Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
EFSA SCIENTIFIC COLLOQUIUM
SUMMARY REPORT
EUROPEAN FOOD CONSUMPTION DATABASE: CURRENT AND MEDIUM
TO LONG-TERM STRATEGIES
28-29 April 2005, Brussels, Belgium
TM-A
D-07-004
-EN
-C
Largo N. Palli 5/A I-43100 Parma Italy
Tel: +39 0521 036 111Fax: +39 0521 036 [email protected]
EF
SA
SC
IEN
TIF
IC C
OL
LO
QU
IUM
SU
MM
AR
Y R
EP
OR
TE
UR
OP
EA
N FO
OD
CO
NS
UM
PT
ION
DA
TAB
AS
E
3IS
SN
1830-4737
3
3 EFSA SCIENTIFIC COLLOQUIUM
SUMMARY REPORT
EUROPEAN FOOD CONSUMPTION DATABASE – CURRENT AND MEDIUM TO LONG-TERM STRATEGIES28-29 April 2005 - Brussels, Belgium
4. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
© European Food Safety Authority – March 2008
Reproduction is authorised, provided the source is acknowledged, save where otherwise stated.
The views or positions expressed in this booklet do not necessarily represent in legal terms the official position of the European Food Safety Authority. The European Food Safety Authority assumes no responsibility or liability for any errors or inaccuracies that may appear.
ISBN: 978-92-9199-070-2
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 5.
About EFSA
The European Food Safety Authority (EFSA) was established and funded by the European Community as an independent agency in 2002 following a series of food scares that caused the European public to voice concerns about food safety and the ability of regulatory authorities to fully protect consumers.In close collaboration with national authorities and in open consultation with its stakeholders, EFSA provides objective scientifi c advice on all matters with a direct or indirect impact on food and feed safety, including animal health and welfare and plant protection. EFSA is also consulted on nutrition in relation to Community legislation. EFSA’s work falls into two areas: risk assessment and risk communication. In particular, EFSA’s risk assessments provide risk managers (EU institutions with political accountability, i.e. the European Commission, European Parliament and Council) with a sound scientifi c basis for defi ning policy-driven legislative or regulatory measures required to ensure a high level of consumer protection with regard to food and feed safety.EFSA communicates to the public in an open and transparent way on all matters within its remit.Collection and analysis of scientifi c data, identifi cation of emerging risks and scientifi c support to the Commission, particularly in case of a food crisis, are also part of EFSA’s mandate, as laid down in the founding Regulation (EC) No 178/2002 of 28 January 2002.
For more information about EFSA, please contact:
European Food Safety Authority
Largo N. Palli 5/AI-43100 ParmaItaly
Tel: +39 0521 036 111Fax: +39 0521 036 [email protected] www.efsa.europa.eu
6. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 7.
CONTENTS
I INTRODUCTION 9
II SUMMARY OF PRESENTATIONS 11
1. Food Consumption data needs for exposure assessment 11
2. Comparison of food consumption for food groups 12
3. Common methodology for collection of food consumption data in the EU – Issues identifi ed at the EFCOSUM project 12
4. Potential of the standardized EPIC dietary methodology for exposure assessment in Europe 13
5. Concise EU Food Consumption database, under development by EFSA for preliminary exposure assessment 13
6. WHO Experience with food consumption data 14
7. EUROSTAT activities in the fi eld of food consumption 15
III SUMMARY OF DISCUSSION RESULTS 16
1. Introduction 16
2. Methodological issues in the use of food consumption data for dietary exposure assessment – Report from Discussion Groups 1 & 2 17
3. Are different data needed for the various areas covered by EFSA e.g. microbes, pesticides, food contact materials – Report from Discussion Group 3 & 4 19
IV CONCLUSIONS & RECOMMENDATIONS 26
V ANNEXES 29
Annex 1: Programme of the EFSA Colloquium 33
Annex 2: Participants at the Colloquium 37
Annex 3: Presentations made at the Colloquium 43
Annex 4: Slides of the discussion groups 141
8. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 9.
Introduction
The European Food Safety Authority (EFSA) has a role in promoting and co-ordinating the development of uniform risk assessment methodologies. Exposure assessment is a crucial and integral part of risk assessment and the quality of available data both on food consumption and on occurrence levels may have a major impact on the outcome of risk assessment. For European exposure assessment it is important to make use of existing national data that is somewhat comparable at European level. At present however there is a lack of internationally comparable data on food consumption. The ultimate aim is to have harmonised food consumption data collected and available in the future. In the meanwhile, the establishment of a common database on food consumption based on data currently available would improve the consistency and reliability of exposure assessment carried out by the various panels of EFSA and by other experts in Europe. Such database will also improve the consistency of EU participations in international assessments.
To support the establishment of a common database on food consumption, EFSA organised a scientifi c colloquium on the 28th and 29th of April 2005 in Brussels, Belgium. The objective of the colloquium was to have an open scientifi c debate on the state of the art of harmonised approaches to food consumption data collection and the development of a comprehensive database on food consumption at European and international level. Consideration was given to approaches to distinguish vulnerable subsets of the population and the diversity of needs for the various areas of food safety.
Format of the colloquium
The colloquium programme is presented in Annex 1. About 100 participants (Annex 2) representing relevant expertise in the area of food consumption and exposure assessment participated in an active debate. The format of the colloquium included an introductory plenary session with presentations (Annex 3) following which participants were split into discussion groups to address specifi c aspects in more detail. Participants were provided with guidance on the remit of the discussion groups (Annex 4).
On day 1 discussion groups focused on addressing the following two issues:
Methodological issues in the use of food consumption data for dietary �exposure assessment
Are different data needed for the various areas covered by EFSA e.g. �microbes, pesticides, food contact materials?
The outcome of the debate from each group was presented back to the plenary (see section 3.0, Summary of Discussion Results).
I. INTRODUCTION
10. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
On day 2 discussion groups focused on:
The implications for the concise EU food consumption database for �preliminary exposure assessment and for medium and long-term strategies.
The fi nal session was dedicated to discussion on the fi nal outcomes of the discussion groups and focused on arriving at the conclusions and recommendations of the colloquium.
Summary of Presentations
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 11.
II. SUMMARY OF PRESENTATIONS
In order to focus the colloquium in the area of exposure assessment and food consumption, a number of brief presentations were provided by those with expertise in the area. The presentations highlighted the current situation (including limitations and potential) in relation to:
the needs of food consumption data for exposure assessment, a. the different results that can be obtained when using different data sources, b. experience from research projects and international institutions (EFCOSUM, c. EPIC-SOFT, WHO, EUROSTAT), and information from the EFSA concise EU food consumption database. d.
These presentations served to generate much debate and information for the discussion groups that followed. The main points of each presentation are summarised below and a copy of the slides is provided in Annex 3.
2.1 Food consumption data needs for exposure assessment (Catherine Leclercq)2.2 Comparison of food consumption data for some food groups (Wulf Becker)2.3 Common methodology for collection of food consumption data in the EU – Issues
identifi ed at the EFCOSUM project (Karin Hulshof)2.4 Potential of the standardized EPIC dietary methodology for exposure assessment
in Europe (Nadia Slimani)2.5 Concise EU Food Consumption database, under development by EFSA for
preliminary exposure assessment (Philippe Verger)2.6 WHO Experience with food consumption data (Gerard Moy)2.7 EUROSTAT activities in the fi eld of food consumption (Lourdes Llorens-Abando)
2.1. Food consumption data needs for exposure assessment Catherine Leclercq, Research Group Food Safety-Exposure Analysis, National Research Institute on Food and Nutrition, INRAN, Rome, Italy.
Catherine Leclercq provided an overview on exposure assessment within the risk analysis process. In particular, the needs in terms of food consumption data were presented and discussed. Ideally, in order to assess exposure, food consumption data should be recent, codifi ed at a high level of detail and representative of the whole population and, eventually, of subsets particularly at risk (e.g. infant, pregnant women, etc.). However incomplete food consumption data sets can be used to estimate exposure by making conservative assumptions. Access to raw data (data per subject) and availability of a database manager with good knowledge of the data are crucial to perform dietary exposure assessments. In any case it is very important to always characterise uncertainty of the food consumption data in a transparent way.
12. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
2.2. Comparison of food consumption data for some food groups Wulf Becker, National Food Administration, Uppsala, Sweden
Wulf Becker has given a presentation on the different food consumption data sources available in the EU. In particular, he presented the different characteristics of the following data sources: Food Balance Sheets (e.g. FAO), Household Budget Surveys (e.g. DAFNE-project) and Food Consumption / Dietary Surveys (e.g. national surveys and EPIC-project).The implications of using such data for exposure assessments were highlighted. It was noted that the use of food consumption data from different sources may produce highly variable exposure estimates.The levels of consumption were presented for a number of different foods (milk, yogurt, cheese, fruit and vegetables, and fi sh) showing that (i) different data sources may give different pictures of food consumption patterns in a country and that (ii) food patterns differ between males and females, age groups, etc., but not consistently across countries. It was fi nally concluded that harmonised data collections are needed to enable consistent exposure assessment across Europe.
2.3. Common methodology for collection of food consumption data in the EU - Issues identifi ed at the EFCOSUM projectKarin Hulshof, TNO Quality of Life, Zeist, The Netherlands
Karin Hulshof introduced the audience to the European Food Consumption Survey Method (EFCOSUM) project which was performed from 2000 to 2002. It represents a framework of the EU Programme on Health Monitoring and its fi ndings and recommendations were published in 2002 on the European Journal of Clinical Nutrition (issue 56, suppl.2).The aims of the EFCOSUM were (i) to defi ne a method for monitoring food consumption in nationally representative samples of all age-sex categories in Europe in a comparable way, (ii) to indicate how to make existing food consumption data comparable and, (iii) to defi ne a (minimum) set of dietary components which are relevant indicators of health.Activities included plenary sessions, desk research and working group activities, building on existing experience from and in co-operation with the EPIC and DAFNE project on food consumption, food classifi cation and food composition.The EFCOSUM consensus provided recommendations for a European Food Consumption Surveillance in terms of dietary method (two non-consecutive 24-h recalls and a food list to assess the proportion non-users for infrequently consumed food), data collection software, sample size, sampling procedures, recruitment, fi eld work procedures, biomarkers, training / quality control and statistical aspects.
Summary of Presentations
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 13.
2.4. Potential of the standardised EPIC dietary methodology for exposure assessment in EuropeNadia Slimani, Nutrition and Hormone Group, International Agency for Research on Cancer (IARC-WHO), Lyon, France
Nadia Slimani presented the standardised dietary methodology developed within the European Prospective Investigation into Cancer and Nutrition project (EPIC). In particular, she drew attention to the computer-assisted 24-hour dietary recall method (EPIC-SOFT) used to collect measurements from 10 different countries within the study. The objective of EPIC-SOFT was to standardise procedures across countries to describe, quantify, probe and calculate food and beverage intakes (including food supplements). Data from nutritional studies which have used EPIC-SOFT were presented to highlight the potentialities of this instrument for exposure assessment studies. Certain aspects of EPIC-SOFT were discussed in more detail (e.g. the food description system, the food quantifi cation system and validity measurements) together with current limitations.The EPIC-SOFT needs to be upgraded according to the comments received from end-users and the EU recommendations (EFCOSUM) and also according to changes on the food market or food habits (e.g. foods, recipes, food portions, food supplements, brand names, etc).A new release of the EPIC-SOFT software and databases is therefore planned and needs for its use in exposure assessment studies have already been identifi ed. In particular, it was highlighted the necessity to validate further EPIC-SOFT according to designs relevant for risk exposure assessments.
2.5. Concise EU food consumption database, under development by EFSA for preliminary exposure assessment Philippe Verger, INRA, Research unit 1204 – Met@risk
Philippe Verger presented rationale for the development of a concise EU food consumption database for preliminary exposure assessment by EFSA.EFSA panels need international food consumption data for risk assessment, but not all opinions need an accurate international estimation of the intake. Currently there are food consumption data available in the majority of European countries, however, it is diffi cult to compare the data directly between the countries due to the wide varying methodologies used in the surveys (e.g. sampling methodologies, survey duration, clustering of different age groups and using different categorisation systems for food groups). International committees currently use data based on the Global Environment Monitoring System (GEMS) - regional food diets that have been developed comprising raw and semi-processed food commodities - as well as on SCOOP data (Scientifi c cooperation tasks).
14. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Two possible ways for improvement were proposed: (i) a refi nement of the GEMS Food diets and (ii) a “combination” of national data, considering similar age groups defi nitions and food categorisation systems. Preliminary fi sh consumption results from 6 European countries were presented and compared with the corresponding GEMS data. Provisional conclusions show such data to be consistent at that level of aggregation for the considered countries independently of the survey methodology. Moreover, GEMS food consumption values did not represent systematically an overestimate for food consumption. Methods of supporting short term improvements were presented and, in order to move forward, it was suggested to (i) collect data in the EU by EFSA (ii) develop a more comprehensive database (considering more detailed food categories, subpopulations, etc.) and (iii) create a network of database managers in the EU.
2.6. WHO Experience with Food Consumption DataGerard Moy, Food Safety Department, World Health Organisation (WHO)
Gerard Moy presented the food regional diet approach developed within the Global Environment Monitoring System (GEMS) programme. He explained that all sanitary measures for food shall be based on a risk assessment and that the risk assessment shall take into account risk assessment techniques developed by the relevant international organisations. The importance was stressed to consider also the food chain (whether the food is raw, semi-processed, or processed) when consumption data are used to assess dietary exposure.In the GEMS / Food regional diet approach, raw and semi processed agricultural commodities are considered. The food daily intake in grams per capita is characterised for fi ve different regions: the Middle Eastern, Far Eastern, African, Latin American and European type by using the FAO food balance sheets. In 1997, new diets were also developed based on a cluster analysis approach using major food groups. Five-year average intakes for most foods in the 13 GEMS / Food Consumption Cluster Diets were available from the FAO Food Balance Sheet data for the period 1997-2001.The risk assessment bodies working with GEMS / Food diets are the Joint FAO / WHO expert committee on food additives (JECFA), the Joint FAO / WHO expert meetings on pesticides (JMPR) and other ad hoc study groups and consultants. The Codex Committee also uses GEMS / food-based estimates provided by JMPR and JECFA.Results obtained by the GEMS / Food regional approach concerning corn consumption were demonstrated as an example and some problems discussed. Some of the issues which arose during this exercise included failure of matching Codex and FAO codes, missing commodities and processed foods, obsolete and non representative data, together with exclusion of high percentile consumers.
Summary of Presentations
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 15.
2.7. EUROSTAT activities in the fi eld of food consumptionLourdes Llorens-Abando, DG EUROSTAT, Luxembourg
Lourdes Llorens-Abando introduced the activities of EUROSTAT in the fi eld of food consumption data collection. The Food Safety statistics working group decided to include food consumption as one of the priority themes and a Task Force on Food consumption was included in the EUROSTAT working programme for 2005. The general objective of this group is the identifi cation of the needs for food safety statistics and the analysis of the possibility to establish regular data collection on food consumption within the Environmental Stewardship Scheme (ESS).To analyse food consumption patterns and their statistical distribution, the real food intakes have to be considered as a main variable whereas retailers’ sales in volume, purchases by households and food availability, have to be considered as proxy variables. The Task Force already identifi ed a number of basic needs for food consumption data which were systematically sorted into the following fi ve categories: food market management, health and nutrition, diet related diseases, food safety and environmental issues.In particular, it was decided that environment issues and consumer’s needs must be taken into consideration when collecting food consumption data. Data must be updated every fi ve years for Health, Nutrition and Food Safety issues whereas economical variables should be updated on a yearly basis. A standard food classifi cation system is needed but disaggregation might be necessary for some special cases. Finally, consumption data must be discriminated with respect to geographic location, type of settlement, education, age and sex. Populations at greater risk should be given a particular focus in particular children, pregnant women and elderly people.Three main statistical sources of the data were mentioned. Advantages and problems were discussed in detail in relation to dietary surveys, household budget surveys and supply balance sheets.Based on all of these considerations, the main priority for EUROSTAT is to design a strategy to collect Food consumption statistics within the ESS. A mixed data collection strategy has been considered as advisable. This would mean (i) to keep annual data availability with respect to Supply Balance Sheets, (ii) to collect data on household food purchases in a systematic and harmonised way, and (iii) to harmonise the dietary surveys techniques and food items classifi cations. EUROSTAT will produce a document on “ESS strategy on food consumption statistics” and launch grants to develop a harmonised way to collect food consumption statistics at EU level.
16. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
III. SUMMARY OF DISCUSSION RESULTS
3.1. Introduction
Following the introductory presentations (Annex 3), participants were split into discussion groups to debate specifi c issues in more detail. Participants were provided with guidance on the remit of the discussion groups via a presentation by Juliane Kleiner (Annex 4).
On day one, participants were divided into four groups to allow two parallel discussion groups to address each of the following themes:
Methodological issues in the use of food consumption data for dietary �exposure assessment
Are different data needed for the various areas covered by EFSA e.g. �microbes, pesticides, food contact materials?
The discussion group themes and questions were as follows.
Specifi c Issues Methodological issues in the
use of food consumption data
for dietary exposure
assessment
Are different data needed for
the various areas covered by
EFSA e.g. microbes, pesticides,
food contact materials?
Groups Discussion Group 1Discussion Group 2
Discussion Group 3Discussion Group 4
DAY 1 1) What are the minimum quality criteria for a food survey to be used for Exposure Assessment?
2) Is there a pragmatic approach to deal with high percentiles?
3) How to deal with limitations of data?
4) How to address consumer loyalty to particular food items?
1) What are the general needs for exposure assessment?
2) What are specifi c needs for exposure assessment in the various areas?
3) Which food consumption data are available for specifi c purposes?
4) What are the possible ways for improvement of data quality?
Summary of discussion results
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 17.
The outcome of the discussion from each group was presented back to the plenary for general discussion. On day 2 discussion groups continued using earlier discussions to focus on:
The implications for the concise EU food consumption database for �preliminary exposure assessment and for medium and long-term strategies.
DAY 2 5) What are the implications for the concise EU food consumption database currently under development by EFSA for preliminary exposure assessment?
6) What are the implications for medium and long-term strategies?
5) What are the implications for the concise EU food consumption database currently under development by EFSA for preliminary exposure assessment?
6) What are the implications for medium and long-term strategies?
The fi nal session was dedicated to the fi nal outcomes of the discussion groups and to reach conclusions and recommendations of the colloquium. This is presented in section 4.0 of this report.
A summary of the discussion is presented below under the specifi c themes. The reports of the discussion groups are presented in Annex 4.
3.2. Methodological issues in the use of food consumption data for dietary exposure assessment
The aims of the exposure assessment and of the food consumption survey were raised as points to consider in determining the minimum quality required for a food survey. The quality of the food survey data will depend on the aim of the exposure assessment, whether assessing acute or chronic exposure or whether being used to estimate exposure by conservative screening (using crude food consumption data) versus more accurate assessments of exposure (using actual food consumption data). Regarding the food consumption survey, it was raised that consideration needs to be given to whether it is being conducted for traditional nutritional purposes, exposure assessment or both. If it is being conducted for nutritional purposes primarily, consideration should be given as to whether the data will be acceptable to estimate the exposure of certain food chemicals.
18. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
The minimum quality criteria required for a food survey to be used in exposure assessment were intensively debated. It was fi rst recommended that guidelines should be developed on the minimum set of data to be collected in food consumption surveys. First of all, the target population needs to be represented by means of an adequate sample size. The reliability of the sample population was raised as being more important than the accuracy of the recorded amounts. It is essential that the whole diet is addressed, this means that also food eaten outside home, e.g. supplements and drinking water, must be considered. All aspects relevant to describe the food (e.g. processing, handling, preparation, packaging, meal situation, etc.) have to be taken into account. Where necessary, consideration should also be given to the days of the week surveyed and the time of year the study is conducted, in order to address variations in food consumption due to week-days and seasons.It was also highlighted that a large number of food groups is not needed as part of minimum quality criteria. The estimation of the average, median and high percentiles of exposure is needed together with the consumption levels of the relevant food items/ categories. Details need to be provided on the uncertainty of all these measurements.
Limitations of data, under-reporting and consumer loyalty
When dealing with a limitation of food consumption data it was suggested to consider the use of different dietary assessment methods. For example a Food Frequency Questionnaire (FFQ) could be particularly useful for food that is only rarely consumed. Probabilistic modelling was also raised as being useful but the importance of the modelling validation was also highlighted. Specifi c limitations can also be resolved by means of existing statistical methodologies developed to adjust the data. This is the case, for example, of the Nusser method that is used to estimate the distribution of usual dietary intake by short-term measurements. The promotion of new research on how to estimate chronic high exposure from short-term measurements should be considered. In any case, the underlying uncertainty needs to be characterised so that it can be taken into account by risk managers. When data on body weight are missing, it was suggested to use register data or a conservative mean.
Gross under-reporting can infl uence the usefulness of dietary data as a basis for exposure assessment. It was considered fundamental to reduce under-reporting at the minimum during the survey. This objective can be met by a proper training of fi eldwork staff and by food intake probing. In order to have a better understanding of under-reporting, it could also be useful to collect additional information on subjects apart from weight and height (e.g. information on subjects who are dieting or going on a diet, physical activity data, state of health, etc.). It is also important to consider the impact of under-reporting on the exposure assessment in terms of the estimations (e.g. effect on high percentiles).
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 19.
Consumer loyalty to particular food items can have a signifi cant impact on exposure levels, but it is extremely diffi cult to identify by means of the available food consumption data. It was therefore recommended to collect more detailed consumption data (e.g. brand names, organic labelling, etc.). In alternative, the use of worst case deterministic methods and probabilistic modelling has been suggested. In particular, probabilistic modelling can be used by merging food consumption data with data from different sources, such as those from marketing research companies, consumer panels or household budget surveys.
Moreover in some cases, the agencies that requests risk assessment studies and food consumption surveys operate independently. Thus the potential uses and the resources used in a food consumption survey may not be used to their full capacity – more co-operation on such initiatives would be more cost effective and productive for both and would support the collection of more appropriate data.
3.3. Data needs for exposure assessment in various areas covered by EFSA
The fi rst point arose with respect to this theme is the level of detail required on the description of the products consumed. Different options can be adopted, for example consumption levels can be reported for food as eaten, food disaggregated into the main ingredients or into the primary agricultural products. It was highlighted that the level of food consumption data details required for an exposure assessment will depend on the agent in question (e.g. microbes, pesticides, nutrients, food contact materials, etc.) and on whether data are needed for screening purposes or more detailed exposure assessment. Many questions may be answered by means of crude screening methods and only if the estimated crude exposure assessment would give rise to any safety concern more refi ned methods are needed. It should be noted that careful consideration to the required level of detail is essential, since increasing the level of detail may also result in increasing uncertainties and amplifying the effect of variability. The agent under question study also determines the level of details needed from specifi c food or food group categories or specifi c aspects associated with the food (e.g. treatment, packaging, origin, fortifi cation, etc.). It was also pointed out that detailed information on specifi c types of food can be added separately from the consumption data only if this is really useful. The combination of data from different sources (e.g. at consumer or industry level) could be useful especially to check the validity of the available information.
Summary of discussion results
20. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Specifi c needs and availability of Food Consumption data
Data are available in Europe but are not always complete and comparable, especially regarding the food groups and the food descriptions. The adequacy of current databases was discussed. In general, detailed food consumption data should be accessible to all stakeholders. The only data that covers practically all European countries are household budget surveys, although some may cover expenditures only. National food consumption databases are available but should be kept separately and fed into a European database. Consideration needs to be given to how regional differences will be covered.
With respect to nutrition, the information on food composition is often inadequate for specifi c components, in particular for trans-fatty acids, dietary fi bre, some vitamins, and trace elements. It may be adequate for screening purposes, but sometimes composition data are completely missing for important foods, e.g. fast-food fries, producing an underestimate. Fortifi ed food and beverages can also be a problem since the levels of fortifi cation may be proprietary information. Data on the composition of nutritional supplements are also needed for proper intake assessment. Available food consumption data may not be suffi cient to assess exposure to food additives, fl avourings and food contact materials. There is little or no information on packaging or home storage containers. Especially in the case of food additives, brand name information is very useful. When dealing with food contact materials it is at least necessary to know whether food is packaged in order to perform a crude screening. Information on the type of packaging material (e.g. glass, plastic, paper, cans) is preferable including food contact material before / after purchase and surface to volume ratio.At present it is diffi cult to assess exposure to pesticide residues at a European level as the national databases differ in details provided.
Due to its aggregation level, existing food consumption survey data are generally not very useful for microbiological exposure assessment. In particular, meat consumption should be broken down at least to the level of animal species. In this case, the information considered as important, and that could be useful to obtain, are the following: heat treatment (including tap water) during industrial processing and preparation, storage conditions (time, temperature, etc.) after purchase and after preparation. Information on hygiene can theoretically be obtained from questionnaires but the reliability of these data is questionable.Moreover, the frequency of consumption and the portion sizes were also mentioned as necessary for microbiological exposure assessment.
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 21.
Summary of discussion results
The main problem in the case of chemical contaminants is that in most of the cases, the food categories used in food consumption survey do not match those in regulations. In particular, the origin of the food is often unknown.
Possible ways to improve the quality of food consumption data
In order to improve the quality of food consumption data, it was proposed to establish a network of food consumption survey managers / owners under the aegis of EFSA. This network could be used as a platform for exchange of information, validation, and provision of data, including national and other data at the European level. Information exchange with other relevant bodies interested in food safety issues (e.g. WHO, FAO, DG Research, EUROSTAT, industry, and other stakeholders) should also be maintained and promoted. An active cooperation with all relevant FP6 projects (e.g. EuroFir, SAFEFOOD, etc.) would also be useful to avoid duplication. The quality of food consumption data can also be increased by identifying the information that can be reported by the consumers in a reliable and cost-effective way. Specifi c data could be captured in separate databases from those collected within food consumption surveys, by means, for example, of special surveys (GMO, radiation, etc). The databases could then be linked to the food consumption data by means of statistical methodologies. This could be for example possible in the case of food brand name. However, missing information on food descriptions (i.e. treatment, packing, etc) are diffi cult to add to existing consumption data. A complete database could be obtained through modelling data from specifi c surveys or by a new pan-European food consumption survey.
When dealing with problems of food classifi cation, it could be useful to consider food regrouping with existing data. National database managers should be able to incorporate several food groups for different purposes.
3.4. The EFSA concise EU food consumption database
The discussion groups supported EFSA’s initiative related to the development of the concise EU food consumption database for preliminary exposure assessment. This was considered as a practical step in a tiered approach. Harmonisation and transparency were highlighted as being very important in the development of the database. It was proposed to quantify the level of conservativeness of this instrument. The concise database should be less conservative than previous screening methods such as the Budget method and should be based on real consumption data that is equivalent to the ones required for the 2nd step in exposure assessment.
22. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
The possible applications of the concise EU food consumption database were discussed and a number of suggestions were proposed. In particular, it was considered necessary that the database is publicly available, despite the fact that this could be a possible barrier for the data owners to provide data. The development of guidelines on how data can be used (and cannot) is fundamental to reassure data owners worried about the possible misuse of the information they provide. The limitations of this database need to be clearly defi ned. For example, if the diet was recorded for only 1 day high percentiles are likely to be overestimated.However, the concise food consumption database should not replace other current efforts for harmonised food consumption data collection. In particular, the use of WHO procedures in all exposure assessments was recommended. Comparison with GEMS / Food regional diet was also recommended in order to perform a sort of benchmarking. It was highlighted as being important to check whether the concise database works in the tiered approach for exposure assessment.
The national database managers need to be involved in the food group identifi cation and content process. A meeting is required with database compilers before the task of completing the database commences. A comprehensive user manual is required for compilers which should include details such as for example: survey characteristics, food groups (e.g. which items are included in each category, including details of country specifi c names). These details must be explicit for the database managers. Data from each survey should be entered separately, without aggregating all national data into one national dataset.National database managers should examine the possibility to disaggregate / group their foods according to different systems and participate in the validation of the database. The database could be improved by giving consideration and advice on the food grouping, in particular some more sub grouping was considered necessary for pesticide risk assessment. As a general suggestion, the EFCOSUM food codifi cation system was proposed as a standard. Moreover, a food consumption survey inventory should be completed starting from the work done within the EFCOSUM project. This inventory should include of all new food consumption surveys, e.g. Germany.It was also suggested to consider the possibility to organise the database according to clusters of countries using as an example the GEMS / Food Consumption Cluster Diets. The choice of age ranges to be considered in the concise database was also tackled. It was proposed to fi rst consider what is currently available.
The implications for medium and long-term strategies
The discussion groups agreed that in the short to medium term the concise database is needed. Feed-back and a report from the Member States and EFSA panels after use of the concise database would be valuable to the development of future strategies.
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 23.
In the medium to long term more refi ned data are required which involves better standardisation of data. The development of a comprehensive database, in addition to the concise database, was proposed. With this prospective, the following points were considered as important:
Data at the level of the individual. �
Raw data and not summarized data. �
National / regional data. �
Data that were previously validated at national level. �
Demographic data on consumers (e.g. age, weight, sex, etc.) should be �included.
For the determinants of food choice, additional demographic information �(ethnic background, country of residence, education, household income, specifi c food preferences, e.g. organic food) are required to inform risk management strategies.
The food descriptors should be more detailed. The following information �should be included: processed or raw at purchase, detailed information on preparation and cooking, brand information and origin of primary products.
The long term needs and strategies were widely discussed. In particular, the incorporation of existing data into a refi ned database with no loss of existing data was recommended. The development of a EU prototype for monitoring studies of food consumption for risk assessment was recommended. This should be a multi-source collection database that includes various types of food consumption data such as individual surveys, household-based surveys (e.g. commercial, HBS, including DAFNE) and market share data.The discussion groups recommended that a Pan-European food consumption study should be carried out. In particular, it was suggested that:
Co-ordination should be performed by EFSA, e.g. within an exposure �assessment unit. The work should be linked to the other important aspects of exposure assessment such as occurrence data. Co-ordination should include regular and consistent reporting.
A stakeholder’s network should be established including the national food �consumption database managers, European Commission and EUROSTAT.
Funding at European level is required. �
Summary of discussion results
24. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
A preliminary inventory of existing European studies (including total diet �studies) should be conducted.
Vulnerable groups should be included and consideration should be given to �the need for biomarkers for some components.
A strategy for the collection and access to food consumption data at the �European level needs to be developed. Details such as sampling, methods and frequency need to be addressed.
With respect to the food consumption surveys already in place, data collection strategies should be adapted to allow a number of uses of the information collected at national, EU and international level. This is the case, for example, of the food consumption data that can be used to address food safety issues as well as nutritional issues for a population.In this context attention must be given to the level of detail needed on the collected data. Accurate descriptions of food are essential to allow it being correctly identifi ed and classifi ed. Different type of information may be required to allow assessment of issues in relation to the different types of substances. For example, information about packaging could be useful in the case of indirect food additives (contact material), brand name for food and feed additives and the geographical origin for contaminants and pesticides.
