Objective. To identify major food sources of nutrients and dietary constituents for US children.
Methods. Twenty-four-hour dietary recalls were collected from a nationally representative sample of children age 2 to 18 years (n = 4008) from the US Department of Agriculture's 1989–1991 Continuing Survey of Food Intakes by Individuals. For each of 16 dietary constituents, the contribution of each of 113 food groups was obtained by summing the amount provided by the food group for all individuals and dividing by total intake from all food groups for all individuals.
Results. Milk, yeast bread, cakes/cookies/quick breads/donuts, beef, and cheese are among the top 10 sources of energy, fat, and protein. Many of the top 10 sources of carbohydrate (yeast bread, soft drinks/sodas, milk, ready-to-eat cereal, cakes/cookies/quick breads/donuts, sugars/syrups/jams, fruit drinks, pasta, white potatoes); protein (poultry, ready-to-eat cereal, pasta); and fat (potato chips/corn chips/popcorn) also contributed >2% each to energy intakes. Ready-to-eat cereal is among the top contributors to folate, vitamin A, vitamin C, iron, and zinc intakes. Fruit drinks, containing little juice, contribute ∼14% of total vitamin C intakes.
Conclusions. Fortified foods are influential contributors to many vitamins and minerals. Low nutrient-dense foods are major contributors to energy, fats, and carbohydrate. This compromises intakes of more nutritious foods and may impede compliance with current dietary guidance.
- CSFII =
- 1989–1991 Continuing Surveys of Food Intakes by Individuals •
- USDA =
- United States Department of Agriculture
In 1984, Batcher and Nichols1 introduced the idea of “important” as opposed to “rich” sources of nutrients in an analysis of data from the 1977–1978 Nationwide Food Consumption Survey. Whereas rich sources were those foods with the greatest concentration of a nutrient, important sources were those that contributed the most to a population's intake. Important sources of nutrients can be strongly influenced by both the nutrient density of a given food and its frequency of consumption. For example, although yeast breads are relatively poor sources of calcium (white bread contains about 27 g of calcium per average slice), they are consumed so frequently by US adults that they contributed nearly 9% to total calcium intake in 1989–1991.2 Conversely, although very few people consume liver and other organ meats, they are so nutrient-dense that they contributed ∼5% to vitamin A intake. Such knowledge helps to tailor nutrition education messages and design methods for assessing nutrient and food intake in the United States.
In 1985, Block and colleagues3 ,4 provided a more comprehensive listing of dietary sources of nutrients among a representative sample of US adults. These data—from the Second National Health and Nutrition Examination Survey, conducted from 1976 to 1980—and subsequent analyses5–13 have been used by nutritionists, health educators, epidemiologists, and industry to assess which foods contribute most to intakes of nutrients by US adults.
Comparable data for children have been sorely lacking. Only one nationally representative study has been reported12 and that was limited to intakes of fats and cholesterol among 2- to 5-year-olds. The purpose of this article is to examine the contributions of specific foods to nutrient, fiber, and cholesterol intakes of US children, 2 to 18 years of age, using nationally representative data, accounting for the myriad of ways in which foods are consumed. These findings are directly comparable with those reported elsewhere for adults.2
This study used the 1989–1991 Continuing Surveys of Food Intakes by Individuals (CSFII), in which a separate 12-month survey was conducted in each of the 3 years and combined to create one large nationally representative dataset.14 Each year the survey used two independent multistage stratified clustered samples of housing units in the US: an all-income sample and a low-income sample. All were nationally representative. Data for the 3 survey years and all samples were combined to maximize sample size, and weighting factors were used to account for differential rates of selection and for nonresponse. The overall household response rate was 67.2%, and 85.7% of eligible individuals in those households provided 24-hour recall data. Women were more likely than men, and older persons more likely than younger persons, to respond. The sampling weights, which were designed to make the estimates more representative of the US population, adjust for important demographic variables known to be determinants of food consumption behavior.
