed at the last session of the Commission. Among the various decisions taken was that the structures and mandates of all Codex Committees should be reviewed by 2004, improvements implemented in the processes for standards' management and monitoring and to review principles regarding observer participation in Codex meetings. Importantly, a special fund is being established to facilitate the participation of developing countries in the Codex work.

COFI AND ITS SUB-COMMITTEES

The Committee on Fisheries (COFI) is the main governing body for the Fisheries Department of FAO. It convenes every other year as does its two advisory committees, one on trade and the other on aquaculture. Safety and quality issues are regularly discussed at these meetings.

Reporting on the progress of HACCP and risk assessment initiatives are regular items at these meetings. All these committees have emphasized the need for continued capacity building - as the international regulatory framework is becoming more demanding in terms of expertise. The issue of equivalency of safety management systems was considered to require specific attention. The Sub-Committee on Fish Trade has regularly been requested to monitor scientific developments relating to the current restrictions in the use of fishmeal for animal feeds and its alleged link to Bovine Spongiform Encephalopathy (BSE). Such a report will be presented to the subcommittee at its meeting in 2004. A request was made to FAO to conduct a survey on Salmonella surveillance programmes for fishmeal and such a study is under way.

The Aquaculture Sub-Committee made the point that there is a need to better promote Good Aquaculture Practice (GAP) and HACCP at the aquaculture production level. Many exporting member countries expressed concerns that they were often unable to keep up with ever changing international export standards, not the least in light of the fast emergence of increasingly sensitive detection methods with regard to food contaminants. FAO was urged to assist in the harmonization of standards for health and safety of aquaculture products through the Codex process.

PERFORMANCE MONITORING?

At our last meeting in Vancouver, the issue of performance monitoring was brought up. Are we perhaps coming closer to being able to monitor the performance of the whole fish production process with respect to safety? What is the reliability of the food production system? These are questions that will come to mind once HACCP becomes main stream. Running under optimal conditions, every manufacturing process has an expected non-conformity level in relation to product specifications. In optimum manufacturing conditions, a basal non-conformity level indicates the performance of the process taking into account the various elements that relate to its design and operation. Subsequently, reliability is an indication of the consistency of the manufacturing process in maintaining the non-conformity levels within specified limits while meeting the product specification. To determine what is the performance and reliability of HACCP based systems we would have to take up random surveillance programs of the finished products. The sampling could be of low frequency, but very importantly, would have to be based on random sampling.

Today, most sampling regimes by regulatory authorities are targeted and thus not usable for this purpose. This approach would focus on the overall system performance rather than clearing of individual lots and thus be an estimate of how well the preventive approach is working for a given product. Such an approach comes close to the ideas on Food Safety Objectives, that should measure the "prevalence and/or the concentration" of the negative factors in the final product. It could be foreseen that we could build up information bases with such information, on an anonymous basis, for industry and the authorities to consult to see what the food production systems can achieve. That could contribute an important answer to the question "what we can reasonably expect" from a fish production process.

TRAINING AND TECHNICAL ASSISTANCE

FAO has taken various initiatives to assist member countries in developing national capacity and programs. FAO is the specialized UN agency that has the mandate to address food safety and quality issues along the "farm to the fork" continuum. Assistance in fish safety and quality is the mandate of the Fish Utilization and Marketing Service (FIIU) of the Fishery Industries Division (FII) of the FAO Fisheries Department. The assistance has included direct training of fish control personnel from government and industry, provision of technical advice on legislation and the organization of fish inspection and quality assurance as well as assisting in building national research capacity. From 1995 to 1999, FIIU carried out workshops to train more than 1 300 professionals (Lupin, 2000). For the last two years, this training has mostly been through Technical Cooperation Projects funded by FAO and a special FAO programme termed the WTO Umbrella programme.

This training coupled with provision of technical advice has enabled many countries to obtain recognition by importers, not the least the EU. The current EU "List 1" of 73 countries from which fish imports to the EU are authorized, 63 are developing countries. Yet, it is evident that far more training is needed for HACCP implementation and assessment as well as for the emerging risk analysis approach. It is clear that HACCP implementation takes time, even for developed countries. This can be seen from progress made in that area in the USA - 31% of seafood plants were implementing HACCP plans in the required way in 1998 but that figure had risen to 44% in 1999 (US Accounting Office, 2001).

