J.R. Allen
Jane Allen is a research dietitian, James Fairfax Institute of Paediatric Clinical Nutrition, Royal Alexandra Hospital for Children, Camperdown NSW 2050
In clinical research in nutrition, food composition data are used to evaluate the adequacy of the nutrient intake of individuals or groups. Adequacy means the probability of individuals or groups having an adequate nutrient intake (Beaton 1986). It is difficult to assess nutritional status from food intake data alone. Nutrient intake data are usually correlated with biochemical and anthropometric measurements. The final step is to test for a relationship between the observed nutrient intake and the health or disease state of the study sample. Accuracy in the measurement of food and hence nutrient intake is essential to establish true relationships (Pennington 1988).
The accuracy of the nutrient intake measurement is dependent on the quality of food composition data (Guthrie 1989). The quality of the food tables is determined by the number of food samples analysed, the analytical methods used and the quality control in the laboratory. However, the variability in the measurement of food intakes is greater than the variability in the food composition data (Paul & Southgate 1988).
Sometimes there is a need for direct food analysis, that is, to create the laboratory's own set of food tables. This situation occurs when there is no available nutrient information about a particular food or when the nutrients being investigated cannot be calculated accurately from food tables (vitamin C, sodium, folates). In metabolic balance studies, direct food analysis is necessary to know the intake of particular nutrients accurately.
Food tables
Errors in the use of food tables can be minimised by following a few simple rules, eg read all the descriptive notes accompanying a set of food tables; note particular items which vary between food tables, such as the factors which have been used to calculate the metabolisable energy in the diet from protein, fat and carbohydrate, the multiplication factor(s) for nitrogen to calculate the protein content of different foods and so on; observe the method for calculation of niacin and retinol equivalents for these may also vary between food data bases. The convention for missing values also needs to be known as a calculation of nutrient intake which leaves missing values as zero will be an under-estimate (Anderson 1986). The extent of this underestimation has been calculated using the British food tables (Paul & Southgate 1978) for a group of pregnant and lactating women (Black & others 1985). It is better to estimate a value from a related food than leave a zero.
Dietary records
There should be a check that the dietary records to be analysed contain enough specific information to enable the correct food item to be identified in the tables. When this specific information is missing it is necessary to make rules about the choice of a food. These rules need to be documented and used consistently in all data coding. For foods or nutrients not listed, information may be obtained from other food tables, manufacturers, food analysis, or substitution with a similar food or a combination of foods to achieve a particular nutrient balance. The methods of food analysis for the missing food need to be similar to those used for the other foods in the diet otherwise an inaccurate calculation of nutrient intake will result.
Recipes require accurate information on the weights of all the uncooked ingredients and the cooked weight of the total recipe. The reduction in weight with cooking is mostly water loss. If the total cooked weight of the recipe is missing, the proportion of the whole recipe eaten can be used in the calculation. When foods gain fat it is preferable to know the fat content before and after cooking (Paul & Southgate 1988).
Manufacturers' information
Information from manufacturers needs to be continually updated as food ingredients are frequently changed, often for economic reasons. It is important to record the date of the manufacturers' nutrient analysis. Usually only a partial nutrient analysis is supplied so the missing values need to be obtained from other sources, as outlined above. If a proportionate list of ingredients (including water) is given, the nutrient content can be calculated from a recipe.
Poor knowledge of nutrition
Another source of error which is difficult to avoid is the use of food tables by people without the appropriate knowledge of food and nutrition to make correct decisions in nutrient analysis of food intake data. This problem seems to have increased with the arrival of computerised food tables.
For the future, food tables may need to include non-nutritive components such as pesticide levels (Anderson 1986). Other needs are ways to reduce time and cost in data processing, improvement of analytical procedures for certain nutrients (eg folates), and a comparison of the different food tables in use. A large unresolved problem in translating food composition data into actual nutrient intakes is the lack of allowance for nutrient bioavailability or interaction between different foods. More information is needed about the proportion of the nutrients actually absorbed from food as eaten.
References
Anderson, SA (ed). 1986. Guidelines for use of dietary intake data. Bethesda MD: Federation of American Societies for Experimental Biology.
Beaton, GH. 1986. Toward harmonization of dietary, biochemical, and clinical assessments. The meanings of nutritional status and requirements. Nutr. Rev. 44: 349–58.
Black, AF., Paul, AA & Hall, C. 1985. Footnotes to food tables. 2. The underestimations of intakes of lesser B vitamins by pregnant and lactating women as calculated using the fourth edition of McCance and Widdowson's The composition of foods. Hum. Nutr. Appl. Nutr. 39A: 19–22.
Guthrie, HA. 1989. Interpretation of data on dietary intake. Nutr. Rev. 47: 33–8.
Paul, AA & Southgate, DAT. 1978. The composition of foods. London: HMSO.
Paul, AA & Southgate, DAT. 1988. Conversion into nutrients. Cameron, ME & Van Staveren, WA (eds). Manual on methodology for food consumption studies. Oxford: Oxford University Press. 121–44.
Pennington, JAT. 1988. Associations between diet and health: The use of food consumption measurements, nutrient databases, and dietary guidelines. J. Am. Diet. Assoc. 88: 1221–4.