In general, cooperation was raised as being crucial to future progress at national, EU and international levels with the following points being highlighted:
Full support and co-operation of Member States is essential. �
Co-operation is required with relevant international organisations and �different projects and programmes at national and international level.
All stakeholders need to be involved – industry, consumers, others. �
Close collaboration is required with European statistical systems to �improve information exchange, data collection and harmonisation and avoid duplication.
Consider the collection of data from other national or regional sources. �
EFSA has a role in monitoring ongoing research programmes and utilise �their outputs.
Networking needs to improve to ensure all risk assessors are aware of work �being conducted and to support progress and developments in this area.
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 25.
Summary of discussion results
The role of EFSA in the enactment of many of the proposed strategies was highlighted in relation to a number of areas:
to co-ordinate a Pan European food consumption survey. �
to provide feedback and a report on the use of the concise database on �exposure assessment.
to co-ordinate the use of modelling tools to estimate exposure. �
to work with Member States, Commission and others on medium to long �term strategies.
to monitor ongoing research programmes and utilise their outputs �
to disseminate information on EFSA food consumption database to �EFSA panels.
26. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
IV. CONCLUSIONS & RECOMMENDATIONS
The discussions of the participants led to the agreement that harmonisation of food consumption data was the ultimate requirement in addressing exposure assessment at the European level. It was recommended however that consideration be given to how food consumption data could support other aims of a Member State such as the collection of data for nutrition and public health purposes. The Colloquium was in favour of a Pan-European food consumption survey and recommended that EFSA take a lead role in the co-ordination and completion of associated tasks in meeting this initiative. A number of specifi c recommendations were made in moving towards this ultimate goal.
The Colloquium acknowledged the importance of co-operation and communication 1. in the success of this initiative at national, EU and international levels. Co-operation and communication need to be addressed by means of strategic recommendations on harmonising food consumption data at EU level.
The Colloquium recommended that the inventory currently being conducted on 2. food consumption data and other data available in Member States of the European Union should be completed. This would establish the baseline situation in terms of availability of food consumption data and the type of data available and would help to develop a strategy for access to data.
A European network of national food consumption database managers should be 3. established. Such a network founded with an ethos of collaboration would support managers in their role, support consistency and greater accuracy in the collection of data and support a commitment to the development of a Pan-European food consumption survey.
Comprehensive Guidelines need to be developed to support food consumption 4. database managers in the type of data to be collected.
The Colloquium recommended that work is to be conducted on the use of modelling 5. of exposure for risk assessment and sensitivity analysis to determine the main contributors to exposure.
The Colloquium concluded that recommendations on how to ensure validation of 6. the food consumption surveys needs to be defi ned.
A major proposal of the Colloquium is to develop recommendations and guidance 7. on the description, classifi cation and categorisation of food to be assessed in any food consumption survey.
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 27.
In terms of the 8. EFSA concise food consumption database – the Colloquium supports the completion of this initiative.
In terms of a 9. comprehensive food consumption database – the Colloquium recommends the development of a strategy to develop a more comprehensive food consumption database. The expertise of other Member States in their collection of food consumption data was acknowledged as being fundamental in supporting the development of a comprehensive food consumption database.
28. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 29.
ANNEXES
30. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 31.
V. ANNEXES
Annex 1: Programme of the EFSA Colloquium
Annex 2: Participants at the Colloquium
Annex 3: Presentations made at the Colloquium
Annex 4: Slides of the Discussion groups
Annexes
32. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Annex 1 - Programme of the EFSA Colloquium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 33.
Annex 1: Programme of the EFSA Colloquium
EFSA Scientifi c Colloquium on European Food Consumption Database:Current and medium to long-term strategies28-29 April 2005, Brussels, Belgium
PROGRAMME
Chair: Bo JanssonCo-chair: Ada KnaapRapporteurs: Paul Finglas, Gerhard Heinemeyer
Thursday 28 April 2005
10:00-10:30 Briefi ng meeting with overall chairs and discussion group chairs and rapporteurs
10:30-12:50 Session 1: INTRODUCTORY PLENARY SESSION
10:30-10:50 Introduction to EFSA and to the work of the Exposure group of the Scientifi c Committee
Herman Koëter
10:50-11:10 Food consumption data needs for exposure assessment
Catherine Leclercq
11:10-11:20 Discussion
11:20-11:40 Comparison of food consumption data for some food groups
Wulf Becker
11:40-11:50 Discussion
11:50-12:10 Common methodology for collection of food consumption data food in the EU - Issues identifi ed at the EFCOSUM project
Karin Hulshof
12:10-12:20 Discussion
12:20-12:40 Potential of the standardized EPIC dietary methodology for exposure assessment in Europe
Nadia Slimani
12:40-12:50 Discussion
12:50-13:50 LUNCH
34. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
13:50-14:30 Continuation of session 1
13:50-14:10 Concise EU Food Consumption database,under development by EFSA for preliminary exposure assessment
Philippe Verger
14:10-14:20 Discussion
14:20-14:30 Introduction to discussion groups Juliane Kleiner
14.30-16.00 Session 2: DISCUSSION GROUPS (DG) Four parallel discussion groups to address questions related to:
DG 1 & DG 2
Methodological issues in the use of food consumption data for dietary exposure assessment
DG 1
Chair:Rapporteur:
Max FeinbergWulf Becker
DG 2
Chair:Rapporteur:
Alan ReillyJoseph Shavila
DG 3 & DG 4
Are different data needed for the various areas covered by EFSA, e.g. microbes, pesticides, food contact materials
DG 3
Chair:Rapporteur:
Anders MøllerRuth Charrondière
DG 4
Chair:Rapporteur:
Sam PageArie Havelaar
16:30-17:00 COFFE / TEA BREAK
Annex 1 - Programme of the EFSA Colloquium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 35.
17:00-18:20 Session 3:REPORT BACK FROM DISCUSSION GROUPS TO PLENARY
17:00-17:10 Report back from discussion group 1 Wulf Becker
17:10-17:20 Report back from discussion group 2 Joseph Shavila
17:20-17:40 Discussion
17:40-17:50 Report back from discussion group 3 Ruth Charrondière
17:50-18:00 Report back from discussion group 4 Arie Havelaar
18:00-18:20 Discussion
20:00 DINNER
Friday 29 April 2005
09:00-10:00 Session 4:
PLENARY
09:00-09:20 WHO experience with regional diets Gerry Moy
09:20-09:30 Discussion
09:30-09:50 EUROSTAT activities in the fi eld of
food consumption
Lourdes Llorens Abando
09:50-10:00 Discussion
10:00-13:00 Session 5:
CONTINUATION OF DISCUSSION GROUPS TO DISCUSS
IMPLICATIONS FOR THE CONCISE EU FOOD CONSUMPTION
DATABASE FOR PRELIMINARY RISK ASSESSMENT AND FOR
MEDIUM AND LONG-TERM STRATEGIES
11:00-11:30 COFFE / TEA BREAK
36. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
13:00-14:00 LUNCH
14:00-16:30 Session 6:
REPORT BACK FROM DISCUSSION GROUPS TO
PLENARY AND FINAL DISCUSSION AND CONCLUSION
14:00-14:20 Report back from discussion group 1
and discussion
Wulf Becker
14:20-14:40 Report back from discussion group 2
and discussion
Joseph Shavila
14:40-15:00 Report back from discussion group 3
and discussion
Ruth Charrondière
15:00-15:20 Report back from discussion group 4
and discussion
Arie Havelaar
15:20-16:30 General discussion and conclusions and recommendations
16:30 COLLOQUIUM ADJOURNS
Annex 2 - Participants at the Colloquium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 37.
Annex 2: Participants at the Colloquium
Name Affi liation Country
Discussion
Group
(DG)
Dr. Davide Arcella National Research Institute for Food and Nutrition
IT 1
Prof. Herman Autrup University of Aarhus DK 4
Dr. Martine Bakker National Institute of Public Health and the Environment (RIVM)
NL 1
Dr. Wulf Becker National Food Administration SE 1
Mr. Philippe Becquet FEFANA BE 1
Mr. Sami Belkhira Dow Corning BE 3
Mrs. Maaike Bilau University of Ghent BE 2
Dr. Achim Boenke European Commission BE 3
Mr. Alexander Boss Food Standards Agency UK 4
Mr. Bernard Bottex International Life Sciences Institute (ILSI)
BE 3
Ms. Daniele Bubits Agency for Health and Food Safety
AT 2
Mr. Patrice Buche Institut National de la Recherche Agronomique (INRA)
FR 2
Ms. Irene Bustos European Environmental Citizens Organisation for Standardisation
BE 3
Mrs. Esther Camenzind-Frey
Federal Offi ce of Public Health CH 1
Ms. Ute Ruth Charrondière
Food and Agriculture Organization (FAO)
IT 3
Dr. George Chryssochoidis
Agricultural University of Athens
GR 4
Ms. Rosanna D’Amario European Commission BE 2
Dr. Jan Demyttenaere European Flavour & Fragrance Association
BE 1
38. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Name Affi liation Country
Discussion
Group
(DG)
Mr. Claude Dequatre BP Chemicals FR 2
Mrs. Catherine Dervin Institut National de la Recherche Agronomique (INRA)
FR 1
Mrs. Anna-Maria De Smet
The Brewers of Europe BE 1
Dr. Stefanie De Vriese Institute for Public Health BE 2
Mr. Dan Dils European Flavour & Fragrance Association
BE 2
Mr. Alexander Döring European Feed Manufacturers Federation
BE 3
Dr. Roberto Fanelli Mario Negri Institute IT 2
Dr. Max Feinberg Institut National de la Recherche Agronomique (INRA)
FR 1
Dr. Paul Finglas Institute of Food Research UK 4
Dr. Eva Gelencser Central Food Research Institute HU 1
Mrs. Ieva Gudanaviciene National Nutrition Centre under Ministry of Health
LI 2
Mrs. Anne-Marie Hamelton
PlasticsEurope BE 3
Dr. Andy Hart Central Science Laboratory UK 3
Dr. Arie Havelaar National Institute of Public Health and the Environment (RIVM)
NL 4
Dr. Gerhard Heinemeyer German Federal Institute for Risk Assessment
DE 2
Mrs. Marise Herve Federation of European Food Additives and Food Enzymes Industries
BE 2
Mr. Klaus Hinrichs Product Safety and Regulation DE 4
Dr. Tero Hirnoven National Public Health Institute FI 2
Dr. Karin Hulshof TNO Quality of Life NL 1
Mrs. Inge Huybrechts University of Ghent BE 2
Annex 2 - Participants at the Colloquium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 39.
Name Affi liation Country
Discussion
Group
(DG)
Prof. Bo Jansson University of Stockholm SE 2
Mr. Christian Jassogne CEFIC BE
Dr. Lars Johansson Directorate for Health and Social Affairs
NO 1
Dr. Michael Kaethner Bayer Crop Science DE 4
Mr. Martin Klatt Food and Consumer Product Safety Authority
NL 4
Dr. Ada Knaap National Institute of Public Health and the Environment (RIVM)
NL 3
Prof. Kitty Koelemeijer Nyenrode University NL 1
Dr. Mariella Kuilman DSM Food Specialties NL 1
Mr. Lionel Lafay French Food Safety Agency (AFSSA)
FR 2
Dr. Jean-Charles Leblanc Institut National de la Recherche Agronomique (INRA)
FR 2
Dr. Catherine Leclercq National Research Institute for Food and Nutrition (INRAN)
IT
Dr. Cinzia Le Donne National Research Institute for Food and Nutrition (INRAN)
IT 4
Mrs. Petra Lehner Federal Chamber of Labour AT 4
Dr. Pierre Le Neindre Institut National de la Recherche Agronomique (INRA)
FR 3
Dr. Rene Lesuis Food and Consumer Product Safety Authority
NL 4
Mr. Oliver Lindter Federal Institute for Risk Assessment
DE 4
Dr. Thomas Lyall Freshfel Europe BE 4
Ms. Lourdes Llorens Abando
European Commission BE 2
Dr. Sinéad McCarthy Institute of European Food Studies
IE 1
40. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Name Affi liation Country
Discussion
Group
(DG)
Dr. Ulrich Moser DSM Nutritional Products Europe
CH 2
Dr. Anders Møller Institute for Food and Veterinary Research
DK 3
Dr. Gerald Moy World Health Organization (WHO)
CH 1
Mrs. Dana Müllerova University of Pilsen CZ 2
Dr. Marga Ocke National Institute of Public Health and the Environment
NL 2
Dr. Peter Oldring Valspar UK 2
Dr. Samuel W. Page World Health Organization (WHO)
CH 4
Prof. Kimmo Peltonen National Veterinary and Food Research Institute (EELA)
FI 3
Mr. Pedro Picciochi Food Safety Agency PT 1
Dr. Monique Raats University of Surrey UK 3
Dr. Udo Rabe BASF AG DE 1
Mr. Alan Reilly Food Safety Authority IE 2
Dr. Jiri Ruprich National Institute for Public Health
CH 1
Dr. Dace Santare Food and Veterinary Service LT 3
Dr. Annette Schaefer European Commission BE 4
Dr. Bettina Schmidt-Faber
Federal Offi ce for Consumer Protection and Food Safety
DE 3
Dr. Joseph Shavila Food Standards Agency UK 2
Dr. Derek Shrimpton European Federation of Astions of Health Product Manufacturers
UK 4
Dr. Wolfgang Sichert-Hellert
Research Institute for Child Nutrition
DE 1
Mr. Martin Slayne European Commission BE 4
Dr. Nadia Slimani International Agency for Research on Cancer
FR 1
Annex 2 - Participants at the Colloquium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 41.
Name Affi liation Country
Discussion
Group
(DG)
Dr. David Tennant Food Chemical Risk Analysis UK 1
Mrs. Clara Thompson CIAA BE
Dr. Ariane Titz Association of the European Self-Medication Industry
BE 3
Mrs. Jessica Tressou Institut National de la Recherche Agronomique (INRA)
FR 3
Mrs. Aida Turrini National Institute of Public Health and the Environment (INRAN)
IT 1
Prof. Mieke Uyttendaele University of Ghent BE 3
Mrs. Sirje Vaask Tallinn University of Technology EST 2
Dr. Sandrine Valentin French Food Safety Agency (AFSSA)
FR 4
Dr. Jacob Van Klaveren Institute of Food Safety (RIKILT) NL 3
Prof. Herman Van Oyen Institute of Public Health BE 3
Dr. Philippe Verger Institut National de la Recherche Agronomique (INRA)
FR 3
Mrs. Jeannie Vergnettes European Commission BE
Mr. Jean-Luc Volatier French Food Safety Agency (AFSSA)
FR 1
Mr. Georg Bernhard Zechendorf
European Commission BE 4
Dr. Marcel Zwietering Wageningen University NL 4
42. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
EFSA Staff
Dr. Jan Bloemendal International and Institutional Affairs
Mrs. Marie-Noëlle Costa Scientifi c Committee
Mrs. Anja Friel Panel on Pesticide Risk Assessment Peer Review
Dr. Leng Heng Panel on dietetic products, nutrition and allergies
Dr. Claudia Heppner Panel on contaminants in the food chain
Mrs. Laura Soriano Scientifi c Committee
Dr. Liisa Vahteristo Panel on additives and products or substances used in animal feed
Mrs. Vanessa Descy Scientifi c Expert Services
Dr. Juliane Kleiner Scientifi c Expert Services
Dr. Djien Liem Scientifi c Committee
Ms. Valérie Rolland Scientifi c Committee
Dr. Herman Koëter Director of Science
Ms. Irère Van Geest Communication
Annex 3: Presentations made at the Colloquium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 43.