FISH SAFETY AND QUALITY KNOWLEDGE BASE - FISHPORT

To conduct a proper hazard analysis it is necessary to integrate information from several scientific disciplines to implement adequate controls or mitigation strategies. Although the internet offers a wide range of scientific information, finding adequate and succinct information can be a perplexing experience for the novice user of the internet. Today, the delivery of information is less of a challenge than updating and linking of accurate and pertinent information. One way to achieve this is to have on-line active participation of the scientific and technical community. The concept of using a "Food Continuum" implies the integration of causes and sources of hazards from production to consumption. As discussed before, an integrated approach, in terms of risk management, for the entire food chain is therefore the logical next step in the production of safe and acceptable food products.

To work towards such an integrated internet based system for aquatic products, FAO has launched the Aquatic Food Products Initiative in partnership with other international organizations. The initiative takes a long term view and is seeking participation of scientists world wide to create an integrated knowledge base on safety and quality of aquatic food products.

The objective is to make a "level playing field" for developed and developing countries in activities such as hazard analysis, risk assessment, quality evaluation and product development. The system will create a knowledge base of synoptic scientific and technological information for practical application and training. Contributions are submitted through password protected access and presented under the name of the contributor's institution, both as a means to ensure proper recognition for the work and to ensure appropriate level of quality assurance. In addition, the author and the institution assume the maintenance of their information. Under the supervision of FAO, all contributions are peer reviewed and all texts must be fully cited.

The FishPort website (www.fishport.org) uses "EcoPort" technology which has been developed for over 10 years in close collaboration with scientists mainly involved in plant pest and crop management areas (www.ecoport.org). It operates under the auspices of the FAO, the Smithsonian Institute and the University of Florida. EcoPort is a species based system that combines the biological information of species within an ecological context. FishPort use the species information in the form of entities that are designed to describe food hazards, food quality attributes, processing methods and products. Today, EcoPort has some 1 200 contributors from 130 institutes and universities around the world. It currently holds some 35 000 photographs and almost 600 000 references.

FAO is contacting scientists and institutes to act as "gatekeepers" and contributors in the area of safety and quality of fishery products, with plans to extend the database to processing methods and composition of products at a later stage. The first meeting of gatekeepers has recently been held in Rome.

CONCLUSION

Despite apparent rises in the reported incidence of food borne diseases, excellent progress is being made in the implementation of preventative HACCP based systems. We are also making good progress in risk analyses. What has been achieved is to place clear responsibilities with food processing establishments, without being too prescriptive, and yet keep effective controls in line with government regulations. The assigning of responsibilities to companies has created incentives for them to be innovative in achieving "constant improvement". The message that modern safety and quality systems can produce safe food - but not without any risk - seems to be getting through to politicians and the public. Yet this needs to be repeated, that no food production system can produce food that is risk free for everybody at all times. Similarly, the responsibilities of the consumers themselves are becoming recognized. This is in line with the commonly accepted notion of traffic safety, that the authorities, the car manufacturers and the travelling public all have responsibilities to human safety.

The main reason why the move from "command and control" approaches towards more "self control" became so widely accepted is simply that it made good sense. To prevent the disease causing agents getting into the products in the first place was a convincing argument, perceived as fair, reasonable and achievable.

The scientific underpinning of the process has been a slow one - but has progressed to a point where there is now a large body of experience and large amounts of data on hand. The underlying philosophy of placing more responsibility on the industry itself has created an incentive for constant improvements to produce foods "as safe as technologically and economically practicable".

REFERENCES

CDC. 2002a. Food borne disease outbreaks reported to CDC January 1, 1990 through March 15, 2002. Online at http://www.cdc/gov/ncdod/dbmd/outbreak/cdc_reported.htm.

CDC. 2002b. Preliminary Food Net data on the incidence of food bore illness - selected sites, United States, 2001. Morbid Mortal Weekly, Rep. 51, 325-29.

FAO. 2002a. The State of World Fisheries and Aquaculture. Fisheries resources: trends in production, utilization and trade. p.3-45. FAO, Rome.

