|
|
|
| Food consumption, in the broadest sense, can be estimated at different levels, including population level (e.g. food balance sheets, which are based on food availability or market supply), household level (e.g. household budget surveys, which usually record household food purchases), or at the individual level (dietary surveys). | |
| Dietary surveys at the individual level are one of the main tools for assessing dietary adequacy. They collect information on the types and amounts of food consumed by a (usually) representative sample of individuals within a population. This information is combined with data on the nutritional value of each food to estimate the intake of nutrients of interest. In some cases, blood samples are also collected in order to assess nutritional status. | |
| Reliable, up-to-date and representative food composition data are essential to ensure that nutrient intakes estimated from dietary surveys are robust. Errors may lead to incorrect public health measurements and conclusions concerning the relation between diet and disease. | |
|
Dietary assessment
Dietary assessment may be undertaken in a number of formats, including:
-
at a national level, in order to monitor the effectiveness of government policies and to help formulate appropriate dietary advice (e.g. the comprehensive National Diet and Nutrition Surveys in the UK, the Dutch National Food Consumption Survey, and many others).
-
on specific population sub-groups, particularly vulnerable sub-groups (e.g. hospital patients, lone parents, homeless people) or regional groups.
-
as part of major international nutrition epidemiological research studies (e.g. European Prospective Study of Cancer and Nutrition (EPIC)). Specific considerations for international studies will be the subject of a future web feature.
-
as part of individual nutrition research projects to test specific hypotheses (e.g. the impact of specific nutrients on bone health).
The dietary assessment technique chosen will depend on the aim of the survey and the desired outputs. Those commonly used include:
-
24 hour dietary recall, in which an interviewer questions the participant about the food and beverage consumed during the previous day. To give a more representative picture of an individual’s diet, the 24-hour recall may be repeated on a number of separate occasions.
-
Dietary records, in which respondents record the foods and beverages consumed over a specified number of days. Portion sizes may be determined by weighing or by estimation using household measures.
-
Food frequency questionnaires, in which respondents are asked about the frequency with which they consumed foods listed in a questionnaire over a specified period of time. Information on portion size may also be collected.
A fourth method, which is unusual in not requiring food composition databases, is the duplicate diet method, in which an exact duplicate of an individual’s diet is analysed for components of interest. Although this method is accurate, it is costly and time-consuming, and is seldom used. | |
|
Dietary survey nutrient databanks
Nutrient databanks for dietary surveys are usually based on available food composition databases, often national or official databases. However, they differ in a number of ways, primarily:
-
in general, dietary survey nutrient databanks have values assigned for all nutrients in all foods
-
they usually contain a larger number of foods, when additional detail is needed, or to incorporate data on home-made dishes
Assigning nutrient values | |
| Nutrient databanks for use in dietary surveys generally have values assigned to all nutrients in all foods, unlike many food composition databanks, which may have missing values where a definite value could not be assigned. It is important that missing nutrient values/gaps in data should not be treated as zero in dietary surveys because this would result in an underestimation of nutrient values. Missing values are therefore usually estimated, often by referring to nutrient levels in similar foods. | |
| There may be additional checks on manufacturers’ data (e.g. food labels) for nutrients added for fortification, colouring, or antioxidant purposes, especially if the survey is focussing on specific micronutrients. In addition, manufacturers’ data may be used to update information that may be out of date in food composition databases (and particularly printed tables), which cannot keep up with changes in industrial foods. Examples include the trans fatty acid content of fat spreads and the sodium content of processed foods, both of which are being reduced by food manufacturers. | |
| For home-made dishes consumed by participants in dietary surveys, recipe calculations are undertaken if there are no suitable matching products within the nutrient databank. These calculations are based on the nutrient content of the individual recipe ingredients, with adjustments for moisture and nutrient losses on cooking. | |
|
Food classification and coding | |
| Foods in dietary survey nutrient databanks are classified at different levels, including individual food level (e.g. bananas), and food group level (e.g. fruit). Individual food code numbers are used to code the dietary records that have been used to collect information on the foods consumed. | |
| It is clearly important that foods consumed are correctly identified and coded in order to ensure that nutrient intakes are as accurate as possible. Errors might occur, for example, in foods with similar names (e.g. squash can refer to a vegetable, or, in the UK, to a soft drink), or if cooked foods are coded as raw. | |
| The level of detail used for food coding reflects the dietary survey methodology, availability of nutrient data, and the purpose of the survey. For example, consumption data collected by food balance sheets, household budget surveys, and food frequency questionnaires would generally have a less detailed food coding frame than that used in a weighed dietary record at the individual level. | |
| Food coding frames generally become more complex as more nutrient data become available; for example, initially data might only be available for ‘apples’, but new analytical data on individual cultivars would enable the main cultivars to be coded individually. However, there also needs to be a consideration of the level of detail likely to be provided by participants in a dietary survey, since it would not be beneficial to have codes on individual cultivars if respondents only recorded that they consumed ‘apples’. | |
| Dietary surveys focussing on specific nutrients might have coding systems that provided more detail on rich sources of these nutrients. In addition, food consumption data are sometimes used for estimation of intakes of non-nutrient components. Thus for example, there might be separate food codes for canned and bottled soft drinks, even though they are nutritionally similar, if there is an interest in contaminants present in packaging materials. | |
A number of outputs of the EuroFIR project will benefit those working in the field of dietary surveys. For example, one of the activities that has been undertaken in work package (WP) 2.2 (Composite, processed and novel foods) within EuroFIR was the development and harmonisation of guidelines for imputing the composition of composite foods. In addition, the aim of WP 1.6 (Food description) is to harmonise existing food identification and description systems for use in food composition databank systems in order to conform to European dietary habits and the needs of intake assessments.
| |
|