INTRODUCTION TO EFSA
HERMAN B.W.M. KOËTER
EFSA Deputy Executive Director
and
Director of Science
44. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
EFSA’s Mission and Tasks [Reg 178/2002] :
… provide � scientifi c advice and scientifi c and technical support… [Art. 22. 2]
… shall provide � scientifi c opinions… [Art. 22.6]
… provide the best possible scientifi c opinions in all cases provided for by �Community legislation and on any question within its mission… [Art. 23(a)]
… � collect and analyse data to allow the characterization and monitoring of risks… [Art. 22.4]
… promote and co-ordinate the � development of uniform risk assessment
methodologies [Art. 23(b)]
… commission � scientifi c studies… [Art. 23(d)]
… undertake action to � identify emerging risks… [Art. 23(f)]
Provide � scientifi c and technical assistance with a view to improve co-operation [Art. 23(i)]
EFSA stands for
Independency �
Scientifi c excellence �
Openness and transparency �
Co-operation �
Scientifi c activities (work theme 1) :
Providing scientifi c opinions, guidance and advice in response to questionsMost questions from the Commission, but also from European Parliament �and Member States;
Questions can be broad (e.g. guidance document on allergenicity assessment �of GMO’s) or focused (e.g. risk assessment of SEM in baby food in glass);
Questions have ‘terms of reference’ and deadlines for response. �
Annex 3 – Presentation by H. Koëter
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 45.
Scientifi c activities (work theme 2) :
Assessing the risk of regulated substances and development of proposals for risk-related factors
Chemical categories include : �
Food additives, smoke fl avourings, enzymes, food supplements; �
Food dye stuffs; �
Food contact materials; �
Additives for use in animal nutrition; �
Existing and new Pesticides; �
Genetically modifi ed organisms in food and feed �
Risk related factors include MRL’s �
Deadlines for opinions are strict and often very short (e.g. for GMO’s, feed �additives)
Scientifi c activities (work theme 3) :
Monitoring of specifi c risk factors and diseasesGeographical BSE risk assessment; �
BSE / TSE testing and validation of tests; �
Monitoring of zoonoses and zoonotic agents; �
Containment and eradication of animal diseases (e.g. AI, foot and mouth �disease, rabies)
Scientifi c activities (work theme 4) :
Development, promotion and application of new and harmonized scientifi c approaches and methodologies for hazard and risk assessment of food and feed
Harmonization of detection methodology for chemical and microbiological �contaminants in food / feed;
Harmonized approach for environmental hazard and risk assessment; �
Harmonized approach for human health risk assessment; �
46. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Harmonization of methodology for the monitoring and reporting �of animal diseases;
Exposure assessment modeling and data collection �(chemical and microbiological);
Improving transparency of the risk assessment process. �
Projects initiated by the Scientifi c Committee :Assessment of genotoxic carcinogens; �
Use of benchmark dose levels for assessment; �
Assessment of botanicals and botanical products; �
Applying the QMRA; �
Early warnings of emerging risks; �
Assessment of micro-organisms deliberately introduced in food (QPS); �
Guidance on exposure assessment �
Working Group on Exposure assessment was established in December 2003 �
Members include : 9 members (2 members from the SC and 7 members �from Panels)
Draft opinion was discussed by SC on 28 February – 1 March 2005 �(Guidance Document on human exposure assessment);
A number of changes were recommended on format and presentation of �the GD as well as on the recommendations on how to share responsibilities in this area
It was alos suggested to consider the outcome of the 3rd Colloquium in �the revised draft Guidance Document
Invest in fore-front science through :
Organisation of open scientifi c EFSA meetings, to discuss in-depth topical �and sensitive issues related to EFSA’s mission : EFSA Scientifi c Colloquiums;
Adequate follow-up on EFSA Scientifi c Colloquiums (e.g. development of �Guidance Documents).
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 47.
Annex 3 – Presentation by H. Koëter
The Colloquium is:
An interactive event rather than only a passive listening to lectures; �
A platform for scientists to have in-depth discussions on current and medium �to long-term strategies for food consumption data collection and the development of a comprehensive international database on food consumption;
How to distinguish vulnerable subsets of the population and the diversity �of needs;
A way to build common views and understanding and for EFSA to pick �your brains.
The Colloquium is not:
An attempt to agree on the details of a preferred strategy or approach, if any �
An attempt to fi nalise a blue print for the work ahead of us; �
A “who is right and who is wrong” discussion. �
EFSA Colloquium 1
“Methodologies and Principles for Setting Tolerable Intake Levels for Dioxins, Furans and Dioxin-like PCBs”
18-29 June 2004, Brussels
EFSA Colloquium 2
“Micro-organisms in Food and FeedQualifi ed Presumption of Safety – QPS”
13-14 December 2004, Brussels
48. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 49.
WHICH ARE OUR NEEDS IN TERMS OF FOOD
CONSUMPTION DATA WHEN WE PERFORM
EXPOSURE ASSESSMENT ?
CATHERINE LECLERCQ
Research Group Food Safety-Exposure Analysis
INRAN
National Research Institute on Food and Nutrition
50. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
RISK ANALYSIS
Risk assessment �exposure assessment
Risk management �
Risk communication �
High intake of any substance (natural or artifi cial, nutrient or non nutrient) can constitute a health hazard.
Chemical hazards need to be ranked not only according to their toxicity but according to the level of risk deriving from their presence in the diet.
Heavy metalsToxic substances
Preservatives
lead
cadmium
mercury
dioxine
acrilamidemycotoxins
Allyl caproate
Eucalyptol
Isoa
lyl a
ceta
te
Furfuryl alcohol
Flavourings
Natural toxicants
cyanogenic glycosides
solanine
lectins
canthaxanthin
tartrazine
xanthophylls Colours
nitrite
nitratesodium sulfite
Sweetening agents
sucralose
aspartame alitame
acesulfame K
saccharinantimony trioxide
polyethylene silver zeolite A
Food ContactMaterial
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 51.
Annex 3 – Presentation by C. Leclercq
Exposure is compared with the safety limits established for the various substances present in the diet that may constitute a health hazard for the consumers (risk characterization).
Exposure assessment is therefore crucial in order to discriminate the substances that are likely to pose a safety risk from those which are not.
A safety limit for CHRONIC TOXICITY: the Acceptable Daily Intake (ADI)
“The amount of a food additive, expressed on a mg/kg body weight basis, that can be ingested daily over a lifetime without incurring any appreciable health risk, and is based on an evaluation of available toxicological data”
Scientifi c Commitee for Food, 1991.
Intake must be assessed as mean intake within a long period of time: CHRONIC EXPOSURE
A safety limit for ACUTE TOXICITY:the Acute Reference Dose (expressed per kg body weight)
Exposure must be assessed within a restricted time interval (meal or day): ACUTE EXPOSURE
Dietary exposure =
Sum (Food chemical concentration x Food consumption)
body weight
52. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Impact of body weight on exposure
CYCLAMMATE ADI: 7 mg/kg body weightMaximum Permitted Level in sugar free beverages: 400 mg/l
30 kg
ADI reached through daily consumption of 0.5 litre of beverage
50 kg
ADI reached through daily consumption of 0.9 litre of beverage
60 kg
ADI reached through daily consumption of 1 litre of beverage
Age Body weight Average energy requirement
Child (3-4 years) 15 kg 94 kcal/kgChild (9-10 years) 30 kg 68 kcal/kgSedentary Adult 60 kg 35 kcal/kg
European RDAs (Nutrient and energy intakes for the European Community, 1993)
The ratio of food consumption / body weight is higher in children than in adults. If the chemical concentration is the same in their diet, exposure is higher in children.
In order to protect the health of all consumers there is the need to ensure the safety of the diet of the whole population, including high consumers (for hazardous chemicals) and low consumers (for nutrients) of the different food items.
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 53.
Annex 3 – Presentation by C. Leclercq
We need to estimate the highest percentiles
of mean daily intake/kg body weight
Example of a food chemical intake distribution
The issue of what proportion of the population would have to exceed the ADI before action (i.e. risk management decision) is considered necessary to reduce exposure is:
a � scientifi c question: (sample size needed to assess the 95°, 97.5°, 99° percentiles of exposure)
a � political / ethical question (can we exclude 5%, 2.5%, 1% of the European population from the risk assessment?)
The number of chemical substances that should be monitored is huge
For example:
The food additives authorized in the EU are more than 300 �
More than 2800 different fl avouring substances are claimed by industry �to be added to foods and beverages in Europe
95thMean
90th
99th
High percentiles
54. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Exposure varies with AGE and PHYSIOLOGICAL STATUS
Working group on exposure assessment
European Food Safety AuthoritySCIENTIFIC COMMITTEE
Panel on food
additives,
fl avourings,
processing aids
and materials in
contact with food
Panel on plant
health, plant
protection
products and their
residues
Panel on additives
and products or
substances used
in animal feed
Panel on
genetically
modifi ed
organisms
Panel on dietetic
products,
nutrition and
allergies
Panel on
contaminants in
the food chain
Panel on
biological hazards
(including
TSE / BSE issues)
Panel on animal
health and
welfare
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 55.
Annex 3 – Presentation by C. Leclercq
Exposure varies with FOOD CHOICE
Exposure varies with GEOGRAPHICAL AREA
56. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Stepwise procedure:
In order to target detailed dietary exposure estimates to chemicals which might be of health concern for the general population or for certain critical groups.
At step 1, exposure is fi rst assessed by using methods based on conservative
assumptions. If the conclusion is “no safety concern” no further assessment is performed.
Source: REPORT OF THE JOINT FAO / WHO CONSULTATION ON FOOD CONSUMPTION AND EXPOSURE ASSESSMENT OF CHEMICALS, Geneva, 10-14 February 1997.
At the fi rst step of chronic exposure, hypothetical food consumption data are
used. The techniques are easy to use and pragmatic but they differ according to
the category of substance.
They need to be validated:
Are they really conservative? Which percentile of exposure do they assess?
They could be harmonized:
In order to harmonize the level of protection of consumers for the different categories of substances.
Source: REPORT OF THE JOINT FAO / WHO CONSULTATION ON FOOD CONSUMPTION AND EXPOSURE
Improved Consumption Data Quality
Improved Residue Data Quality
Best
EstimateMethods Used For Estimating Exposure
Model Diets
Regional Diets
National Diets
Estimate
First
Maximum Levels in Standards
Monitored Levels
As Consumed Levels
Household andIndividual Diets
Single Point Method Probabilistic Method
Least quality,greatest uncertainty,
and least cost.
Best quality,least uncertainty,and greatest cost
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 57.
Annex 3 – Presentation by C. Leclercq
Food consumption data are needed to check for the underlying assumptions
of step 1 methods
Residues of food contact materials: a 60 kg subject may consume 1 kg of food /
beverage in contact with 600 cm2 of the relevant food contact material -> need to assess the ratio of contact surface to mass in packaged food available on the market.
Flavourings: each single fl avouring is consumed by 32 millions of consumers in Europe (“per capita” method) -> need to assess the market of fl avoured food and their distribution among foods.
If exposure at the fi rst step is high, more accurate assessments of exposure are needed, based on actual food consumption data.
The study of exposure from diet must take into special consideration individuals who consume relatively large quantities of foods containing higher concentrations of the food chemical of interest.
Food consumption data needs in terms of data quality:
recent �
individual level �
representative of the population �
adequate dimension of the sample �
covering the whole diet (foods eaten out of home, supplements, drinking �water)
referred to short term and long term consumption �
with individual body weights �
codifi ed into highly disaggregated food categories: �
brand (additives and fl avourings) �
type of packaging (residues of packaging material) �
cooking method (toxic substances related to processing) �
58. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Food consumption data needs in terms of data availability:
In order to perform refi ned exposure assessment to single substances we need
easy access to raw data (food consumption for each single subject) �
database manager (to assign chemical concentration to the correct �food groups)
Food consumption data available at EU level for exposure assessment are highly variable:
Household Budget Surveys �
Food consumption surveys aimed at assessing nutrient intake (dietary �history, Food Frequency Questionnaire, 24 h recall, 48 h recall, 2 day record, 3 day record, 7 day record)
14 day consumer surveys at brand level �
How can we use incomplete data sets?
By making conservative assumptions:
total brand loyalty of subjects to the food items containing the highest �concentration of substance can be assumed
low standard body weight (60 kg in the adult) can be used instead of �measured body weight
databank of large portion sizes may be used. �
By correcting data:
chronic exposure can be derived from short term surveys �
individual consumption can be derived from household consumption �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 59.
Annex 3 – Presentation by C. Leclercq
Need to account for uncertainty
“uncertainties… should be explicitly considered at each step in the risk assessment and documented in a transparent manner. Expression of uncertainty… may be qualitative or quantitative, but should be quantifi ed to the extent that is scientifi cally achievable”
Codex Working Principles on Risk Analysis, 2003
Uncertainties in food consumption data
Measurement uncertainties �
food types, weights and portion sizes �
under-reporting and recall error �
Sampling uncertainty �
when survey is biased or too small to assess extreme consumers �
Extrapolation uncertainty �
extrapolation to different years, seasons, regions, population groups etc. �
extrapolation from short to long term exposure �
A good survey design allows to reduce uncertainty
Characterising uncertainty in exposure assessment
Qualitative: �
list sources of uncertainty affecting the assessment �
evaluate direction and magnitude of each �
discuss combined impact on exposure estimate �
Quantitative: �
uncertainty factors & worst case assumptions �
sensitivity analysis �
probabilistic analysis, e.g. Monte Carlo or Bayesian methods �
60. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium60. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 61.
Annex 3 – Presentation by C. Leclercq
The EFSA Journal (2004)46
The EFSA Journal (2004)46
62. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
The EFSA Journal (2004)46
Which are our needs in terms
of food consumption data
when we perform
exposure assessment ?
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 63.
FOOD CONSUMPTION IN EUROPE –
DATA FROM DIFFERENT SOURCES.
IMPLICATIONS FOR EXPOSURE ASSESSMENTS
W. BECKER
National Food Administration
Uppsala, Sweden
64. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Data sources on food consumption
FAO Food Balance Sheets (FBS) �
Food availability / market supply �
Household Budget Surveys �
Food brought into the household �
Food consumption / dietary surveys �
Food and drink eaten during a specifi ed period �
FAO Food Balance Sheets (FBS)
Food availability / market supply �
not � purchase or eating
primary, raw commodities �
Regularly updated �
Inherent unsystematic differences vis-à-vis national data �
Primarily used for trends �
Used in exposure assessment �
Household Budget Surveys
Food and drink brought into the household �
not � eating
May cover only expenditures �
Meals outside home may not be included or recorded as expenditures �
Household composition varies �
Used in exposure assessment �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 65.
Annex 3 – Presentation by W. Becker
Food consumption surveys
Food and drink � consumed by individuals during a specifi c period by individuals
Various methodolgies affect results �
Record, Interview, Food Frequency Questionnaires �
Level of detail of food coverage and description �
Population sampled �
Mis- and / or under-reporting is common �
Nationally representative surveys not available in all countries �
Data available in Europe
FBS: FAO, Eurostat, (OECD) �
HBS: all countries (?), DAFNE-project �
Food Consumption Surveys �
national surveys in several countries, with different methodologies �
multi-centre studies: SENECA, EPIC �
National food consumption studies used
Name Age Method
N 1997 Norkost 1997 16-79 yr FFQ
SE 1997-98 Riksmaten 18-74 yr 7-d rec
DK 2000/01 Danskernes kostvaner 15-75 yr 7-d rec
D 1998 GeNuS 18-79 yr DH
UK 2000-01 NDNS 19-64 yr 7-d rec
NL 1997-98 DFCS 1-65+ yr 48-h rec
F 1999 INCA 15-65+ yr 7-d rec
66. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
EPIC Collaborating Centres
Availability of milk, FAO FBS 2000
66 Summary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium
N SE DK D UK NL B-Lx F P E GR I0
100
200
300
400
500
600
700
800
900
1000
1100
ml/p
ers/
year
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 67.