FAO. 2002b. Risk Assessment of Campylobacter spp. in broiler chickens and Vibrio spp. in Seafood. Report of a joint FAO/WHO Expert Consultation, Bangkok, Thailand 5-9 August 2002, FAO. Food and Nutrition Paper 75.

FAO/WHO. 2002. Principles and guidelines for incorporating microbiological risk assessment in the development of food safety standards, guidelines and related texts.

ISO. 2003. 12th Cycle of the ISO Survey. Online at www.iso.org.

Lupin, H. 2000. Fish quality and safety. In: The State of World Fisheries and Aquaculture, p.47-51. FAO, Rome.

Sumner, J., Ross, T. & Ababouch, L. 2003. Application of risk assessment in the fish industry. FAO Fisheries Technical Paper (in press). 91 pp.

United States Accounting Office. 2001. Food Safety: Federal oversight of seafood does not sufficiently protect consumers. GAO-01-204.

WHO. 1999. Food safety issues associated with products from aquaculture. Report of a joint FAO/NACA/WHO Study Group. WHO Technical Report Series No. 883. Geneva. 55p.

WHO. 2001. Food safety and food borne illness. Fact Sheet 237. Online at www.who.int.

Developments in the US Seafood Safety Program

Philip Spiller
Office of Seafood, United States Food and Drug Administration, College Park, United States of America

INTRODUCTION

Thank you. It is a pleasure to be here today.

I have had the honour of participating in three of the four congresses held to date and look forward to this one. It has been my experience that these sessions provide us with an opportunity to learn about what is going on in the world of fish inspection and seafood safety programs generally; be a little thought provoking; and hopefully come away with a few ideas to integrate into our own programs.

Later today and tomorrow two of my United States colleagues will offer their perspectives on some developments in aquaculture and on HACCP in the United States.

We look forward to hear what others have to say on those topics and on others that will be addressed this week.

DEVELOPMENTS SINCE THE LAST CONGRESS

I have been asked to discuss what has happened in the U.S. seafood safety program since the last Congress - how our inspection regime has evolved.

In other words, "where are we?".

I appreciate being asked that kind of question because it compels me to take a pause from the day-to-day hustle and think about it. And, after some pondering, I came to the realization that in a number of respects we are at the beginning of things. Some of these "beginnings" involve actions that we are taking or about to take. Some may be a little further into the future. But what intrigues me the most about them is that they seem to involve new ways of thinking for us and new ways of looking at the world around us. And you never know where that sort of thinking can take you. Moreover, regardless of where it takes you, there is always the possibility that getting there will involve an interesting ride.

This morning I would like to share two examples of "beginnings".

As a preview, on Wednesday, I will provide an additional example of a "beginning", in the area of food security - an area that has received a considerable amount of thought and attention in my country since the last congress. We've all become comfortable with using HACCP and HACCP-based thinking to help us prevent unintentional contamination, but we have discovered that HACCP alone is not well suited to controlling harm that is intentionally caused. So we are using some novel forms of risk assessment modelling to help us anticipate and prevent intentional harm to the food supply. But, that's for Wednesday.

For today, I've chosen another example of a "beginning" in the area of risk assessment, unrelated to food security. As background, two years ago at the congress in Vancouver, I noted that risk assessment, and especially microbiological risk assessment, appeared to be an idea whose time had come.

But I also noted that risk assessment poses some significant challenges to the risk manager to understand what it is and what it is not, as well as how to use it rather than be used by it. Now, two years later, we are starting to figure out how to use risk assessments in ways that could enable us to make better risk management decisions than we would be able to do without those risk assessments. I would like to share that experience with you this morning.

A second example of a "beginning" that I will talk about this morning is in the emerging area of nutrition. That's very different from my other examples, but it does suggest that the beginnings that I am seeing cover a lot of ground.

RISK ASSESSMENT AND "WHAT IF" SCENARIOS

So lets get started.

My agency is about to publish in final form two major risk assessments relevant to seafood products: one on Vibrio parahaemolyticus and one on Listeria monocytogenes. There are also "international" risk assessments now underway or contemplated under the auspices of the FAO/WHO that have drawn resources from a number of countries.

