Beneris project description

This page is a encyclopedia article. The page identifier is Op_en3219
Moderator:Henrik (see all)
Upload data {{#opasnet_base_link:Op_en3219}}

Objectives

The general objective of this proposal is to create a framework for handling complicated benefit-risk situations, and apply it for analysis of the benefits and risks of certain foods. The first food commodity to be used in the development of the methodology is fish.

The specific objectives of this proposal are

Objectives in developing benefit-risk analysis methods

• To develop Bayesian belief networks (BBN) to handle complicated benefit-risk situations, and to develop a decision support system (DSS) based on BBN.
• To develop improved methods for dose-response assessment, combining epidemiological and toxicological data, and apply them in combining epidemiological and toxicological information on fish contaminants (esp. dioxins and PCBs).
• To develop an integrated repository of surveillance, nutrient and food consumption data that is capable of receiving, analyzing, and disseminating the accumulated data for benefit-risk analysis and to key stakeholders.

Scientific objectives in food risks and benefits

• To review the existing databases and their availability for chemical contaminant data in Europe, and integrate available data.
• To estimate average nutrient intakes and food consumption in various subgroups based on national registries in three countries and to explore the use of the data in benefit-risk analysis.
• To estimate the health benefits of fish, and understand the effect of fish on different population subgroups (age, health, pregnancy etc.)
• To establish the association between external dose (intake) and internal dose (concentrations in the body) by analysing contaminants (PCDD/Fs, PCBs, PBDEs, organotin compounds, PCNs and Hg/methyl-Hg) from 100-200 placentas.
• To combine existing and new data from food consumption databases with data on levels of contaminants in fish. The special emphasis is on children and the developing foetus.
• To estimate distributions of nutrient intake and food consumption relevant to benefit-risk analysis in a number of populations, and also the variability in exposure among various subgroups in the population.
• To identify food consumption patterns and food choices that determine the intake of those nutrients and contaminants that are related to benefit/risk-balance of a food item.
• To explore the usability of these patterns in another country than in which they were developed.
• To find out the effects of certain policy options on dietary habits and on intake of important nutrients and contaminants (e.g. vitamin D, n-3 fatty acids, dioxins, PCBs). As an example, does a restrictive recommendation on fish eating increase meat consumption?

Objectives in dissemination

• To integrate results into updated benefit-risk assessments, and evaluate the remaining uncertainties and their importance for decision-making.
• To evaluate the integration methodology by all partners and develop it further.
• To develop an internet interface for publishing risk assessment results.
• To develop a method to publish entire benefit-risk models over the Internet using XML.
• To develop methods to collect feedback from end-users about benefit-risk analyses.
• To enhance the availability of existing databases through this interface.
• To disseminate the results and to evaluate the relevance and usefulness of the work done in the project from the perspective of an end-user / authority.

Relevance to the objectives of the specific programme and thematic priority

In order to obtain the general objective of "improving the health and well-being of European citizens through higher quality of food and improved control of food production and related environmental factors" it is necessary to develop tools for measuring simultaneously several benefits and several risks.

There are very few formal means of measuring simultaneous benefits of food. In risk assessment many guidelines and protocols have been developed to measure risks, but usually for specific purposes such as drug or chemical registration. Therefore risk assessment tends to be very fragmented, and there is no good evidence that risk assessment practised in different areas (e.g. in food control vs. air pollution abatement) leads to approximately similar results. Also there is no good evidence that assessment of different risks (e.g. cancer and developmental effects) leads to similar level of protection.

Because there are problems on both sides, i.e. benefit assessment and risk assessment, it is very difficult to cost-effectively “improve the health and well-being of European citizens”, and therefore integrated methodologies are greatly needed. This has been attempted by adopting common measures such as DALYs or QALYs. However, problems exist all the way from measuring the primary effect (e.g. toxicity) through extrapolation from laboratory experiments, or benefits from short intervention studies (e.g. effects of fish diet on serum cholesterol levels) to real-life situations with variable lifetime exposure to different beneficial and harmful factors in food, and their contribution to life-years or diseases. This means that the data used for estimating DALYs or QALYs could also be questioned.

Therefore, in order to be able even to attempt to achieve a global improvement in health and well-being, a holistic top-down approach is needed to illuminate the whole spectrum of causal chains of events from the very specific pieces of information in dietary and health sciences, exposure and consumption studies, toxicology, and epidemiology. This will require “developing and integrating research in a way that food from farming, including fishing and aquaculture, is produced, distributed and consumed along the various stages of food production chain and includes consideration of associated environmental factors and their influence on human health”, as stated in the objectives of this priority, but it also will require development of methodological tools to integrate in a reliable manner all these results which often are not commensurable. Therefore this proposal tends to go a few steps beyond the objectives of the thematic priority, and also to question some of the present methodologies in trying to achieve those objectives.