Annex 3 – Presentation by W. Becker
Milk consumption in Dafne countries, 1990
Consumption of milk and yoghurt, EPIC
0
50
100
150
200
250
300
350
400
Norway
Sweden
Denm
ark
Germ
any
UK NL
Fran
ce
Spain
Greec
eIta
ly
g/d
ay
women
men
68. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Milk and cheese consumption among elderly, SENECA
Consumption of milk according to EPIC and national FCS
68 Summary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium
0
100
200
300
400
500
DK/Rosk
ilde
NL/Culem
borg
B/Ham
me
F/Hag
F/Rom
ans
P/Coim
braP/V
F
E/Bet
anzo
s
I/Pad
ua
g/d
ay
women
men
0
50
100
150
200
250
300
350
400
Norway Sweden Denmark Germany UK NL France
g/d
ay
EPIC
National FCS
women
womenincl.
cheese
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 69.
Annex 3 – Presentation by W. Becker
Availability of vegetables and fruit, FAO FBS 2000
Total fruit and vegetable consumption, EPIC
Summary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium 69
0
200
400
600
800
1000
1200
1400
N SE DK D UK NL B-Lx F P E GR I
g/p
ers/
year
Fruit
Veg
0
100
200
300
400
500
600
700
Norway
Sweden
Denm
ark
Germ
any NL
UK
Fran
ce
Spain
Greec
eIta
ly
g/d
ay
women
men
70. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Fruit and vegetable consumption among elderly, SENECA
Consumption of fruit & vegetables in EPIC and national FCS
70 Summary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium
0
100
200
300
400
500
600
700
800
900
1000
DK/Rosk
ilde
NL/Culem
borg
B/Ham
me
F/Hag
F/Rom
ans
P/Coim
braP/V
F
E/Bet
anzo
s
I/Pad
ua
g/d
ay
F&G, men
F&V, women
0
100
200
300
400
500
600
Norway Sweden Denmark Germany NL UK France
g/d
ay
EPIC
National FCS
women
women
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 71.
Annex 3 – Presentation by W. Becker
Availability of fi sh (round fi sh) in some European countries, FAO FBS - 2000
Fish consumption in EPIC study
Summary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium
0
50
100
150
200
250
N SE DK D UK NL B-Lx F P E GR I
g/p
ers/
year
0
20
40
60
80
100
Norway
Sweden
Denm
ark
Germ
any
UK NL
Fran
ce
Spain
Greec
eIta
ly
g/d
ay
women
men
72. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Fish consumption among elderly, SENECA
Fish consumption according to EPIC and national FCS
72 Summary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium
0
20
40
60
80
100
120
DK/Rosk
ilde
NL/Culem
borg
B/Ham
me
F/Hag
F/Rom
ans
P/Coim
braP/V
F
E/Bet
anzo
s
I/Pad
ua
g/d
ay
women men
0
10
20
30
40
50
60
70
Norway Sweden Denmark Germany UK NL France
g/d
ay
EPIC
National FCSwomen
women
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 73.
Annex 3 – Presentation by W. Becker
Conclusions
Different data sources may give different pictures of food consumption �patterns in a country
Food patterns differ between males and females, age groups etc., but not �consistently across countries
Conclusions: EA
Use of food consumption data from different sources in EA may produce �highly variable estimates
Available food consumption data at the individual level are in many cases not �directly comparable due to methodological issues
Use of available food consumption data in EA should be explored �
Harmonised data collections are needed to enable consistent EA across �Europe
Some References
FAO Food balance sheets for year 2000. FAOSTAT, www.fao.org. �
Anonymous (1998): Zo eet Nederland. Resultaten van de Voedselconsumptie- �peiling 1997-1998. Den Haag: Voedingscentrum.
Becker W , Pearson M. Dietary habits and nutrient intake in Sweden 1997-98. �National Food Administration, Uppsala 2002.
Fagt S, Matthiessen J, Trolle E, Lyhne N et al. (2002): 2. Dietary habits of the �Danish Population 2000-2001. Fødevarerapport 2002:10. Søborg: The Danish Food Directorate.
National Diet & Nutrition Survey 2000-2001. Volume 1. Types and quantities �of food consumed. HMSO 2002.
Johansson L, Sovoll K (1999): Norkost 1997. National dietary survey among �men and women 16-79 years of age (In Norwegian). Report 3/1999. Oslo, National Council on Nutrition and Physical Activity.
74. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Mensink GBM, Thamm M, Haas K (1999): Die Ernährung in Deutschland 1998. �Gesundheitswesen 61, S201-S206.
Volatier JL et al. (2000): Enquête INCA individuelle et nationale �sur les consommations alimentaires. Technique & Documentation, Londres / Paris / New York.
Seneca investigators. Food patterns of elderly Europeans. �Eur J Clin Nutr 1996; 50 (Suppl. 2): S86-S100.
EPIC calibration study, Public Health Nutr. 2002 Dec;5(6B) �
DAFNE Initiative. Public Health Nutr 2001; 4(5B) �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 75.
KARIN HULSHOF
TNO - Quality of Life
COMMON METHODOLOGY FOR COLLECTION
OF FOOD CONSUMPTION DATA IN THE EU.
ISSUES IDENTIFIED AT THE EFCOSUM PROJECT
76. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
EFCOSUM project 2000-2002
European Food Consumption Survey Method (EFCOSUM)
Framework of EU Programme on Health Monitoring
Findings and recommendations published in EJCN 2002;56, suppl.2
Participating European countries
EFCOSUM - starting points
23 participating European countries �
co-operation with EPIC, DAFNE �
building on European actions on food classifi cation and food composition �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 77.
Annex 3 – Presentation by K. Hulshof
Aims of EFCOSUM
to defi ne a method for monitoring food consumption in nationally �representative samples of all age-sex categories in Europe in a comparable way
to indicate how to make existing food consumption data comparable �
to defi ne a (minimum) set of dietary components which are relevant �indicators of health
International comparable data on
consumption of foods as well as nutrients to get insight into average intake �and their distribution in well-defi ned groups of individuals
usual as well as acute intake �
existing as well as new data �
Relevant dietary indicators* based on EURODIET
vegetables, excl. potatoes �
fruit, incl. fresh, excl. bottled / canned juices �
bread �
fi sh, shellfi sh �
saturated fatty acids (%energy) �
total fat (total lipids) (%energy) �
alcohol (g/day) �
biomarkers for folate, iron, vitamin D, iodine, sodium �
* very minimum set
78. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Inventarisation of available food consumption data at the individual level
most European countries have carried out national dietary surveys which �provide valuable information for usage in national nutrition policy and nutritional surveillance
period 1985-2001: 21 European countries - 45 surveys: �huge differences in population groups, age categories, duration of the study and dietary methods
e.g. 23 dietary records (16x 7d; 5x 2-4 d; 2x 1d)
9 dietary recalls (7x 1d; 1x 2d; 1x 3d)
7 dietary history
6 food frequency questionnaire
Harmonisation of existing data
pragmatic guidelines were developed to permit available data to be made �comparable in a transparent way:
population groups - adults
age of survey - 5-10 years
type of survey - 1 day
food categorization - EFG (COST action 99)
� about 15 countries could provide food consumption data that were available and can be made reasonably comparable at the level of food groups (‘raw edible ingredient level’) for the whole adult population
for comparability at the nutrient level it was recommended to wait for a �European nutrient database
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 79.
Annex 3 – Presentation by K. Hulshof
Method for comparable new data
individual data using 24-hour recall and biomarkers �
2 repeated, non-consecutive (computerised) 24-hour recalls, over 1 year, �including all week days
quantifi cation of portion sizes with picture book and additional data �
strictly standardised procedures with (upgraded version of) EPIC-SOFT �as software of choice (if EPIC-SOFT cannot be used at least the rules for standardisation and data control should be used)
FFQ if necessary to obtain % non-users �
use of modelling techniques �
Motivation 24-hour recall method
applicable for broad populations of different ethnicity �
no literacy requirement �
relatively low burden for respondent as well as interviewer �
information possible on meal occasions, place and indoor / outdoor consumption �
open-ended �
procedure does not alter food intake pattern �
administration time is short �
method is cost-effective �
80. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
24-hour recall and intake calculations
suffi ces to get population mean intakes and distributions for observed intake �
allows determination of usual intakes and distributions when using at least �two non-consecutive 24-hour recalls and a food list to assess the proportion non-users for infrequently consumed food
Comparability of food classifi cation systems
level of comparability �
food commodities, ingredients, as consumed �
grouping systems �
EPIC, DAFNE, EFG �
recommendation: coding at ‘raw edible ingredient �level’ EFG system
observedusual
Intake
norm
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 81.
Annex 3 – Presentation by K. Hulshof
Comparability of food composition databases
National food composition databases are not comparable, �not even on macro nutrient level
European Food Information Resource (EuroFIR) is started in 2005. �Principal objective is:“to build and disseminate a comprehensive, coherent and validated databank providing a single, authoritative source of food composition in Europe for nutrients, and newly emerging bioactive compounds with putative health benefi ts”
Operationalisation of a European Food Consumption Surveillance
In EFCOSUM consensus also recommendations have been set up for:
data collection software �
sampling procedures �
recruitment �
fi eld work procedures �
biomarkers �
training, quality control �
Statistical aspects
minimum sample size ~ 2000 adults per country �
2x 24-h recall; FFQ for % users infrequently consumed food �
mean, median, quartiles, P5 and P95 �
82. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Overall conclusions EFCOSUM
available datasets on dietary intake are not comparable on European level �
existing data from 15 countries can be made comparable on the level of �food groups
fi rst choice: a new Pan-European food consumption surveillance on �individual level using 2x 24-hour recall
second choice: 24-hour recall for calibration �
How to proceed?
several European countries adopted the EFCOSUM �approach in new studies
further development and validation of a �trans-European method to be used for estimation of the intake of foods, nutrients and potentially hazardous chemicals within the European population including children
include further developments within EuroFIR �(food grouping, food composition database)
resources… �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 83.
Annex 3 – Presentation by K. Hulshof
Participants
Prof dr JA Abaravicius, LV Dr MRH Löwik, NL
Prof JA Amorim Cruz, PT Dr A Møller, DK
Dr W Becker, SW Prof O Moreiras, ES
Prof dr G Biró, HU Dr L Ovesen, DK
Dr H Boeing, DE Dr J Ruprich, CZ
Ms HAM Brants, NL Dr W Sekula, PL
Dr JH Brussaard, NL Dr N Slimani / Dr UR Charrondière, FR
Dr S De Henauw, BE Ms G Smithers, UK
Sr AMJ van Erp, NL Dr L Steingrímsdóttir, IS
Dr K Holciková, SV Dr A Trichopoulou, GR
Dr KFAM Hulshof, NL Dr A Turrini, IT
Dr S Jacob, CH Dr LM Valsta, SF
Dr L Johansson, NO Dr Ph Verger, FR
Dr A Kaic-Rak, HR Dr JL Volatier, FR
Dr J Kearney, IE Dr DC Welten, NL
Dr JS Koenig, AT
84. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 85.
N. SLIMANI
Nutrition and Hormone Group
International Agency for Research on Cancer (IARC-WHO)
Lyon, France
POTENTIAL OF THE STANDARDIZED EPIC
DIETARY METHODOLOGY FOR EXPOSURE
ASSESSMENT IN EUROPE
86. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
European Prospective Investigation into Cancer and Nutrition (EPIC)
10 European Countries �
23 Centres �
Over 520 000 participants �
Ranging in age from 35-75 years. �
Tromsø
Umeå Malmö
Aarhus Copenhagen
Oxford Cambridge
Potsdam Heidelberg
Utrecht Bilthoven
Paris (nationwide)
Turin Milan Florence Naples Ragusa
Oviedo San Sebastian Pamplona Murcia Granada
Athens (nationwide)
ö
Umeå
(Iarc) Lyon
Paris
Florence
Milan
Ragusa
Turin
NaplesBarcelona
Oviedo
Granada
Murcia
Pamplona
San sebastian
Cambridge
Oxford
Athens
Heidelberg
Potsdam
Aarhus
Copenhagen
Tromsø
Bilthoven
Utrecht
Malm
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 87.
Annex 3 – Presentation by N. Slimani
Ex
tern
al a
ctiv
itie
s
EPIC-SOFT asReference?
Special issue of EJCN, 2002
Concept andversions
development
10 EPIC-SOFTversions
Special issue of PHN, 2002 EPIC Nutrient
DataBase (ENDB)
Analyses on the EPIC-SOFT data
Validation of the EPIC-SOFT
measurements
Field work(n=37,000)
EUROFIRNew BelgiumVersion,
Upgrade of theDutch one
EFCOSUM
Several regional and national surveys
1992 1995 2000 2002 2004 2005
Inte
rnal
act
ivit
ies
Main features of EPIC-SOFT
Standardised procedures across countries to � describe, quantify, probe and calculate food and beverage intakes (incl. food supplements)
11 translated versions with the same interface adapted to each country �
Calculate food quantities in a common basis �
Systematic quality controls before, during and after data collection �
Standardized procedures to classify, store, retrieve and export data �
Standardized procedures to update EPIC-SOFT databases (58 fi les) �
Harmonized nutrient databases (ENDB), outside EPIC-SOFT �
88. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Potentials of EPIC-SOFT satisfying needs for exposure assessment (summarized from C. Leclerq, EFCOSUM, ILSI)
Standardised open-ended dietary methodology on individuals tested in �different study contexts and populations (within / outside EPIC)
Detailed and fl exible system of food description and quantifi cation which can �potentially be adapted to specifi c exposures
Information on � meal occasions, time and place and related frequency of consumption (e.g. for investigating foods taken out of home, meal patterns)
Validated measurements (evaluated at population level in EPIC) �
Detailed and standardised data exports for analyses ( � raw standardized data which can according to different exposures or hypotheses of interest)
Flexible system to upgrade individual databases and prepare new country- or �study-specifi c EPIC-SOFT versions
Nutritional studies using EPIC-SOFT within or outside EPI
See table following pagewing pagSee table follow
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 89.
Annex 3 – Presentation by N. Slimani
Co
un
try
(da
tes)
Stu
dy n
am
es
(in
sti
tute
in
vo
lve
d)
Ty
pe o
f
stu
dy
Su
bje
cts
Inte
rvie
w
me
tho
d
No
of
24
-HD
Rs
Pa
rtic
ipa
tio
n
Ra
tes (*
)
EP
IC(1
99
5-2
00
0)
EP
IC (
23 c
entr
es)
Cal
ibra
tio
n37
,00
0 a
du
lts
Face
-to
-fac
eE
xcep
t N
orw
ay1
57%
to
92
%(h
alf
ce
ntr
es
>8
0%
)
Sp
ain
(20
01-2
002
)S
pan
ish
bla
dd
er a
nd
ca
nce
r st
ud
yV
alid
atio
n21
7, a
du
lts
Face
-to
-fac
e7-
972
% (
60
%)
No
rway
(20
02-2
003
)N
orw
egia
n W
om
en a
nd
C
ance
r st
ud
yV
alid
atio
n28
6, w
om
enTe
lep
ho
ne
49
9%
(5
7%
)
Ger
man
y(2
002
-20
03)
Bav
aria
n f
oo
d c
on
sum
pti
on
su
rvey
IIR
egio
nal
surv
ey10
50, 1
4-8
0 y
Tele
ph
on
e3
85%
(7
1%
)
(At
leas
t 2
x 24
-HD
Rs)
Sp
ain
(20
03-2
00
4)
Bas
qu
e co
un
try
surv
ey a
mo
ng
ch
ildre
n a
nd
ad
ole
scen
tsR
egio
nal
surv
ey10
00,
ch
ildre
n
and
ad
ole
scen
tsFa
ce-t
o-f
ace
2?
