Over the last decade food composition activities have increasingly been undertaken by a variety of agencies, programmes, projects and people, for an ever-growing number of reasons. Many national, regional and international agencies acknowledge the importance of food composition data and the need to interchange information that is unambiguous and useful to all those who need it (Rand and Young, 1983; Rand et al., 1987; West, 1985; Lupien, 1994).
The creation of a food composition database calls for an integrated approach to the generation, acquisition, processing, dissemination and use of food composition data.
INFOODS, the International Network of Food Data Systems, was established in 1983 by the United Nations University (UNU), with an organizational framework and international management structure that included a global secretariat and regional data centres. Its mandate is “to improve data on the nutrient composition of foods from all parts of the world, with the goal of ensuring that eventually adequate and reliable data can be obtained and interpreted properly worldwide” (INFOODS, 2003). In the mid-1990s, FAO joined UNU in the INFOODS effort. The main activities of INFOODS at the international level include development of technical food composition standards, assistance to regional data centres and individual countries in developing their food composition activities, and publication of the Journal of Food Composition and Analysis (Elsevier, 2003).
Most countries in the world participate in international fora and are signatories to international agreements that directly and indirectly relate to food composition. The World Declaration and Plan of Action for Nutrition adopted at the International Conference on Nutrition (FAO/WHO, 1992), the Rome Declaration on World Food Security and the World Food Summit Plan of Action (FAO, 1996), and the World Trade Organization's Agreements on Sanitary and Phytosanitary Measures and Technical Barriers to Trade (WTO, 1998a,b) are examples of such agreements.
Currently, there are 17 regional data centres in operation (see Appendix 1). Regional food composition tables have been prepared, both electronically and in printed form (Dignan et al., 1994; de Pablo, 1999; Puwastien et al., 2000), and many regions undertake regular food composition coordination activities and have established technical task forces that involve individual countries in the region.
Most countries now undertake activities relating to the production of food composition data. A national food composition programme is usually the result of the combination and coordination of activities, within a defined administrative framework, related to food composition data generation, compilation, dissemination and use. Many countries have established a steering committee to facilitate such a framework. A steering, or advisory, committee is ideally composed of individuals directly involved in food composition work, that is, the data users, generators, compilers and disseminators. The involvement of data users– agriculturalists, analysts, health professionals, dietitians, nutritionists, food industry personnel and consumer groups – is crucial to the effectiveness of a steering committee.
Often a single organization has overall responsibility for the management of a national food composition programme, yet it is rare that a single organization accomplishes all the activities itself. Regardless of their affiliations, laboratory-based data generators must interact closely with the data compilers, and compilers must interact closely with data users. Data compilers therefore serve the central function and usually also act as data disseminators (i.e. they publish the data, electronically and/or as printed tables). In most countries there also exist other agencies whose activities are directly or indirectly related to food composition data, and who operate in concert with the national programme. National food composition programmes also operate in conjunction with their regional data centres and with ongoing international activities.
The organizational framework of a national programme will depend on the policies and procedures already being followed in the country or region where the programme is being established. Indeed, the national food and nutrition policy of the country concerned may already favour the establishment or updating of a food composition database (e.g. Langsford, 1979); any new programme should generally aim to fit into the framework of the existing national policy.
Many countries will already have experience in the production of food composition data and their use in tables. In developing a database programme, the aim should be to build on this experience. Existing data on foods with known, relatively stable composition can be used in the new database, provided that these data are evaluated and meet the criteria for inclusion.
A decision to embark on the production or revision of a food composition database may be made by government, or within a research institute or department, by professional groups of users (e.g. dietitians, epidemiologists) or, occasionally, an individual researcher.
The advocacy for newly establishing or revitalizing a database programme can effectively be presented in different ways:
Any submitted document should emphasize the potential benefits of such a programme, especially in terms of community health and welfare, national esteem and economic benefits accruing through reduced health costs and advantage to the food industry, agriculture and trade. The availability and usefulness of any existing data and resources should be stressed. In addition, cost estimates that take into account the costs of administration, analyses, data management and data dissemination will be required.
Any group or individual with responsibility for a database programme should pursue the following objectives:
A food composition database should be defined by the uses for which it is intended. Because such a database is essentially a tool for nutritional work in the widest sense, it must be designed with all immediate and proposed uses clearly defined, and potential users must play a major role in its design.