As to the objectives of this call T5.4.8.3, this proposal attempts to answer exactly those questions laid out in this STREP. Our aim is to develop comprehensive risk analysis methods, which integrate both the adverse and beneficial effects of foods. The potential impacts of environmental contaminants, esp. bioaccumulative pollutants are considered. Development of new predictive methods is planned as well as their availability over Internet to end-users.

In short, this proposal aims to measure, or at least illuminate, the total gain of benefits in certain model situations, and the respective risks involved, and then develop a holistic approach to measure the total health impact in a quantitative manner, emphasising at the same time the uncertainties and their impact by VOI (value of information) analysis. We will then attempt to rephrase the framework developed in a form of a user-friendly interface for disseminating the results of benefit-risk analyses over the Internet. This product would then be available for any other innovative purposes needing a similar analytical capability as our example cases.

Potential impact

Some recent recommendations by scientists and the discussion provoked by these recommendations exemplify very well the confusion in promotion of health benefits versus communication of health risks. Depending on the a priori assumptions one can legitimately come to completely opposite recommendations. For instance, Hites et al. (2004) concluded that because of organochlorine contaminants European farmed salmon should not be consumed more often than once in 1 – 4 months. This conclusion could only be justified, if beneficial effects of salmon consumption are totally ignored, because the restrictions would only save 40 lives (using the quantitative risk assessment model of US EPA), but could cause a net loss in lives of 5000 due to increased cardiovascular mortality (Tuomisto et al, 2004). Hence it is very important for the risk manager to understand, whether his/her aim is only to control exposure to chemicals or also take into consideration the total health impact of the recommendation. This should also be made transparent to the public.

In addition to ambiguity concerning health impacts, the case presented above also exemplifies the economic impacts of a certain decision. An existence of a whole profession (e.g. fish farming in Europe or fishing in the Baltic Sea) could be dependent on the decision, and it would be of utmost importance that the reasoning of any decision is based both on sound science and on open and transparent consideration of all factors involved, and also certainty that no hidden agendas are involved. This will set high demand for the methodology of comparative benefit-risk analysis. Comparisons will be made on a trans-European basis, and information on food consumption, composition, food habits, as well as concentrations of toxic contaminants from different countries are therefore needed. This kind of data collection would be very difficult in one country, and participation of several countries with somewhat different food consumption habits gives clear added value for the project. Also participation of research institutes and food safety agencies as well as SMEs is essential both because of different expert skills, and also different approach to the benefits and risks of food through their normal activities. The partners have good contacts with potential sources of data in different parts of Europe, and also contacts to other related activities such as ongoing EU projects Safe Foods. These collaborations are expected to be of benefit to all parties.

Contributions to standards

Maximum limits and action levels of chemicals in food in the European Union are set by the Commission following consultation with EFSA (European Food Safety Authority). Prior to the establishment of EFSA in 2002 this function was carried out by the EU Scientific Committee for Food and the Standing Committee on Foodstuffs. The opinion of JECFA (Joint FAO/WHO Expert Committee on Food Additives) is also taken into account. As an example of this, the Commission set in 2001 a dioxin limit for fish of 4 ng/kg fresh weight. Several fish species in the Baltic Sea region (Baltic herring, salmon, river lamprey, occasionally bream) exceed these limits, and the Commission has ruled a five-year derogation for Finland and Sweden to sell in their domestic market these fish until 31.12.2006. The basis for these decisions is the general target of decreasing the weekly dioxin intake in Europe below 14 pg/kg.

There has been continuous discussion on to what extent the decisions are based only on toxicity information on these chemicals, and to what extent also beneficial effects are considered. In granting the derogation for Finland and Sweden the Commission explicitly stated that the basis was to take into consideration the health benefits of fish consumption, especially vitamin D and n-3 fatty acids. The decision was, however, clearly on an ad hoc basis, and there is no systematic way of taking the benefits into consideration. Another example of an ad hoc recommendation is that breast feeding is recommended regardless of higher dioxin levels in breast milk than in infant formulas. Again the basis is the generally accepted advantages of mother's milk to the infant, but there is no formal way to estimate these advantages. This kind of approach leads to arbitrary regulations.

It is hoped that in this project a systematic framework can be created for acknowledging in a scientifically valid and transparent manner both benefits and risks of food in decision-making situations. This would greatly help all decision-makers both in European Union and nationally and help to establish more evidence-based standards rather than working on arbitration.

Contributions to policy developments

Beneris does not attempt to develop recommendations to any particular open policy question. The role of this project is to estimate risks and benefits, and develop approaches to estimate them in a satisfactory and balanced way. However, this work requires real-life case studies, and the case studies are selected in a way that some aspects are currently on the political agenda of the EU or member states. This is because the research on science-policy interface requires that there are ongoing scientific and policy process on the two sides of the interface. The results will therefore most probably have policy implications. It is the aim of this project to study the roles of scientific information and interpretation on one hand, and political considerations and risk assessment policy on the other hand. The policy-relevant case studies are an important method to achieve this aim. However, the most important result of the case studies will be an improved process of utilising scientific information in decision-making, not the recommendations to particular policy cases.