These risk assessments are intended to be important tools to help us understand more about the hazards we would like to control. But like all tools, we must learn how to use them effectively and, if your experience has been like mine, you know that it might not always be immediately clear how to do so. It is imperative that every risk assessment is designed to answer questions that the risk manager needs answers to in order to make risk management decisions.

A risk manager cannot sit back and wait for a risk assessment to tell him or her what to do. These things are not "magic bullets." They do not make decisions for you.

But I think we are beginning to understand two things:

1. First, how to design the risk assessment, and

2. Second, how to use the risk assessment to maximize the opportunity for the risk manager to obtain important insights that might otherwise go unrecognized, and that might even facilitate some "breakthrough" thinking in designing food safety systems.

Here is the basis for my optimism.

In 1997-1998, the United States experienced very unusual outbreaks of Vibrio parahaemolyticus from the consumption of oysters, involving hundreds of cases. The Food and Drug Administration worked with the states to develop a national "interim" control plan to prevent this kind of occurrence from happening again. The control plan involved a substantial sampling effort for the particular pathogenic strain of Vibrio parahaemolyticus that caused most of the outbreaks in 1997-1998. Because the strain was so novel in the United States, the plan involved new methodologies, training and resources.

The strain has not been found since. Moreover, the search for that strain has not had any impact whatsoever on controlling the sporadic cases of Vibrio parahaemolyticus that occur every year. In defence of the control plan, it wasn't designed to do that, but consistent with its design, it doesn't!

So we need to ask ourselves:

• Is the work misdirected?

• Should the control plan be redesigned?

• If so, how?

As a result of the outbreaks, the Food and Drug Administration decided to conduct a risk assessment for Vibrio parahaemolyticus, largely as an academic exercise in the very new field of microbiological risk assessment.

The risk managers - us - originally questioned this decision because we already had a national control plan in place - even though it was interim - and we didn't know what the risk assessment could tell us that we did not already know.

One reason why we questioned it was because the risk assessment could not tell us much about the outbreaks that generated the risk assessment in the first place. We were frustrated at first, because we wanted to understand why the outbreaks had occurred. We still do, actually. The outbreaks essentially came and went. They could have been caused by bilge dumping from cargo ships or by El Nino, or by some combination. But they are not something we see in a normal year and we have not seen them since.

On the other hand, the risk assessment could tell us a lot about the sporadic cases that occur every year in the United States, even though we were not originally interested in them because of the perceived very low level of occurrence. For one thing, it suggested that we have a much bigger problem in the United States than we realized: several thousand illnesses per year rather than less than 50 that were previously assumed based on low rates of illness reporting.

Moreover, the risk assessment was able to identify a number of risk factors that seem to have a cause and effect relationship with levels of sporadic illness. The risk factors are things that could be managed more effectively than taking a sample and hunting for a strain of Vibrio parahaemolyticus that we could not find.

Things like:

• Water temperatures,

• Season of the year,

• Percentage of pathogenic strains in the water versus non-pathogenic strains,

• Time from harvest to refrigeration.

The risk assessment could also rank the risk factors for us, from most to least likely to impact illness. But what the risk assessment did not do initially was give us an idea of how much impact each risk factor might have on illness.

Even for the most highly ranked risk factors, we could not tell whether manipulating them through regulatory controls would have a significant impact on illness or a small impact, or whether it might be necessary to manipulate risk factors in combination. For that reason, we continued to be unsure of how or whether the risk assessment could really help us.

The breakthrough might have come once we started realizing how we could essentially interact with the risk assessment by asking what we call "what if" questions about the risk factors.

For example:

• What if we could cut off harvest at a very specific water temperature? Or at a slightly different temperature? What would be the impact on illnesses?

• Or what if we could manipulate the time from harvest to refrigeration to be a specific length of time - what would the effect be on illness?

• Or what if we allowed harvest only up to a certain level of organisms in the shellfish?

• Or what if we manipulated multiple risk factors?

We are using the results of as many "what if" scenarios as we can dream up to model the effect of various mitigation strategies and reconsider our initial control strategy for Vibrio parahaemolyticus. To date, these "what ifs" suggest that preventive control strategies such as reduced time from harvest to refrigeration and certain relaying practices could result in substantial reductions in sporadic cases. These results point us in very different directions from our original sampling strategy.