Net
her
lan
ds
(20
03)
Pilo
t st
ud
y o
f th
e D
utc
h
nat
ion
al c
on
sum
pti
on
su
rvey
Pilo
t st
ud
y75
0, y
ou
ng
adu
lts
Tele
ph
on
e2
95%
(4
2%
)
Bel
giu
m(2
00
4-2
005
)B
elg
ium
Fo
od
co
nsu
mp
tio
n s
urv
eyN
atio
nal
surv
ey32
00,
15
and
o
lder
Face
-to
-fac
e2
? (4
2%
)
Net
her
lan
ds
(20
05-2
006
)T
he
Du
tch
Co
nsu
mp
tio
n
Su
rvey
in C
hild
ren
Nat
ion
alsu
rvey
128
0, 2
-6 y
2-d
ay p
re-s
tru
ctu
red
dia
ry(E
PIC
-SO
FT
as
dat
a en
try
syst
em)
In p
rep
arat
ion
Ger
man
y(2
005
-20
06 ?
)G
erm
an N
atio
n-w
ide
Foo
d
Co
nsu
mp
tio
n S
urv
ey (
NV
S II
)N
atio
nal
surv
ey20
,00
0, 1
4-8
0 y
Tele
ph
on
e2
In p
rep
arat
ion
(*) :
Par
tici
pat
ion
rate
s to
the
EPIC
-SO
FT 2
4-H
DR
inte
rvie
w (
s) (
gre
y fi g
ure
s), a
nd
over
all p
arti
cip
atio
n ra
tes
(ora
ng
e fi g
ure
s)
90. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
What can we learn from these different study designs?
EPIC-SOFT can be used in different geographical regions, study designs �and on representative samples
Relatively high participation rates to the dietary interviews (57% to 99%), �despite different study designs, populations and number of repeated measurement and overall participation
Possibility to collect single or repeated 24-HDRs using alternative �approaches, i.e. face-to-face and telephone, data entry (?)
Potential to evaluate the reliability of measurements in other study groups �and contexts then EPIC
Short-time interview (~30-35 minutes) �
Importance of these studies as benefi cial experiences for improving EPIC-SOFT and possible study design(s) for pan-European monitoring surveys to be further investigated
Some selected aspects of the EPIC-SOFT programme or measurements
The EPIC-SOFT food description system �
The EPIC-SOFT food quantifi cation system �
Some insights on the validity of the EPIC-SOFT measurements �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 91.
Annex 3 – Presentation by N. Slimani
Food / recipe search and identifi cation
Food description: facets / descriptors approach
92. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Food quantifi cation
Selected photo from the EPIC-SOFT picture book
EPIC-SOFT picture book containing 140 common photo series…
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 93.
Annex 3 – Presentation by N. Slimani
Final collected 24-HDR interview
94. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Flexibility of the food description system…
PRESERVATION METHOD
NEW DESCRIPTORS ADDED !!!
ADD NEW FOOD !!!
NEW PHOTOS ADDED !!! 3000 to 13,000 food + facet / descriptor strings
PHYSICAL STATE
COOKING METHOD
FACETS / DESCRIPTORSFOOD LIST
- Fish n.s.- Anchovy- Bass, sea- Carp .- Salmon
- Undefi ned- Canned- Frozen - Fresh . .
- Undefi ned- Fillet- Whole- Small pieces- Slice / steak .
- Cooked n.s.- Fried- Boiled.- Steamed
.
.
.- Griddled fl at,
over heat
.
DONELESS
Veal thin (photo 1) Veal thin (photo 2)..Veal thick (photo x)
NEW ADDED FACET !!!
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 95.
Annex 3 – Presentation by N. Slimani
EPIC-SOFT facets / descriptors and (methodological) food classifi cation system
Facets (No of descriptors) Food groups (No sub / sub-groups)
• Source (21) • Unclassifi ed (0)
• Physical state (28) • Potato and other tubers (3)
• Cooking method (28) • Vegetables (10)
• Preservation method (14) • Legumes ( (14)
• Packing medium (24) • Fruits (5)
• Flavoured / added components (33) • Dairy products (12)
• Sugar content (5) • Cereals and products (10)
• Fat content (29) • Meat and products (19)
• Type of packing (4) • Fish and shellfi sh (4)
• Food production (12) • Eggs and products (2)
• Enriched / fortifi ed (6) • Fat (7)
• Brand name / product name (open list) • Sugar and confectionary (10)
• Skin consumed (3) • Cakes (3)
• Visible fat consumed (3) • Non-alcohol beverages (10)
• Type of fat used (open list) • Alcoholic beverages (8)
• Type of milk / liquid used (12) • Condiments and sauces (10)
• Soups, bouillon (3)
• Miscellaneous (6)
Standardisation and (relative) validity of the EPIC-SOFT at population level (n=37,000)
This has been evaluated by three main ways:
1. Level of standardisation of dietary measurements across the interviewers �(Slimani et al. 2000 )
2. Degree of under- or over-estimation by comparing mean total energy �intake and EI:BMR and its confi dence intervals across centres (Ferrari et al., 2002)
3. Using independent biomarkers such as urinary nitrogen �(Slimani et al. 2003 ) or plasma carotenoids (Al-Delaimy, submitted)
96. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Ad
juste
d c
en
tre
me
an
en
erg
y i
nta
ke
s i
n t
he
24
-HD
Rs,
aft
er
exclu
sio
n: W
om
en
mmary Report EFSA Scientific Colloquium 3 28-29 April 2005 - Brussels Belgium
Ferr
ari
et a
l., 2
002
1500
1600
1700
1800
1900
2000
2100
2200
Greec
eAstu
rias Gra
nada M
urcia Nav
arra
San S
ebas
tian
Flore
nce Vares
e RagusaTu
rin Naples Fr
a: North
-Eas
t
Fra:
North-W
est
Fra:
South
Fra:
South co
ast Heid
elber
gPo
tsdam
Bilthove
n Utrech
t
UK (G) UK (H)
Aarhus
Copenhag
en
Malm
ö Umea
Nor: South
& E
ast
Nor: North
& W
est
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 97.
Annex 3 – Presentation by N. Slimani
We
igh
ted
Pe
ars
on
’s c
orr
ela
tio
n b
etw
ee
n g
en
de
r-sp
ecifi
c c
en
tre
alg
ori
thm
ic m
ea
ns
of
24
-h u
rin
ary
N a
nd
me
an
N f
rom
24
-HD
Rs (
N=
22
)
Summary Report EFSA Scientific Colloquium 3, 28-29 April 2005 - Brussels, Belgium 97.
Slim
ani e
t al
., 20
03
Cam
Hei
Oxf
Pot
Gre
Flo
Bil
Tur
Var
Rag
Bil
Cam
Flo
HeiPa
r
Nap
Oxf
Pot
Rag
Tur
Var
Gre
1.8
22.2
2.4
2.6
2.8
3
1.8
22.
22.
42.
62.
83
Loga
rith
mic
mea
n N
uri
ne
Logarithmic mean N 24-HDRs
Y =
1.2
9x -
0.9
7
r =
0.8
6
ra =
0.94
Mal
e
Fem
ale
98. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Sp
ea
rma
n ’s c
orr
ela
tio
n b
etw
ee
n m
ea
n 2
4-H
DR
s a
nd
DQ
s f
ruit
an
d
ve
ge
tab
le i
nta
ke
s a
nd
me
an
pla
sm
a c
aro
ten
oid
leve
ls (
n=
16
)
Lute
inZ
eaxa
nth
inB
eta
Cry
pto
xan
thin
Lyco
pen
eA
lph
aca
rote
ne
Bet
aca
rote
ne
Tota
lca
rote
no
ids
Fru
its
& V
eg.
0.8
4
0.6
8
0.8
7
0.6
8
0.8
3
0.6
7
0.6
7
0.5
6
n.s
.n
.s.
n.s
.n
.s.
0.7
1
0.5
6
Veg
etab
les
0.69
n.s
.
0.68
n.s
.
0.61
n.s
.
0.54
n.s
.
n.s
.n
.s.
n.s
.n
.s.
0.64
n.s
.
Fru
its
0.8
0
0.7
6
0.8
7
0.7
1
0.8
4
0.6
8
0.64
0.6
4
n.s
.n
.s.
n.s
.
n.s
.
0.6
7
0.6
1
Cit
rus
fru
its
0.58
0.6
9
0.67
0.6
4
0.9
0
0.7
0
n.s
.0
.61
-0.6
7n
.s.
n.s
.n
.s.
n.s
.n
.s.
Tota
l car
rot
-0.6
4n
.s.
-0.7
8 n
.s.
-0.8
0n
.s.
n.s
.n
.s.
0.7
1
0.6
1
n.s
.
n.s
.
n.s
.
n.s
.
All
tom
ato
coo
ked
0.71
n.s
.
0.53
n.s
.
0.52
n.s
.
0.7
8
0.6
9
n.s
.n
.s.
n.s
.n
.s.
0.74
n.s
.
: S
tan
dar
diz
ed 2
4-H
DR
s :
No
n-s
tan
dar
diz
ed E
PIC
DQ
s
Al d
elai
my
et a
l., s
ub
mit
ted
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 99.
Annex 3 – Presentation by N. Slimani
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 99.
Current limitations
EPIC-SOFT versions are not available for all EU countries �
EPIC-SOFT and its databases need to be adapted to different �population groups (e.g. children, adolescents, ethnic groups)
In Children, EPIC-SOFT might be used as data entry system � feasibility study on going in NL (Ocké & Hulshof)
EPIC-SOFT measurements need to be validated for the specifi c �needs of risk exposure assessment
EPIC-SOFT needs to be reprogrammed on a modern window environment �
Additional EPIC-SOFT modules need to be developed for an easier �maintenance of the EPIC-SOFT databases (58 fi les per version)
Future needs for risk exposure assessment
Evaluate the (specifi c) needs for exposure assessments precisely �and the possibilities of upgrading EPIC-SOFT databases (and structure)
Upgrade EPIC-SOFT and its databases according to comments from �EPIC-SOFT end- users and EU recommendations (EFCOSUM) to improve further the programme and the interview procedures
Upgrade the EPIC-SOFT databases according to changes on the food �market or food habits (e.g. foods, recipes, food portions, food supplements, brand names, etc…)
Develop other EPIC-SOFT versions and upgrade / adapt the �software / databases to other countries / different population groups
Validate further EPIC-SOFT according to designs relevant for risk �exposure assessments
Find resources… �
100. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 101.
PHILIPPE VERGER
Research unit INRA 1204 – Met@risk
CONCISE E.U. FOOD CONSUMPTION DATABASE
UNDER DEVELOPMENT BY EFSA FOR
PRELIMINARY EXPOSURE ASSESSMENT
102. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Practical defi nition for international dietary exposure assessment
Accumulation of arguments from national data to estimate the “real” �exposure of the general population: in general not based on one fi gure but on consistent results from various origin.
Needs for EFSA panels
Risk assessment needs international food consumption data �
A vast majority of the EFSA opinions’ do not need a �sophisticated assessment
Protecting countries without any data (reasonable overestimation) �
Acceptability of the opinions by stakeholders �
Few opinions need an accurate international estimation of the intake �
Current situationFood consumption data are available in a majority of European countries �
BUT
The comparison of food consumption obtained at national level can not �be done directly
Main reasons for observed differences
Various survey methodologies �
Various sampling methodologies �
Various survey duration �
Various clustering of age groups �
Various categorisation systems �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 103.
Annex 3 – Presentation by Ph. Verger
Data currently used by international committees
GEMS Food diets (5 diets) �
SCOOP data (Scientifi c cooperation tasks) �
Task 4.1 specifi c for food intake data �
Other tasks (not always consistent) �
Possible improvement: 2 ways
“Disintegration” of the GEMS Food diets (13 diets or more) �
“Combination” of national data �
Same age groups �
Same food categorisation �
Adults and children
Two age groups �
3 to 14 years old �
15 and over �
104. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
European Food Grouping (EFG)
• Main categories • Sub categories
• Cereals & cereal products • Fruit juices
• Sugar & sugar products • Soft drinks, excl fruit juice
• Fats (vegetable and animal) • Bottled water
• Vegetables, nuts and pulses
• Starchy roots or potatoes • Beer
• Fruits • Wine
• Fruit juices, soft drinks and bottled water • Other alcoholic beverages
• Coffee, tea, cocoa
• Alcoholic beverages • Meat and meat products
• Meat and meat products, offal • Offal
• Fish and seafood • Seafood
• Eggs • Fish
• Milk
• Cheese and other milk products • Control of consistency
• Tap water • All solid foods
• All beverages
• Food non classifi ed
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 105.
Annex 3 – Presentation by Ph. Verger
Preliminary results: fi sh consumption in Europe
Data from 6 countries �
Italy �
France �
UK �
Sweden �
Norway �
Netherlands �
Total fi sh consumption (all fi shes including those not identifi ed �by the consumer)
Average consumption, high percentiles and portion size �for consumers only.
Percentage of consumers �
Mean intake (consumers only)
Sweden: 29 ± 21 g �
Netherlands: 34 ± 25 g �
France: 36 ± 29 g �
UK: 43 ± 36 g �
Italy: 48 ± 41 g �
Norway: 77 g (FFQ with standard portion) �
GEMS Food Europe (5 diets): 47 g �
GEMS Food (13 diets): �
Cluster B: 41 g (Italy) �
Cluster E: 25 g (France & UK) �
Cluster F: 45 g (Sweden) �
106. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
High percentiles (95th or 99th)
Sweden: 113 g �
Netherlands: 123 g �
UK: 136 g �
France: 141 g �
Norway: 174 g �
Italy: 190 g �
GEMS Food Europe + WHO procedure: 3 * 47 g = 141 g �
Average portion size
Netherlands: 99 g �
Sweden: 114 g �
France: 118 g �
Italy: 129 g �
Percentage of consumers for fi sh
Netherlands*: 15 % �
France: 78 % �
Italy: 84 % �
Sweden: 98 % �* Two days
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 107.
Annex 3 – Presentation by Ph. Verger
Fis
h c
on
su
mp
tio
n b
y s
pe
cie
s
Co
nsu
mp
tio
nM
ean
FR
g/d
ay
Mea
n
ITg
/day
Mea
n
NL
g/d
ay
Mea
n
NW
g/d
ay
Mea
n
SW
g/d
ay
Mea
n
UK
g/d
ay
99th
FR
g/d
ay
99th
IT
g/d
ay
99th
N
Lg
/day
95th
N
Wg
/day
99th
S
Wg
/day
97.5
th
UK
g/d
ay
Sal
mo
n15
1419
85
1963
6065
1427
66
Her
rin
g11
2030
35
1942
6694
730
53
Tun
a11
1514
83
1631
6155
3127
52
Tro
ut
3230
05
-22
5587
014
-53
Car
p0
80
00
-0
80
00
-
Pilc
har
d-
Sar
din
a10
220
00
1547
700
00
62
Mac
kere
l13
2516
40
1754
245
4010
047
Oth
er fi
shes
3143
33-
18-
9919
010
1-
75-
No
n s
pec
ifi e
d
fi sh
29-
39-
--
129
-10
7-
--
Tota
l Fis
h36
48
3477
2943
141
190
123
174
113
136
108. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Provisional conclusions
Data on average intake, high percentiles and average portion size �for consumers only are consistent at that level of aggregation for the considered countries independently of the survey methodology
GEMS Diets do not represents an overestimate for fi sh consumption �
Short term improvements
Comparable estimation of frequency of consumption �
Comparable estimation of percentage of consumers �
Table on large portions �
Estimation of high intake for chemicals in more than one food category �(SCOOP 4.2)
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 109.