Three aspects are fundamentally important:
When a governmental committee decided to revise the database presented in The composition of foods (Paul and Southgate, 1978), a steering panel was set up to define the requirements of users. The panel consisted of users (government departments, dietitians and research nutritionists) and compilers, as well as the person in charge of the analytical work and those responsible for the design of the computerized database. The steering panel consulted major users of the existing tables (dietitians, researchers, food industry) by questionnaire (Paul and Southgate, 1970) and in personal discussions, and invited comments by advertisements in the scientific and food press. The compilers collated this information and used it to plan the revision.
A user questionnaire was also used in the early stages of the Pacific Island Food Composition Programme (Bailey, 1991). Other methods for obtaining suggestions from users are to hold a public meeting (Greenfield and Wills, 1981) or national conference (Food and Nutrition Research Institute/National Research Council of the Philippines, 1985), or to solicit submissions from scientific societies (Bernstein and Woodhill, 1981).
Users' contributions to the programme should be continuous, to ensure that the database is both relevant and practical. It may therefore be useful for professional associations of users (or a consortium of them) to form a committee that would continue to supply information and monitor the programme. Including a session or workshop on the subject at an annual national or regional nutrition conference (e.g. the Sociedad Latinoamericano de Nutrición conference), or holding food composition conferences of the type held annually in the United States (USDA, 2003b), may be useful as a forum for this purpose.
This overall strategy in the design of a database programme and definition of users' requirements is illustrated in Figure 2.1.
The stages of an ideal food composition database programme are set out in Figure 2.2. Funding must be obtained and procedures established for communication between all relevant parties. All existing food database programmes and facilities in the country should ideally be coordinated, because much of the analytical work can be done cooperatively by government, research institutes, or industry laboratories working in food research or related fields. Facilitation of this collaboration should be an early, important priority.
Obviously, a budget will have to be drawn up; Box 2.1 lists the various items that need to be provided for.
Figure 2.1 Initiation of database programme: definition of users' requirements
Figure 2.2 Establishing priorities for sampling and analysis
|Box 2.1 Major elements in the budget of a food composition database programme|
Usually, information on the composition of locally available foods already exists, even in countries that have no formal national tables of food composition. The first stage is therefore to evaluate this information, both published and unpublished, for its suitability as data sources (see Chapter 10 for the principles guiding this evaluation). Consideration of user requirements reveals what new information is required, and proposals for new sampling and analytical programmes are made. In most countries it is necessary at this stage to define priorities; this will require further input from the users of the data system.
Sampling and analysis should be considered together, both to ensure data quality (Chapters 5, 6, 7 and 8) and also because the resources required for sampling and analysis need to be estimated together.
In developing the sampling plan and protocols (Chapter 5), a considerable range of inputs is essential, and the compilers need to consult widely. If, as happens in many countries, part of the programme is assigned to a contractor, the compiler must ensure that the contractor is aware of user requirements and the quality standards that have been set for data entering the system.
Sampling and analytical programmes are most conveniently focused on specific foods or groups of foods. This focus on specific foods is also useful in defining the experience required of groups invited to tender contracts. This stage is shown schematically in Figure 2.3. The proposed time scale for the work will determine resource requirements, and logistical factors need to be considered carefully. Once these factors have been assessed it is possible to estimate the costs of the different sections of the programme and submit a budget for approval.
Analysts must plan carefully to ensure that a balance is kept between personnel, laboratory space, equipment, running costs, and so on. Analysts preparing budgets or submitting contract proposals should highlight funds needed for meeting any specific requirements for their laboratories, as it is unlikely that any laboratory will already be ideally suited to carry out the work. Budgetary considerations will vary from country to country. Where labour is expensive, investment in automated equipment may be most advisable. Where labour is inexpensive, more staff can be employed. Wet chemical methods may be more appropriate if it is difficult to service and obtain parts for instruments.
Figure 2.3 Development of sampling and analytical programmes
Tasks in addition to chemical analyses include the regional collection of foods, determination and preparation of edible portions of foods, estimation of serving sizes and consideration of cooking methods (see Chapter 3). Groups with the appropriate technical facilities can carry out this work separately from the analytical programme, if necessary.
In principle, the concept of data quality is built into the analytical procedures (Chapters 7 and 8), and the users' steering group will ensure that the analysts are aware of the detailed requirements of users. Nevertheless, it is useful to review analytical programmes regularly to reinforce the overall objective of the analyses – the construction of a food composition database for many different types of user.
Conversely, analysts should keep the users' steering group informed of both the limitations of, and improvements in, analytical methodology, in order to ensure that the group works with realistic expectations.
