2.3.1 Food balance sheets
The food balance sheets (FBSs) assembled by the Food and Agriculture Organisation (FAO) describe the current and developing structure of the national dietary patterns, in terms of the major food commodities that disappear from the national markets (www.fao.org). A food balance sheet is completed at national level, on the basis of the annual food production, imports and exports, changes in stocks and the agricultural and industrial uses within a country. When these have been taken into account, the remaining quantities represent the food that can be assumed to have been available for human consumption in that country (Kelly et al, 1991).
Since 1949, FBSs are regularly collected on a world-wide basis and, in spite of their limitations, countries with no routine information on the food consumption of their population and those interested in comparing their national dietary patterns with those of other populations have traditionally used them (Helsing, 1995).
International comparisons based on the time series FBS data, in conjunction with information from other sources, can help to indicate trends in the food available to the overall population of one country in relation to others, and have thus been used for ecological correlations of food patterns with the morbidity and mortality of nutrition-related diseases. The user of these data, however, should bear in mind their constraints and interpret comparisons with due caution (Southgate, 1991). The accuracy of recording differs considerably between countries and commodities. Although data on their own food production are collected in some countries, these sources of information can be largely under-recorded. Waste and food given to pets may also be sources of error, since they are considerably dependent on time, cultures and type of commodities. Lastly, the conversion of foodstuffs into nutrient equivalents by the application of factors derived from various sources must be prudently treated.
The household budget surveys (HBSs) are periodically conducted by the National Statistical Offices of most European countries in nationally representative samples of households. By recording the values and quantities of household food purchases, the HBSs can adequately depict the dietary patterns prevailing in representative population samples. Moreover, the concurrent recording of demographic and socio-economic characteristics of the household members may allow exploratory analyses on the evaluation of their effects on dietary choices. One of the main advantages of the multi-purpose HBSs is their periodic undertaking by Governmental Services, making them a readily available and thus an affordable source of dietary information in developed and developing countries (Trichopoulou, 1992).
The HBSs can be thought of as occupying a position between the FBSs and the specially designed individual food consumption surveys. Like food balance sheets, the HBSs allow intercountry comparisons on a regular basis but, moving from total population to household level, they further allow the calculation of both the mean and the distribution of food availability within the population and specific subgroups (Trichopoulou et al, 1999).
Issues of comparability can be raised when using HBS data for international comparisons. The data collection methodology is uniform enough to allow such comparisons, but the food information recorded in the various countries may be of different forms and levels of detail. The methodology, however, for addressing these discrepancies has been developed in the context of the DAta Food NEtworking (DAFNE) project (Lagiou et al, 2001; Friel et al, 2001). However, since HBSs are not primarily designed to collect nutritional information, the food data bear limitations, which need to be considered when they are used for nutritional purposes (van Staveren et al, 1991; Southgate, 1991; Trichopoulou, 1992). The following points should be borne in mind:
Despite their limitations, the HBSs provide a resource for the conduct of a wide range of nutritional analyses. They also constitute a reasonable alternative to specially designed individual-based nutrition surveys for most Mediterranean and central/eastern European countries. HBS data could help highlight issues such as differences in dietary patterns (Byrd-Bredbenner et al, 2000), high risk population groups on account of their nutritional habits, relationships between diet and morbidity/mortality data (Lagiou et al, 1999) and dietary intakes of additives and contaminants.
The specially designed individual dietary surveys (IDSs) primarily aim at the collection of information on the food intake of free-living individuals over a specified period. The individual surveys, when intakes of the subject are recorded as adequately as possible, are expected to provide evidence on the food quantities consumed and to allow the calculation of both the mean and the distribution of food and nutrient intake among the whole or segments of the population.
The quantification of foods consumed and the selection of items to be included in the food list, in the case of closed lists, are critical components of data collection. Standard, natural and household units, three-dimensional food models, photographs, drawings of foods and geometric shapes are often used for documenting portion sizes.
Recall methods, in comparison to the record ones, do not require literacy; they are not expected to cause alterations in the eating behaviour of the subject, since the information is collected after the fact; and they have minimal respondent burden. Nevertheless, recall methods are subject to respondents' memory, a limitation not present in food records. In recent surveys, dietary recalls are collected using computer software programmes that allow data to be uniformly collected, by prompting interviewers to ask all the necessary questions, and may further reduce the cost of data collection and processing.
The food records and the 24-hour recall may be used to estimate the absolute intake of energy, macronutrients and some vitamins and minerals that are commonly found in the food supply. Both methods are frequently used in describing the mean intake of aggregated food groups and in validating food frequency questionnaires. These short-term methods are completely open ended, they accommodate any food or food combination reported by the subject and they allow recording information at various levels of detail including the type of food, the food source, the food processing and preparation methods. They are therefore particularly useful for estimating intakes of culturally diverse populations. One single day of intake, however, is highly unlikely to be representative of usual intake. For this reason the collection of multiple days of intake is required in order to estimate as adequately as possible the subjects' usual intake.
Food frequency questionnaires are food lists of differing length and the information collected can refer either to the frequency of consuming certain foods and beverages, or to both the frequency and estimates of the portions consumed. The underlying principle of the food frequency method is that the average long-term diet reflects the conceptually important exposure, and therefore makes the food frequency questionnaires the method of choice for measuring dietary exposures in epidemiological studies. In constructing a food frequency questionnaire, careful attention must be given to the format of the food frequency section, the selection of foods that will be included in the food list and the clarity of the questions. Food frequency questionnaires can be administered to large population groups; they can be applied as interviews or in a self-administered form and are relatively easy and less time consuming to complete when compared to other dietary assessment methods. It should, however, be borne in mind that food frequency questionnaires including a restricted food list may result in reducing the true variance of intake.
For most investigations of nutritional epidemiology, the relative ranking of individuals according to their food and nutrient intakes is adequate for determining correlations of relative risks. In such cases, food frequency questionnaires constitute the primarily selected dietary assessment method. In situations, however, when the aim is to compare the nutrient intakes of various populations or to evaluate compliance with dietary recommendations, estimates of the absolute energy and macronutrient intakes may be required. In such instances, records or 24-hour recalls are generally the methods of choice (Willett, 1998).
Was this article helpful?