Time will tell how all this works out and whether this kind of analysis lends itself to other product/hazard combinations.

Even so, there are three initial lessons from this experience that I would like to pass on:

1. Tell the risk assessors that the major outputs of risk assessment should be the factors that affect risk.

2. Then tell the risk assessors that you will want to see how variations in these factors could effect the rate of illness. These are your "what ifs."

3. All risk assessments contain assumptions to fill in gaps in the data. However, I strongly recommend that you avoid attempting to generate either risk factors or "what ifs" on the basis of these assumptions if the assumptions will have a significant effect on the outcome. If data are needed, get the data first!

One reason why our risk assessment was able to generate risk factors and "what ifs" is because we had a lot of data on aspects like water salinity, water temperatures at various times of the year around the country and time from harvest to refrigeration. We really were in great shape to conduct a risk assessment, although we didn't understand that at first.

And as a final reminder: the results of our "what if" scenarios will never tell us how to manage. That decision will still come down to judgment and values based on the cost and consequences of regulatory burden balanced against the amount of illness we are willing to tolerate.

NUTRITION

The idea of risk management as a product of judgment and values (in addition to being science-based, of course) provides me with a transition into my next example of a "beginning", nutrition. It essentially takes the form of a question:...And that is, whether and how we might take nutritional outcomes into account when we assess risk and make risk management decisions involving human food safety.

The Commissioner of the US Food and Drug Administration, Dr. Mark McClellan, is an advocate of the preventive value of good nutrition. For that reason, his strategic action plan for the FDA includes a strong emphasis on the nutritional component of FDA's food program, with an initial focus on the control of obesity.

What has been characterized as an obesity epidemic has become a major topic of discussion in the United States. It is being extensively covered in the media and it is affecting policy at public health agencies such as mine.

To show how no stone is being left unturned, the National Academy of Sciences, the institution that advises the government on scientific matters of the highest importance, recently issued a 450-page report on the "Nutrient Requirements of Dogs and Cats". According to the study, 25% of dogs and cats in westernized societies are obese.

And on an even lighter note - no pun intended - a few weeks ago I read an article in the Washington Post about a popular restaurant in Seattle, Washington, that recently made the decision to serve the most fattening dessert it could design (called "the Bulge") and that a customer who orders it must first sign a waiver that he or she will not bring an obesity-related lawsuit against the restaurant. The owner of the restaurant has been interviewed by many foreign radio stations and repeatedly asked: "What are you crazy Americans doing now?"

Well let me share with you one serious thing that we are doing, as expressed by Commissioner McClellan in an interview that can be found on the FDA web site. In the interview, he points to significant studies that indicate a benefit for the heart from consuming a diet high in the omega-3 fatty acids that are found in fish, although he also cautions that this is not a settled scientific issue. He goes on to say that shoppers in the US might soon be seeing salmon, which is high in omega-3 fatty acids, with a "qualified" health claim on the label that would make clear that scientific evidence suggests a relationship between consumption of omega-3 fatty acids and health, but that the evidence for that relationship is not conclusive.

The use of this kind of "qualified" labelling is relatively new in the U.S. The intent behind it is to improve public health by making sure that consumers are using the best available scientific evidence to guide their food and dietary choices.

So we may be at the beginning of new health claims on seafood labels. Time will tell. My office has not traditionally involved itself with either health claims or labelling. We focus on risks and hazards and how to control them. But I cannot help but wonder whether and how we could take nutrition into account in the management of risk.

For example, when we decide to reduce risk either by removing products from commerce, or by establishing tolerable levels for contaminants, or by issuing advisories, should we consider the nutritional consequences of the actions that we take, either positive or negative? And should we take those nutritional consequences into account when we take regulatory actions?

Commissioner McClellan's strategic action plan for FDA contains an initiative on effective risk management that says our "decisions will weigh, where appropriate, not only the risk-to-benefit profile of the product itself, but also the risk versus the benefit associated with Agency actions".

With that in mind, let me share with you some of the things that I have seen recently.