Annex 3 – Presentation by Ph. Verger
Step forward
Collection of data in the EU by EFSA �
Registration in an EFSA AMI list �(see EFSA web under Opportunities / Procurements)
Calls for tender for each country addressed to eligible experts / Institutes �
Validation procedure for 3 countries �
Test assumption high consumer of max. 2 categories �
Preliminary comparison with GEMS Food regional diets �
Test applicability through different compounds �
EFSA staff contract 3 months (up to 18 months) �
Initial screening of intake �
Accessibility to EFSA Panels / SC for preliminary exposure assessment as �well as publication on EFSA web
Further development �
More comprehensive database (food categories, subpopulations etc.) �
Creation of a network of database managers �
Starting point for a network with food consumption database managers �in the EU
EFSA contact : [email protected]
110. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 111.
GERALD G. MOY
Food Safety Department
World Health Organization
WHO EXPERIENCE WITH
FOOD CONSUMPTION DATA
112. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 113.
Annex 3 – Presentation by G. Moy
Risk Analysis Paradigm
SPS Agreement
Agreement on the Application of Sanitary and Phytosanitary Measures �
All sanitary measures for food shall be based on risk assessment (Article 5.1) �
Risk assessment shall take into account “risk assessment techniques �developed by the relevant international organizations”
Risk Assessment
* Science basedRisk Management
* Policy based
Risk Communication
* Interactive exchange of information and opinions
114. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Exposure Assessment
The qualitative and / or quantitative evaluation of the likely intake of biological, chemical, and physical agents via food as well as exposures from other sources if relevant.
Codex, 1997
Exposure Assessment (single food)
Dietary Exposure = C x F
C = Concentration of food chemicalF = Amount of food consumed
Hazard Characterization Exposure Assessment
Risk Characterization
Hazard Identifi cation
Risk Assessment Process
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 115.
Annex 3 – Presentation by G. Moy
Exposure Assessment (multiple foods)
Dietary Exposure = Σ Ci x F
i
Ci = Concentration of food chemical in a specifi c foodFi = Amount of that food consumed
Potential exposure scenarios
Toxic Concern Exposure Route
single chemical single food
single chemical multiple foods
single chemical multiple media
multiple chemicals with the same mechanism of action (toxicity)
single food
multiple chemicals with the same mechanism of action (toxicity)
multiple foods
multiple chemicals with the same mechanism of action (toxicity)
multiple media
116. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Exposure
Assessment
TARGET GROUP / BODY WEIGHT- Foetus- Infant- Children mean or
standard deviation- Adults mean or
standard deviation
FOOD CHEMICAL DATA- Maximum
permitted- Highest reported- Trial mean or
median- Monitoring data- Processing factors
applied- As consumed
FOOD CONSUMPTION(including drinking water)- Highest reported- High percentile- Mean (eaters only)- Mean population
OTHER FACTORS- Nutritional status,
esp. malnutrition- Occupation- Health status- Age- Gender- Genetics- Multimedia
HAZARD CHARACTERIZATION- ADI- PTWI / PTDI- Potency- Acute Reference
Dose
EXPOSURE TIME- Lifetime- Year- Month- Week- Day- One meal
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 117.
Annex 3 – Presentation by G. Moy
Hazard Characterization
Acceptable Daily Intake �
Provisional Tolerable Daily Intake or Weekly Intake �
Carcinogenic Potency �
Acute Reference Dose �
Target Group Body Weight
Foetus �
Infant �
Children �
Adults �
Vulnerable groups �
Exposure Time
Lifetime �
Year �
Month �
Week �
Day �
One meal �
Other Factors
Nutritional status, �esp. malnutrition
Gender �
Genetics �
Multimedia �
Other �
Occupation �
Health status �
Age �
118. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Food Chemical Data
Maximum permitted �
Highest reported �
Mean or median �
Monitoring data �
Processing factors �
As consumed �
Food Consumption
Highest reported �
High percentile �
Mean or median (eaters) �
Mean or median (population) �
Consumption Data
Population methods �
Food balance sheets �
Per capita production �
Household methods �
Purchase records �
Food disappearance �
Individual methods �
Food record survey �
24-hour recall survey �
Food frequency questionnaire �
Meal-based diet history survey �
Food habit questionnaire �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 119.
Annex 3 – Presentation by G. Moy
Food Chain Considerations
Raw agricultural commodities �
Semi-processed commodities �
Processed foods �
Foods as consumed �
Single commodity �
Mixed commodity �
GEMS / Food Regional Diets
Raw and semi-processed agricultural commodities �
Use FAO Food Balance Sheets �
Per capita daily intake �
Grams per person per day �
120. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Eg
yp
tia
n F
oo
d B
ala
nce
Sh
ee
t -
19
98
PE
R C
AP
ITA
PR
OD
UC
TP
RO
DU
CT
ION
IMP
OR
TS
EX
PO
RT
ST
OTA
L S
UP
PLY
FEE
DS
EE
DW
AS
TE
SU
PP
LY (
G/D
AY
)
Wh
eat
6093
,15
7413
,33
7,8
1225
8,6
811
61,5
924
512
91,0
539
7,0
Mai
ze54
30,4
930
54,5
1,04
8183
,96
3593
,77
2974
0,39
151,
1
So
rgh
um
900
0,2
080
0,2
450
571
,48
11,4
Po
tato
es19
84,
0160
,18
233,
918
10,7
919
820
3,22
58,4
Su
gar
Can
e14
352,
783,
3314
349,
4514
3514
467
,3
Su
gar
(R
aw E
qu
ival
ent)
1171
1101
,09
1,57
2032
,47
84,
4
Pu
lses
, Oth
er56
1,6
142,
075,
2967
8,38
113,
4527
35,6
920
,8
Tom
ato
es57
53,2
86,
8721
,357
38,8
557
5,33
214,
4
Veg
etab
les,
Oth
er57
09,4
52,
6290
,99
5662
,95
0,35
566,
4621
1,6
Ora
ng
es, M
and
arin
es18
63,4
60,
9421
7,76
164
6,6
418
6,35
60,6
Lem
on
s, L
imes
300
012
,49
287,
5130
10,7
Ban
anas
600
11,6
0,01
611,
661
,16
22,8
Ap
ple
s38
8,46
38,9
90,
0842
7,37
42,7
116
,0
Dat
es83
9,8
0,23
0,67
839,
368
431
,4
Gra
pes
957,
7314
,61
0,86
971,
49
95,7
836
,2
Fru
its,
Oth
er10
47,6
516
,53
26,7
510
37,4
310
0,05
38,9
Bo
vin
e M
eat
483
,02
130,
90,
5663
8,41
26,5
Po
ult
ry M
eat
522,
560,
060,
3252
2,3
21,7
Fres
hw
ater
Fis
h27
6,62
0,22
027
6,8
411
,5
Pel
agic
Fis
h78
,07
230,
190
308,
2531
,58
11,5
Ric
e (M
illed
Eq
uiv
alen
t)29
68,
310,
68
417,
6930
50,3
690
,04
83,
3818
0,39
110,
8
Su
gar
& S
wee
ten
ers
1250
,22
1106
,81
5,07
2113
,91
87,7
Pu
lses
588,
8516
4,07
9,45
723,
4711
3,45
33,4
536
22,4
Veg
etab
le O
ils11
1,8
724,
057,
46
814,
2917
,3
Milk
- E
xclu
din
g B
utt
er34
81,2
632
2,24
8,4
636
09,4
957
7,79
174,
0611
8,5
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 121.
Annex 3 – Presentation by G. Moy
Five GEMS / Food Regional Diets
Middle Eastern �
Far Eastern �
African �
Latin American �
European-type �
Risk Assessment Bodies Using GEMS / Food Diets
Joint FAO / WHO Expert Committee on Food Additives (JECFA) �
Joint FAO / WHO Expert Meetings on Pesticide Residues (JMPR) �
Ad hoc Study Groups and Consultations �
Corn Consumption by GEMS / Food Region (g/person/day)
Middle
Eastern
Far Eastern
African
Latin
American
European
0
20
40
60
80
100
120
122. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Problems with the GEMS / Food Regional Diets
Codex and FAO codes do not match �
Some commodities missing �
Some processed food missing �
Too old �
Not representative �
Does not include high percentile consumers �
GEMS / Food Consumption Cluster Diets
FAO / WHO Exposure Assessment Consultation, February 1997, Geneva �
Peterson and Barraj paper �
Use average 1997-2001 FBS �
9 original clusters �
13 proposed consumption cluster diets �
Proposed Revised GEMS / Food Regional / Cultural Dietary Clusters
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 123.
Annex 3 – Presentation by G. Moy
Corn Consumption (g/person/day)
GEMS / Food Cluster H
250 versus 40
Bolivia � Mexico �
El Salvador � Nicaragua �
Fiji � Panama �
Guatemala � Paraguay �
Haiti � Peru �
Honduras � Saint Kitts & Nevis �
St Vincent & Grenadine �
0
50
100
150
200
250
A B C D E F G H I J K L M
Proposed Revised GEMS / Food Region
124. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
GEMS / Food Cluster I
250 versus 100
Botswana � South Africa �
Cape Verde � Swaziland �
Ghana � Togo �
Kenya � Tanzania �
Lesotho � Zambia �
Malawi � Zimbabwe �
Mozambique �
Namibia �
Codex Committees
Codex Committee on Pesticide Residues �
Uses GEMS / Food-based exposure estimates provided by JMPR �
Presented with proposed new diets in 2005 �
Will issue Circular Letter requesting missing data �
Will consider again with worked examples in 2006 �
Codex Committee on Food Additives and Contaminants �
Uses GEMS / Food-based exposure estimates for contaminants �provided by JECFA
Uses GEMS / Food-based exposure assessments to decide �which commodities maximum levels will be considered
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 125.
Annex 3 – Presentation by G. Moy
97.5 Percentile Consumption of Corn Products (g/person/day)
97.5 Percentile Consumption of Corn Products (g/kg bw/day)
0
50
100
150
200
250
300
350
400
Sweet Corn Corn Flour Corn Popcorn
Children
General
0
2
4
6
8
10
12
Sweet Corn Corn Flour Corn Popcorn
Children
General
126. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 127.
LOURDES LLORENS-ABANDO
Eurostat
EUROSTAT WORK ON
FOOD CONSUMPTION STATISTICS
Conclusions of the Task Force meeting and the way forward
128. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Contents
Background �
Mandates �
Needs for food consumption statistics �
Main Statistical sources �
Priorities �
Working plan �
Background
The Food Safety statistics working group decided to include food �consumption as one of the priority themes
In 2003 the working group approved to carry out a Task Force on Food �consumption in 2005
Food consumption data collection is included in the EUROSTAT working �programme for 2005
The Food Safety working group approved the launching of the Food �Consumption Task Force and the work programme for it in December 2004
Ten countries are working as volunteers in this Task Force �
Mandates
General goal �
Examine the needs for food safety statistics and analyse the possibility �to establish a regular data collection on food consumption within ESS
Specifi c tasks �
Analyse needs �
Analyse statistical sources �
Establish priorities �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 129.
Annex 3 – Presentation by L. LLorens-Abando
Basic needs (1)
The basic needs for food consumption data are based on: �
Commission and other bodies monitoring and assessment of �the policy making
Evaluation and drafting of legislation and actions at the EU level �
The main themes are: �
Food market management �
Health and nutrition �
Food related diseases �
Food safety �
Environmental issues �
Basic needs (2)
Market management �
Main objectives related to economic objectives �
The data will be used to make economic analysis to evaluate the �Community Agricultural Policy
It requires variables linked to economic and social indicators. �
Main variables required �
Retailers’ sales -
Purchases by households -
Food availability -
Prices -
Organic farming products -
It does not require individual real food intake
130. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Basic needs (3)
Health and nutrition Diet related diseases and Food Safety �
Objectives �
Analysis of food consumption patterns and their statistical distribution -
It requires the estimation of derived variables: calorie intake, micro and -macro-nutrients, etc.
Main variables �
Real food intake -
Proxy variables �
Retailers’ sales in volume -
Purchases by households -
Food availability -
Although it can be proxies the individual real food intake is an essential variable
Basic needs (4)
Environment �
Objective: to analyse the infl uence of consumption patterns on �the environment
Consumption of organic products -
Consumption of GMOs -
Waste in the distribution Chain -
Consumer’s needs �
The consumer must be taken into consideration when collecting food -consumption data.
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 131.
Annex 3 – Presentation by L. LLorens-Abando
Basic needs: summary
Needs Main variable Proxies
Market management • Retailers’ sales (volume and value)
• Households’ Purchases(volume and value)
• Food availability• Food prices • Consumption of
products with distinctive marks
Health and nutrition • Individual food intake Sales in volumePurchases in volume
Diet related diseases • Individual food intake Sales in volumePurchases in volume
Food safety • Individual food intake Sales in volumePurchases in volumeRaw availability
Environmental issues • Waste• Consumption of organic
and GMO products
Basic needs (5)
Needs for data updating �
Health, Nutrition and Food Safety: updating of the main variable �every fi ve years
Market management : updating of the economic related variables yearly �
Food items classifi cation �
Health, Nutrition and Food Safety: EFG, proposed by EFCOSUM �
Market Management: more stages within the �production / consumption chain
For some cases special disaggregation might be needed �
132. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Basic needs (6)
Needs for consumption determinants �
Household related �
Geographical -
Type of settlement -
Education -
Individual related �
Age -
Sex -
Education… -
Needs for population segmentation �
Population at greater risk �
Children -
Pregnant women -
Elderly people -
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 133.
Annex 3 – Presentation by L. LLorens-Abando
Main statistical sources
Dietary surveys �
Household Budget surveys and other Household surveys �
Supply balance sheets �
Main statistical sources (1)
Dietary surveys �
Advantages �
Data on individual food intake -
Allows the estimation of the statistical distribution -
Give very detailed information about the individual characteristics -
Problems �
Frequency is very scarce -
They are very expensive -
The data collection techniques are different and they do not produce -harmonised results
The data must be collected by very specialised interviewers -
Harmonisation is diffi cult since every country wants to keep time -comparability
The informant burden is enormous -
134. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Main statistical sources (2)
Household budget surveys �
Advantages �
They are regularly collected -
They give information about household purchases and this uses a -harmonised classifi cation COICOP-HBS
Give very good information about household characteristics and -composition
Problems �
Not individual data. It must be estimated -
Some countries do not collect data about volume -
There is a lack of data of consumption out of home -
There is a risk to lose the food volume data in the future -
Main statistical sources (3)
Supply balance sheets �
Advantages �
They are annually collected -
They use a harmonised classifi cation -
There is an international comparison -
Problems �
They give only food availability -
There are some doubts about their quality -
There is a risk to lose data on some of the items because the agricultural -statistics have a negative priorities
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 135.
Annex 3 – Presentation by L. LLorens-Abando
Priorities (1)
Commission services need data on food consumption harmonised. �
The Member State would use these data to benchmark their own policies �and improve data collection strategies
The main priority is to design a strategy to collect Food consumption �statistics within the ESS
To keep the data already available �
To develop a mixed data collection strategy �
Priorities (2)
The mixed data collection would mean: �
Supply Balance Sheets: to keep annual data availability �
Household Surveys: to collect data on food purchases in a systematic �and harmonised way
Dietary surveys. There are several initiatives to harmonise �the data collection techniques and food items classifi cations. It is crucial to follow with work already done
The food classifi cations and the consumption determinants
should be harmonised
Working plan 2005
EUROSTAT and the experts of the Task Force will produce a document �on “ESS strategy on food consumption statistics”
The strategy will be approved by the Working Group on Food Safety statistics �
EUROSTAT will launch grants to develop a harmonised way to collect food �consumption statistics at EU level
EUROSTAT will follow an open procedure where everybody
willing to collaborate is welcome
136. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 137.
