Eurocode 2 Review: Summary of the Main Findings
Prepared by Ian Unwin, Food Information Consultancy
in collaboration with
Anders Møller, Food Informatics, Danish Veterinary and Food Administration
Introduction
A report entitled Eurocode 2 Review and modified Core Classification is now available from the Danish Veterinary and Food Administration. This note summarises the main findings of the report which are to be discussed at a COST Action 99 Eurofoods workshop being held in Wageningen on 25 October 1998.
Requirements for food coding
The Rationale Document for the Eurocode Food Coding System, produced in 1991, explains the principles that form the basis of the coding system. It describes general aims of food coding systems and the requirements of Eurocode 2 "as a practical code for intake assessment and table cross references". There remain some doubts that it fulfils this objective, which possibly needs more precise definition. A review and restatement of the requirements should form a key and early task in further development work.
Some examples may be helpful. A criterion of "nutritionally important in the European diet" is used in designing the coding system. This limits the usefulness of the system in measuring changes in food consumption patterns in Europe. The requirement for simple codes logically associated with the foods they represent was important a decade ago but this should now be considered obsolete, being superceded by a need for the widespread availability of effective coding support computer systems. Codes assigned to represent the logical order and hierarchy of classification categories are incompatible with the requirement to add new categories to the system when needed. A final example is the matching of coded consumption data with food composition data. Two separate food codes can be linked to a single food composition record but problems will arise if two composition records have identical Eurocodes. The information to determine which is appropriate will have been lost.
The report suggests a possible basic range of requirements for a food coding system as including the recording of a consumed food, support for the retrieval of relevant food records, matching consumption and composition records, aggregating consumption data (both for population groups and food groups), and disaggregating mixed-food information back to ingredient data using recipe information. This set of requirements needs to be confirmed and specified in more detail. A separate set of requirements should be defined for the practical operation of a food coding support system.
Eurocode2 classification hierarchy
The Eurocode2 classification hierarchy consists of three main levels, with a further one for recording a recipe identifier for mixed dishes. At the top level, the classification assigns foods to one of 13 main food groups. These are divided into sub-groups which can be further divided into a food item level (which in some cases may be lower level sub-groups, e.g. code 3.11.6, 'Meats; Meat products; Offals').
For most of the main groups, coding categories have been fitted into this 3-level hierarchy fairly successfully, but it is clearly inadequate in the 'Meat and Meat Products' main group. This is the reason that there is no code for 'Chicken'; the sub-group has been defined as the more general 'Poultry' and the food item level as more specific terms such as 'Chicken breast'. This shortcoming may also cause the inconsistency between defining terms for carcass meat cuts (for 'Beef' and 'Pork/piglet') and just a 'carcass meat' category (for 'Veal' and 'Mutton/lamb'). An intermediate level separating the terms for carcass meat and for other organs would make the organisation of the classification clearer and simplify some aggregations of coded data. Perhaps more importantly, it would allow additional categories to be defined for recording consumed food items at the level of detail which may be reported.
Categorisation and coding policy
It is a stated requirement of Eurocode that there should only be one code that can be assigned to any given food item. Since the primary classification is into the main groups, any food must be unequivocally assignable to a main group. It is difficult to define logical rules, particularly for mixed foods. Anomalies are also created because the categorisation policy differs between the main groups.
So-called 'generic mixed foods' are included in the Eurocode 2 classification with specific codes whereas other mixed foods are coded as dishes based on more general food codes, for example 'Chilli con carne' either as a Bean dish or a Meat dish. The rules to determine whether a given dish corresponds to a particular generic mixed food or should be coded as a dish of a single food are in some cases based on cultural or terminological interpretation. The decision often determines in which main group a dish is coded, although more detail of the ingredients will be available if a reference to the recipe file is recorded using the fourth field of the code.