In the popular press, a story in the Washington Post last year, entitled "The Omega Principle: Some Fish Fats Protect the Heart. What If They Could Also Treat Your Brain?" Some assertions in that article, which I am simply reporting and not endorsing in any way, were:

• Omega 3 fatty acids found in fish are already prized by cardiologists for protecting the heart against inflammation that can lead to blocked arteries.

• A recent article in the journal Circulation found that men who ate fish twice a week had lower heart rates than men who ate fish less than once per week.

• In addition, omega-3s are now under investigation for treating depression, post partum depression, bipolar disease and similar mental disorders.

• A Lancet article in 1998 described a comparison of rates of fish consumption worldwide to rates of depression. Nations with the highest fish consumption had the lowest rates of depression and vice versa.

• A 2002 study published in the Journal of Affective Disorders reported that rates of postpartum depression are 50 times higher in countries where women do not eat fish. It was stated that the reason for this is that during pregnancy, the mother is the sole source of a certain omega 3 fatty acid and the mother's stores become depleted if she does not consume enough of it in her diet.

• A study, published in 1999 in the Archives of General Psychiatry, of people suffering from bipolar disorder reported that omega 3s significantly lengthened periods of remission when taken along with standard treatments.

• A study in England reported that an omega-3 benefited mood in people whose depression was not well controlled with antidepressants.

The Washington Post article reported that the American Psychiatric Association has formed a committee to review the findings on omega-3s for the purpose of making treatment recommendations. As a cautionary note, however, the article also advised that there are studies with omega 3s that have not found significant effects and that a National Academy of Sciences panel has concluded that there is not enough evidence yet to set a recommended daily allowance.

ABC News.com has recently published articles entitled "A Depression Free Diet" and "Prozac from the Sea, Study: Fish Oil Contains Natural Ingredients that Help Treat Depression." And that's just the popular press. The number of journal articles on my desk on this topic is substantial.

Which brings me back to the original question: "What is a seafood risk manager to do in the face of these possibilities?"

One way to look at it is: do not take them into account. Based on this view, the job of the risk manager is to control hazards. And that's it. Nutrition doesn't enter into the equation. Actually, in many situations, it probably does not have to and it may well be that in some countries the risk manager is really not in a position to consider much else unless there are modifications to the country's food safety laws.

I am not interested in exploring that legal issue today. The question that I am posing is as much philosophical and theoretical at this stage as it is anything else. Assuming for the sake of this discussion that one could take nutritional implications into account, how would one go about doing it?

One way, I suppose, would be to do so quantitatively, such as by:

• Assigning a negative numerical value to a hazard/risk combination that could occur from consuming a particular product;

• Then assigning a positive numerical value to the nutritional benefits from consuming that product;

• Then somehow selecting a risk management option that strike a balance between the two numbers.

I admit that I wouldn't know how to do that, although our economists might be reasonably comfortable with this kind of model. In the more immediate future, however, there are things that we can start asking ourselves.

First, should we apply the criteria that have been developed for making nutrition-based health claims on labels? As I alluded to earlier, the US criteria are:

1. "Weight of the evidence" for a qualified health claim; and
2. "Significant scientific agreement" for a claim without qualification.

But when making a safety-based regulatory decision, what if there is a reasonable possibility of adverse nutritional consequences but not significant scientific agreement or even a weight of the evidence? Should you consider it? Should it affect your decision? Where and how would you draw the line?

Many risk management strategies involve the building in of margins of safety, whether numerically calculated or not. Typically, as greater margins of safety are built into a regulatory scheme, the outcome can be that less food is available for consumption or, in the case of advisories to the public, less consumption of certain foods may occur.

So perhaps it is worth asking whether the margins of safety we provide for consumers should take nutritional consequences into account. I cannot predict where we are going with this. But given the potential consequences for seafood, it's a thought process that I am looking forward to.

CONCLUSION

The things that I have talked about today will likely affect all our lives in some way, if they have not already. If my remarks stimulate some thinking and some discussion, please include me. I am just starting on these things, so I could use all the help I can get.

I will leave you today to ponder on the words of the great American baseball player of the 20th century and inadvertent philosopher, Yogi Berra, who said: "If you come to a fork in the road, take it". We have arrived at some forks, for better or worse, and it will be interesting to see where they take us.

I thank you for your time today.