Arrangements must be made for regular reports from the analytical laboratories. Requirements for reports must be carefully specified so that all analytical data are provided. For example, a protein value alone should not be accepted if the method used was nitrogen (N) determination. In this case, the N value and the factor used or suggested by the laboratory should be provided along with the calculated protein value. Units and rounding criteria must also be specified for reports. Policies must be established regarding the publication of laboratory results before their release in the food composition database. It is generally desirable for the work to be published independently so that the scrutiny of referees will strengthen its scientific validity.
Data provided by the analytical laboratories are subjected to initial evaluation (Chapter 9), ideally in discussion between compilers and analysts, to ensure consistency. Difficulties that may have arisen during the execution of the work can also be discussed at this time. Inevitably, problems will have required those involved in sampling or analysis to depart from the formal protocols. It is vital that the compilers be fully aware of such changes.
Once sufficient information has been accumulated, it is desirable to initiate reviews by the users' steering group and by external specialists in the relevant commodity or food. The users' review provides an assessment of whether the objectives defined by the users are being met; furthermore, it provides a means of managing the progress of the programme.
The external review serves as a conventional peer review and ensures that the data being acquired are compatible with specialized knowledge (which may not be nutritionally oriented) regarding the commodity or foods. Where proprietary products are involved it is desirable to submit the data to the manufacturer for comment. This step will identify inconsistencies with the manufacturers' quality-control data and will indicate whether the food samples analysed were representative of normal production.
The compilers should work closely with the users' steering group. A review by users of sections of the database as they are prepared is highly desirable. These reviews enable users to alert the compilers to problems regarding format, user-friendliness and adequacy of data, and enable the compilers to alert users to problems of inadequate data or to indications that further analytical work is needed. As the database nears completion, pilot trials of its operation become desirable. These trials can be organized through the users' steering group.
Once the database starts to be used, a series of operational studies is desirable. Although studies designed specifically to test the database are valuable (see Chapter 10), the real tests come with regular use, and provision should be made to collect and collate information on difficulties or inconsistencies encountered by users. Errors must be centrally recorded so that the database can be corrected. It is especially important that the database maintenance be seen as a continuous operation.
It is also desirable to establish a permanent users' group, familiar with the programme's original criteria, which will periodically consider extension and revision of the database.
Continuous or periodic revision is essential for several reasons. The level of a food's consumption can change, particularly with the appearance of “new” foods (e.g. instant noodles). The nutritional quality of a traditional food may also change (e.g. changes in animal husbandry and butchering affect the fat content and micronutrient quality of meats). New methods for preparing convenience foods may have striking effects on a food's nutrient composition (e.g. extruded potato-based snacks, depleted of vitamin C) or on its nutritional consequences for sensitive individuals (e.g. the swing towards fructose syrups and sweeteners). Moreover, in addition to changes in foods themselves, advances in analytical methodology may indicate a need to re-analyse foods for a particular nutrient. These trends necessitate continuous nutritional monitoring of the food supply (Paul, 1977) and indicate that a database should be revised from time to time or on a continuous basis. The advent of computer database systems simplifies, in principle, the continuous updating of a database and periodic production of derived databases or tables.
In view of the fact that copyright and intellectual property legislation varies from country to country (Ricketson, 1995), database compilers will need to familiarize themselves with the national and interna-tional provisions and abide by them. Such provisions may include the need to seek permission to use the data, the format of acknowledgement required and the payment of a royalty. Further, normal scientific conventions should be followed regarding the acknowledgement of all data sources so that users can refer directly to the original source.
The organization responsible for the food composition programme, with the endorsement of the national steering committee, will generally publish the food composition data in various printed and electronic forms, and may charge users for the material cost of the publications. The USDA National Nutrient Database for Standard Reference (USDA, 2003a) is an example of a database that is freely available in the public domain. At the same time, provision should be made for licensing the data for commercial users (Greenfield, 1991b), such as diet analysis software developers, who may then on-sell their product with the data.
Figure 2.4 Schematic overview of the organization of a database programme
The schematic outline of the programme in Figure 2.4 shows the organizational elements of a food composition database programme and some of the responsibilities of each component. The whole programme requires communication back to the higher level and, indeed, constant interaction as proposals are made, priorities established, work designed and executed, and the final product reviewed. The compilers form the executive members of the programme, ensuring that objectives defined by the users' steering group are met and that quality is maintained.
In practice, the compilers may be several individuals, each responsible for a single area (e.g. literature review, supervision of analytical programmes or data on certain nutrients, commodities or foods). If resources permit this division of labour, which enables specialized knowledge to develop, it is essential to have a good line management so that the senior compiler has a clear overview of the work as a whole.
Continued interaction with the relevant regional data centre is usually helpful in ensuring that standards are maintained and that data are compatible.