The policies for defining main groups and their subdivision can have major effects on the information recorded by the food coding system. In Eurocode 2 there are two main alternatives based on biological source and product type. A main group may switch from one basis to the other in defining subgroups and their food items (e.g. Miscellaneous foods; Soups; vegetable base, etc.) or in defining further subgroups (e.g. Meats; Meat products; Sausage, frankfurter type). A particular anomaly arises with Soybean as a pulse because all its products, including soya milk and soya flour, are coded together, losing the product type information. A further anomaly provides a 'Milk and milk products' code for 'Cheese substitutes' containing reduced or no milk fat. Clear policy statements would help to avoid such anomalies and would also be very useful as part of the coding documentation.
The recipe and descriptor systems
The Eurocode 2 classification is designed for the coding of foods as they are consumed, but there may be a need to record additional information for mixed foods based on recipes or when a food may be consumed in more than one form. These situations are handled through associated parts of Eurocode, the recipe and descriptors systems respectively. These are less developed and tested than the Eurocode 2 classification.
In addition to the rules for coding recipes as mixed foods mentioned above, aspects of the recipe system needing further consideration include:
- the coding of ingredients. Ingredients are recorded as Eurocode 2 codes, plus descriptors if necessary, which may not be specific enough to define the dish or to calculate its composition.
- other recipe information, for example specifying how ingredient amounts, portion sizes and preparation method are recorded.
- management of recipe with minor variations in ingredient information. Standard approaches to handling variations which modify the basic code or create multiple recipes are needed for an effective recipe system.
- recipe file management. The recipe system as currently proposed assumes a national framework but international recipes and widespread availability of recipe files may also be requirements.
The descriptor system provides a simple approach to storing extra information about foods as consumed using terms defined for 5 independent aspects of processing of foods prior to consumption. Some review of the aspects and the assignment of terms to them is required, for example the aspect of 'Thermal treatment at consumption' includes the terms 'Dried' and 'Canned'. Further terms and indeed aspects might be usefully added. For example, a ''Use context" aspect including the term 'Dietetic use' could replace the regular subgroup 'for dietetic use', allowing the basic code to record more detail on the food itself.
Documentation and coding support
The most up-to-date documentation for the Eurocode food coding system is a draft Manual produced in March 1993. It consists of a main body describing the main features and policies of Eurocode with additional explanations, examples and indexes given as appendices. This provides the basis for effective and comprehensive documentation but it does contain a number of errors, inconsistencies and misleading examples. It is important that these are eliminated to help users code accurately and consistently. Some additional explanation of policy and improved examples incorporating the use of the descriptor and recipe systems are also needed. An updated version should be made available on the Internet.
Availability of effective coding support software is considered essential if Eurocode is to be developed further while maintaining compatibility with the present Eurocode 2 classification. It should support coding without the user having to directly assign the storage codes and thus allow the classification to incorporate new codes in a logical order without changing existing storage codes. The full requirements for coding support need to be specified.
The Core Classification
As part of the Eurocode 2 Review, a core classification was proposed to incorporate modifications to the existing classification system based on the study of each main food group. The proposals maintain the existing 3-level hierarchical structure, although the use of extra levels would give an improved classification, particularly in the Meats group. The Core Classification revises the subgrouping level of the classification, without changing the main group of any foods and, for the most part, without dealing with the details at the food item level. It includes extra Product-type subgroups within main groups which are primarily biologically defined, attempting to provide a clearer policy on when the switch from biological to product-type coding occurs. In this, more importance is attached to coding convenience, making the classification as similar as possible to the foods as reported, than to the use of the hierarchy for data aggregation. The latter can be handled by 'profiles' defining the set of codes which contribute to the aggregate.
The Core Classification has been produced in a database format which demonstrates the separation of storage codes from the logical ordering of the classification. This format could provide a basis for developing a prototype coding support system which would be used to further identify the requirements for a fully operational facility.
Conclusions
The Eurocode 2 Review report lists the tasks involved in the further development of the coding system to enable its wider adoption for recording food consumption information. These include some review and restatement of requirements and policies, the better integration of the descriptor and recipe systems into the overall Eurocode 2 food coding system, improvements to the content and availability of the documentation, and preparation of other coding support tools. Together with discussion and agreement of proposed modification to the classification represented by the Core Classification and the extension of this to the food item level, it is considered that the Eurocode 2 system can be evolved into an effective solution for the coding of food consumption data and other food-related information.
Ian Unwin
3 February 1999