| Food Composition Data: A User's Perspective (1987) |
|The uses of food composition data|
|Food composition -a key to dietary appraisal and improvement in the United States|
National nutrition monitoring system
As part of the National Nutrition Monitoring System, HNIS maintains the National Nutrient Data Bank, sponsors related food composition research, and conducts surveys in order to monitor the food and nutrient content of American diets. The conversion of information on food consumption into nutrient consumption is the agency's major use of food composition data. The nutrient consumption of the nation's population is monitored at three levels: food in the US food supply, food used by households, and food eaten by individuals. To make these conversions, nutrient data are needed both for food as purchased and food as eaten; and data for each are needed by appropriate units of weight. Representative values - those for the food as purchased or as eaten throughout the country year-round - are appropriate for such conversions.
The food supply data, estimated for each year since 1909, show trends in quantities of nutrients available to the population and in the food sources of these nutrients. For example, the data show that dietary fat in the food supply has increased since the beginning of the century from 124 g to 166 g per person per day . The increase comes primarily from fats and oils. These statistics require that food composition data reflect the foods available for consumption during given periods. For example, the composition of poultry or pork in the 1980s is not the same as it was two decades earlier, and this series reflects those differences.
The food supply covers fewer than 400 basic foods - foods before they are combined into food mixtures. Nutrient data for these relatively few foods in their primary forms are generally readily available. Because of this, estimates of a nutrient's content in the food supply usually can be made before estimates for diets of households and individuals, which contain many more foods. An example is zinc . Several years ago, we had sufficient data to assess the food supply and recognize that the zinc levels of 11.5 to 12.5 mg per person per day since 1909 are not sufficient to provide Americans with the recommended dietary allowance of 15 mg per day . Sufficient data are only now becoming available on enough foods to allow us to assess zinc levels in household and individual diets. Many other countries make nutritional assessments of their food supplies based on statistics such as these.
Information on household food use and individual food intakes is collected by interview about every 10 years in Nationwide Food Consumption Surveys, the most recent of which was conducted in 1977|78. The next is planned for 1987. In 1977/78, data were collected in the 48 conterminous states for about 38,000 individuals in 15,000 households . The surveys showed the kinds and amounts of foods that households used during one-week periods and the food and nutrient intakes of individual household members for three-day periods.
The food used by the household, expressed in pounds of foods as purchased or brought in, is translated into the nutrient content of the household diet by using tables of the nutritive values of the edible parts of the food as purchased. Vitamin content is adjusted for usual losses during cooking. The food intake by individual household members, expressed in grams of foods eaten, is translated into nutrient content by using tables of the nutritive value of the edible portion of foods. To make these calculations requires representative values for each of the thousands of foods in forms that Americans buy for use at home and of foods that they eat both at home and away from home. Analyses are not available to provide reliable data for all nutrients for all of these foods. Therefore, some values are imputed from another form of the food or from a similar food to provide the best estimates of the nutrient content of diets.
Another data file is required for survey data analysis. It converts the quantity of household food reported in packages of rice, loaves of bread, cartons of milk, heads of lettuce, and cans of beans into pounds. It also converts quantities of food individuals ate - a hamburger pattie, a cup of milk, an orange, a piece of cake - into grams of edible food.
Over the decades, these surveys have required data not only on more and more foods but also on more and more nutrients. For example, data on various types of dietary fibre, carotenes, selenium, and tocopherol are now of special interest. Our nutrient data research is directed toward filling these and other needs. In the meantime, diet appraisals based on composition that are limited must be recognized as tentative and possibly misleading. Another concern in appraising diets - and one that can be addressed at this time only in a general way- is the availability of nutrients to the body as affected by the food carrier and other foods and drugs consumed.
As diets are monitored for more nutrients, the descriptive information required about the foods in the survey increases. For example, if information on only total fat in the diet is required, the specific ingredients in margarine are not of major concern. However, if the levels of fatty acids are to be measured, information on the types of oils in the margarine must be obtained. The sodium content of a vegetable depends on whether it was prepared from food in the raw, canned, or frozen, and on what seasonings were added in processing and preparation.
HNIS has considered such problems in planning the questionnaire and analysis for the Continuing Survey of Food Intakes by Individuals begun in April 1985 . New survey questions were introduced in an attempt to obtain the food descriptions needed for more reliable estimates of the levels of cholesterol, total fat, fatty acids, and sodium in diets. Some of this descriptive information the respondent may not be able to provide. The decision must then be made as to whether the missing information can be estimated with reasonable accuracy or whether such estimates are so poor as to be misleading.
Information on food composition has expanded and improved considerably since the last national survey was conducted in 1977/78. At that time, intakes were assessed for food energy; for the macro-nutrients - protein, fat, and carbohydrate; for four mineral elements - calcium, iron, magnesium, and phosphorus; and for seven vitamins - A, thiamine, riboflavin, niacin, B6, B12, and C.
For the new Continuing Survey, HNIS nutrient data specialists have updated and extended the data file to include 12 more nutrients and other food components than used in the previous survey: three classes of fatty acids, cholesterol, sodium, potassium, zinc, copper, carotenes, vitamin E, folacin, and dietary fibre. This data file will show the content of food energy and some 27 food components for 100-gram portions of each of about 4,000 foods as consumed, and will contain the most up-to-date information in USDA's Nutrient Data Bank. It is important that users of these data and of the survey results based on them recognize that values for these food components are not equal in reliability. For some food components relatively strong data support the values. Data for others- dietary fibre, vitamin E, and carotenes - are less well founded. If analytical data are not available, values are imputed.
The food consumption survey activities discussed are specified in an implementation plan for the National Nutrition Monitoring System sent to Congress by USDA and DHHS in 1981 . The core surveys of the system are USDA's Nationwide Food Consumption Survey (NFCS) and DHHS's National Health and Nutrition Examination Survey (NHANES).
NHANES collects information for one day's food intake and the frequency of consumption of selected foods; its primary focus is on the nutritional health status as determined by physical examination. The dietary data are assessed for nutrient content using composition tables, and the new data file developed for the Continuing Survey will be updated for use in future NHANES. Some analyses of NHANES data attempt to associate the nutrient content of a person's diet with his health status. Such studies might benefit from more precise data on the composition of the food eaten by that person than is given by the representative values in reference tables.
The statistics from NFCS and NHANES tell much about the diets and state of the nutritional health of the population. They make possible the monitoring of the dietary status of the American population to identify populations at risk, problem nutrients, food and eating patterns, and diet determinants. They have shown that many American diets do not meet desired dietary standards : many are short of recommended levels for certain nutrients such as calcium, iron, zinc, and folacin, some are short in vitamins A and C, and many exceed moderate levels of fat, sodium, and added caloric sweeteners.
A systematic exploration of uses of food consumption information was made by the Food and Nutrition Board (FNB) . Its report of June 1984, National Survey Data of Food Consumption: Uses and Recommendations, identifies the primary uses of food consumption survey data and recommends effective means of obtaining those data. Many of the uses listed are secondary uses, in that they require information on the nutrient content of diets and the nutrient contribution of specific foods to diets.
Dietary data from these surveys are used for many purposes: to monitor dietary status, to measure the economics of food consumption, to formulate and show the effects of food assistance and regulatory programmes, and to provide the basis for food selection guides - to name but a few.