JULIANE KLEINER
EFSA
INSTRUCTIONS FOR DISCUSSION GROUPS
(DG)
138. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Organisational Details
4 parallel discussion groups �
14:30-16:30 DG 1st round – specifi c aspects �
17:00-18:20 Report back to Plenary + discussions �
10:00-13:00 DG 2nd round – implications �
14:00-16:30 Final Plenary session – conclusion and recommendations �
Meeting package
Hand-outs of presentations �
Briefs for discussion groups �
Description of concise EU food consumption database currently under �development by EFSA
Additional background information �
The DAFNE initiative provided by Antonia Trichopoulou �
Considerations on methodological issues for food consumption data �by Max Feinberg
Discussion group 1 + 2 Methodological issues in the use of food consumption data
for dietary EA
What are minimum quality criteria for a food survey to be used in EA �
Is there a pragmatic approach to deal with high percentiles �
How to deal with limitations of data �
How to address consumer loyalty to particular food items �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 139.
Annex 3 - Instructions for discussion groups
Discussion group 3 + 4 Data needs for the various areas to be covered by EFSA
General needs for EA �
Specifi c needs for EA in the various areas �
Which food consumption data are available for the specifi c purposes �
Possible ways for improvement of data quality �
Discussion groups 2 nd day
What are the implications for the � concise EU food consumption database (EFSA)
What are the implications for medium and long-term strategies �for the development of a comprehensive food consumption database
Gaps and uncertainties �
Ways for cooperation with other national, European and international bodies �
Conclusions and recommendations �
It is not the aim of the colloquium to reiterate the achievements from the various projects but to concentrate on the issues which can help to
build up a concise food consumption database as a fi rst step and a comprehensive food consumption database at a later stage.
After the Colloquium
Draft summary report of colloquium to be prepared by rapporteurs �
1 � st review by DG chairs and rapporteurs
Review of revised draft by all participants �
Publication of summary report and power point presentations on EFSA �website (10/05) and in EFSA Science Colloquium Report Series
140. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 141.
DISCUSSION GROUP 1
METHODOLOGICAL ISSUES IN THE USE OF
FOOD CONSUMPTION DATA FOR DIETARY EA
DAY 2 - RESULTS
Annex 4: Slides of discussion groups
142. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Minimum quality criteria
Depends on aim: acute vs chronic exposure, screening vs refi ned EA �
Develop guidelines on minimum set of data to be collected in food �consumption surveys
Representative and large enough for target population �
Allow calculation of high percentiles and uncertainty of measurements, �or distribution, of relevant food items / categories
Food coverage �
all foods consumed necessary for the EA, incl. drinking water, supplements �
Study design adequate to capture, where necessary, variations in �consumption due to week-days and season
Cover relevant aspects of foods and food consumption circumstances �in terms of of the components that are assessed
descriptors (e.g. Langual) for processing, handling, preparation, �packaging etc.
meal situation �
Consider the reliability of a sample population rather than the accuracy of �the recorded amounts
Develop ”Quality control index” for existing / new surveys �
High percentiles and Limitations of data
Mis / Under-reporting �
Multiple informants / instruments �
Field work / training: probing of food intake �
Additional information on subjects: slimming, physical activity data, �state of health
Check how it affects EA, e.g. high percentiles �
Annex 4 – Slides of Discussion Groups
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 143.
Long-term exposure from short-term measurements �
Combination of different dietary assessment methods, �e.g. FFQ for foods that are consumed more rarely
Statistical approaches (Nusser) �
Probabilistic methods should be explored and validated �
Relationship between the risk assessment study and the quality of �consumption data. Who decides?
BW: use register data or conservative mean �
Consumer loyalty to food items
Deterministic, worst case, estimates �Modelling using data from other sources, e.g. market share data, consumer panels, HBSCollect data on brand names, organic labelling
Concise database
Support concise DB as a step in a tiered approach. �Comprehensive user manual need for compilers, e.g. details on survey characteristics, data collection, food group specifi c comments (which items are included in each category, incl. country specifi c names).Meeting with compilers before completing DB to be organised.
Age categories / ranges: choice? Depends on what is available �How to report BW, e.g. for an age range of 3-14 yr?
How conservative should we be? �Proposed concise DB less conservative than previous screening methods, e.g. Danish budget method. Based on real consumption data, 2nd level EA. Should not replace current methods.
Transparency important, e.g. for third countries and international bodies �
Clustering of countries in DB? GEMS �
Validation important if it works in the tiered approach. Compare with GEMS. �
Always use WHO procedures in EA �
144. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Long term needs / strategies �
Multi-source collection DB including various types of food consumption data �
individual surveys -
household-based (commercial, HBS, incl. DAFNE) -
market share data -
A Pan-European food consumption study should be carried out. �Establish a network of the DB managers, COM, Eurostat and other stakeholders for this purpose.Coordination by EFSA, e.g. within an exposure assessment unit. Link to work on occurrence data. Regular and consistent reporting.Funding at European level needed. Cover vulnerable groups.Biomarkers needed for some components. Inventory of existing European studies.Total diet studies.
Draft a strategy for collection and access to food consumption data at the �European level. Sampling, methods, frequency.
Adapt data collection strategies to cover various substances: food safety �aspects should be considered in addition to nutrients. Food identifi cation and description important.
Indirect food additives (contact material): packaging �
Food and feed additives: brand name �
Contaminants: geographical origin �
Pesticides: geographical origin, organically produced �
Veterinary drugs �
Nutrients �
Bioactive substances �
Naturally occurring toxins �
Micro-organisms �
Annex 4 – Slides of Discussion Groups
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 145.
Feed-back from MS and EFSA Panels after use of concise DB: �Evaluation report
Encourage research in how to estimate chronic high exposure from �short-term measurements
Develop a EU prototype monitoring studies on food consumption for RA �
individual based �
146. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 147.
DISCUSSION GROUP 2
METHODOLOGICAL ISSUES IN THE USE OF
FOOD CONSUMPTIONDATA FOR DIETARY EA
DAY 2 - RESULTS
148. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Implications of a concise European Food Consumption Database?
Screening step to allow preliminary exposure analyses �(1st step in EA) – tends to give excessive estimate
Rapid exposure assessment (refi ned exposure assessment 2 � nd step)
A more detailed exposure assessment may still be required (3 � rd step)
Most practical way forward �
EFCOSUM codifi cation of foods – aggregation of categories �
Harmonisation �
To enable correlation between detailed MS surveys where they exist and �other MS to make the exposure assessment pan-European
Information on total diet study to be incorporated �
Implication for medium / long-term strategies for the development of a comprehensive food consumption database
Short – medium term concise �
Medium – long term more refi ned requiring better standardisation of data �
EFSA to co-ordinate the use of modeling tools to estimate exposure �
EFSA to work with MS, Commission and others on medium – long term strategy �
Incorporation of existing data into refi ned database with no loss �of existing data
Annex 4 – Slides of Discussion Groups
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 149.
Cooperation: national, EU and international
Full support and co-operation of MS essential �
Co-operation with relevant international organisations �
Involvement of all stakeholders – industry, consumers, others �
Close collaboration with European statistical system to improve exchange �of information, data collection and harmonisation – avoid duplication
Collection of data from other national or regional sources �
EFSA to monitor ongoing relevant research programmes and utilise outputs �
Improve networking among ‘experts’ �
Conclusion and recommendations
EFSA to consider establishing a task force to evaluate the use of modeling �of exposure for risk assessment and sensitivity analysis to determine main contributors to exposure
EFSA needs to communicate their strategy to develop the proposed concise �database to MS and get their support
EFSA needs to communicate relevance of exposure assessment and its �relationship to risk to Member States and consumers
Use of EFSA extranet for rapid compilation of data from MS �
EFSA to develop guidelines for MS database managers �
In the development of more refi ned database, EFSA should consider �how exposure could be used to establish a threshold approach on a case by case basis
EFSA to develop structure / content of any refi ned database with MS �and others
EFSA to request submission of possible food categories and other relevant �parameters in order to develop a harmonised structure for a refi ned database
Food consumption data for children and the amalgamation of existing data �should be a priority for EFSA
In the development of concise database EFSA should include methods of �food preparation etc.
150. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 151.
DISCUSSION GROUP 3
DATA NEEDS FOR THE VARIOUS AREAS
TO BE COVERED BY EFSA
DAY 2 - RESULTS
152. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Questions
What are the general needs for exposure assessment?1.
What are the specifi c needs for exposure assessment in the various areas?2.
Which food consumption data are available for the specifi c purposes?3.
What are possible ways for improvement of data quality?4.
What are the implications for the concise EU food consumption database 5. under development by EFSA for preliminary exposure assessment?
What are the implications for medium and long-term strategies?6.
For which components in risk assessment would we need improvements in 7. the concise EFSA food consumption database and which ones?
Data availability and submission
Requirement: �
individual data, �
national / regional (not necessarily representative), �
enter each survey separately (without aggregating all national data into �one national dataset)
previously validated at national level �
Complete FCS data owners, not present at this colloquium. EFSA should �contact additional data owners so that they register at AMI (not only passively put information on EFSA website and EFSA channels but active contacting)
Complete (newer and older) FCS inventory put together by EFCOSUM. �Examples of new FCS: Germany, regional Spanish, Ireland, Iceland, Italy, Belgium, Netherlands, France, UK, Norway, Finland, Sweden, Denmark, EPIC
Complete tables of the concise food consumption database �
FCS data owners should participate in the validation of the data �
Annex 4 – Slides of Discussion Groups
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 153.
Barrier for data owner to provide data: �
that publicly available (with possible misuse) �
newer data might not be submitted (but as very aggregated this risk might �be diminished)
Network of FCS compilers / owners under EFSA umbrella
Establish network for �
exchange of information, �
validation, �
provide data �
Include national and other data owner �
European level �
Active cooperation with FP6 projects (e.g. EuroFir, SAFEFOOD) to �avoid duplication
Information exchange incl. WHO / FAO, DG RESEARCH, EUROSTAT, industry, �other stakeholders
Use of the concise FCS database
Publicly available �
Provide guidelines of how data can be used (and how not) �
Database will have limitations, e.g. high percentiles of 1 day recall are �overestimated and of 3-7 days are underestimated
Possible improvements for the concise FCS database
The actual food groups are diffi cult to use for pesticide risk assessment. �Some more sub grouping would be necessary. National database managers should examine the possibility to disaggregate / group their foods according different systems. Good if EFSA would ask for that.
Age / sex distribution? �
154. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Medium / long-term strategies propositions
Refi ne FCS percentiles for acute and chronic exposure assessment �
Validation of some assumptions of exposure models though network, �e.g. food additives
Evaluation and validation of models for food contact materials �
Use of mean body weight in screening phase in exposure assessment �(not sure how this would be used)
Deal with uncertainties, e.g. underreporting of certain population groups �(e.g. senior)
Investigate future collaboration of the network with other partners in exposure �assessment (e.g. DG SANCO, DG INDUSTRY, industrial associations)
Encourage harmonised data collection �
Extend study populations, e.g. ethnic groups �
FCS missing for some regions (e.g. Southern Europe, Baltic countries), �for population groups (e.g. infants, children, senior / older people)
Recommendations
Establish the EFSA concise food consumption database and stay �focused on it
Get commitment from EFSA to maintain EFSA concise food �consumption database
Establish strategy to develop a more comprehensive food �consumption database
Clarifi cation of defi nition of terms used �
Wide use of database beyond exposure assessment, e.g. nutrition �
Improve communication and advocacy �
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 155.
DISCUSSION GROUP 4
DATA NEEDS FOR THE VARIOUS AREAS
TO BE COVERED BY EFSA
DAY 2 - RESULTS
156. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
European concise food consumption database
Possible use and suggested additions �
Demographic data on consumers: age, weight, sex, … �
Nutrition
May be adequate for screening purposes �
Missing information fast-food, fries so may underestimate calorie intake �
Food additives – functional
May be adequate for screening whether information on processed or not �at purchase is added
Food additives –non- functional (incl. food contact materials)
May be adequate for fi rst level screening if information is added on package. �Preferably also information on type of packaging material (glass, plastic, paper, cans).
Chemical contaminants
May be adequate for screening level risk assessment of environmental �contaminants if combined with occurrence data.
Not useful for regulated chemicals (e.g. pesticides, veterinary drugs) �or substances with very specifi c use within a major food category (e.g. only corn). Food categories do not match those in regulations.
Micro-organisms
Not useful unless some categories are disaggregated. In particular meat �consumption should be broken down to at least the level of animal species.
Also needs information on heat treatment during industrial processing �and preparation.
N.B. This may imply that it is not necessary to add information on number of eating occasions.
Annex 4 – Slides of Discussion Groups
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 157.
European comprehensive food consumption and behaviour database
General requirements
Needed for risk assessment
Needs level of detail in individual food consumption surveys �
Need full data instead of summary data �
Data should be available in more disaggregated food categories �for all domains
Processed or raw at purchase �
Detailed information on preparation and cooking �
Meal composition and frequency and weight per meal �
Brand information and origin of primary products can be applied in different �domains, but should be rather specifi c (e.g. diet coke vs. regular coke). May be obtained via barcode information but may miss local sales. Traceability barcodes are being developed.
Determinants of food choice to inform risk management strategies
Additional demographic information: ethnic background, country of residence, education, household income, specifi c food preferences, e.g. organic food, …
Nutrition
Fortifi cation, may be proprietary information
Nutritional supplements
Food additives – functional
Brand information is key
158. Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium
Food additives - non-functional (incl. food contact materials)
Food contact material before / after purchase
Surface to volume ratio
Chemical contaminants
Weighed records are essential for acute exposure assessment of pesticides to comply with existing regulations
Micro-organisms
Storage after purchase (time-temperature, sell-by or best before date) and after preparation
Can valid information on hygiene be obtained from questionnaires?
Conclusions and recommendations
Concise database
Concise database is a useful fi rst step towards harmonized European food �consumption data.
Concise database is useful for screening purposes, in particular for �macronutrients and environmental contaminants.
Current concise database does not cover the needs of microbiological �risk assessment.
Concise database can be very useful in threshold of toxicological �concern approaches.
Concise database is not appropriate for use in pre-market regulatory risk �assessment of regulated chemicals.
Annex 4 – Slides of Discussion Groups
Summary Report EFSA Scientifi c Colloquium 3, 28-29 April 2005 - Brussels, Belgium 159.
Comprehensive database
Conduct value of information analysis before investing in additions �to the database.
Development of comprehensive database should fi rst focus on �microbiological hazards.
Develop European food categorization system which embraces all �relevant details for risk assessment in different domains.
Anticipate developments in regulations before deciding on details of �comprehensive database.
EFSA would be the appropriate organization for database coordination. �
Ideally, a comprehensive database would also capture information on �determinants of dietary habits to guide risk management policies.
EFSA SCIENTIFIC COLLOQUIUM
SUMMARY REPORT
EUROPEAN FOOD CONSUMPTION DATABASE: CURRENT AND MEDIUM
TO LONG-TERM STRATEGIES
28-29 April 2005, Brussels, Belgium
TM-A
D-07-004
-EN
-C
Largo N. Palli 5/A I-43100 Parma Italy
Tel: +39 0521 036 111Fax: +39 0521 036 [email protected]
EF
SA
SC
IEN
TIF
IC C
OL
LO
QU
IUM
SU
MM
AR
Y R
EP
OR
TE
UR
OP
EA
N FO
OD
CO
NS
UM
PT
ION
DA
TAB
AS
E
3
ISS
N 1830
-4737
3