Trained interviewers gathered data on household composition and characteristics, conducted 24-hour dietary recalls with household members, and instructed them on how to complete dietary records for the next 2 days. Dietary data included detailed descriptions of all food and quantities eaten. The primary meal planner/preparer was asked to report intake information for any children younger than age 12. For this analysis, only the 24-hour recall dietary data, provided by 4008 children age 2 to 18 years were used. Excluding the 2-day record data provided the largest number of respondents, and preliminary analyses showed no important differences in findings between the use of one 24-hour recall versus 3 days of intake. Additional information regarding the sampling procedures and data collection and nutrient content of foods is available.14
The Survey Nutrient Data Base16 provides data on the nutrient content of foods reported in the survey: the values in the database for carotenes are those used by the United States Department of Agriculture (USDA) in arriving at the values for total vitamin A and are not solely β-carotene. Sixteen nutrients which have public health significance in terms of meeting either nutrient requirements or dietary guidance are reported here. Analyses of 12 other nutrients available on the USDA database were done and are available on request.
A total of 2436 individual food codes were reported by children in this study. These foods were divided into 113 food groups of similar nutrient content or consumption based on the judgment of the investigators. A selected number of these food groups were subdivided to provide clarification for findings for certain nutrients. For example, a single food group for yeast bread is reasonable for assessing the contribution of yeast bread to energy intakes, but subdividing yeast bread into white- and whole-grain yeast breads provides more useful information for assessing contribution to fiber intake. Detailed information regarding individual foods making up each food group is available on request.
Of the 2436 individual foods consumed by children in this dataset, 885 (or ∼36%) were designated as food mixtures and selected to be disaggregated using USDA's food grouping system17 to account for the contributions of the various ingredients. These foods included casseroles, pizza, soups, frozen dinners, ethnic food mixtures, salads, vegetables with added fats (including french fries), and sandwiches. Other foods, such as yeast breads, baked goods, and ice cream were not disaggregated because such foods are not viewed as food mixtures in dietary guidance18 or, based on the judgment of the investigators, by the general population. Decisions regarding the grouping and selecting of foods to be disaggregated were based on the judgment of the investigators and have implications for findings. For example, disaggregating pizza will give a truer estimate of how cheese contributes to nutrient intake but does not allow for knowing how pizza itself contributes. Therefore, analyses were conducted both with and without disaggregation of food mixtures to assess both the impact of the method and the contribution of important food mixtures. Data presented in the Tables represent analyses with food mixtures disaggregated. Analyses before disaggregation are discussed selectively in the text and are available on request.
The food grouping system developed by the USDA allows for the capability to disaggregate food mixtures into component parts at several levels. For example, disaggregation of a cheeseburger can go as far as the bun, cheese, ground beef, and condiments, or it can be taken to the next level by additionally disaggregating the bun or condiments into ingredients such as flour, yeast, tomato paste, and so forth. Generally, we disaggregated mixtures to a level at which foods would be classified easily into food groups of the food guide pyramid.19 The USDA food grouping system also takes into account both the yield of the recipe and the effect of preparation method on retention of nutrients.
The contribution of each food to the intake of each dietary component was obtained by summing the amount of the component provided by the food for all individuals and dividing by the total intake of that component from all foods for all individuals. Thus, the weighted percentage contribution for all children is given by: where F i = the amount of the dietary component contributed by the particular food for the ith individual, T i = the total amount of the dietary component from all foods for the ith individual, andw i = the sample weight for theith individual. This formula provides the percentage contributed by each food to children's total consumption of each component.20
In this study, food sources of nutrients were assessed for all children and four age or age–gender subgroups: 1) both genders, 2 to 5 years (n = 1107); 2) both genders, 6 to 11 years (n = 1475); 3) males, 12 to 18 years (n = 710); and 4) females, 12 to 18 years (n = 716). These groups correspond to designations used in nutrition monitoring.15
The Appendix lists all the food groups used for these analyses.Table 1through Table 16 show the food sources for 16 nutrients and dietary constituents. For each nutrient or dietary constituent, food groups are listed in descending order by the percentage they contributed to total intake for all children 2 to 18 years of age. Therefore, the sources within a column for any age–gender subgroup may not appear in descending order. Each Table has been limited to food groups that contribute at least 1% of total intake for all children.
Energy, Macronutrients, and Cholesterol
Table 1 through Table 6 show nutrient sources for energy, macronutrients, and cholesterol. Milk, yeast bread, cakes/cookies/quick breads/donuts, beef, and cheese are among the top ten sources of energy, fat, and protein, reflecting their ubiquitous consumption. Many of the top ten sources of carbohydrate (yeast bread, soft drinks, sodas, milk, ready-to-eat cereal, cakes/cookies/quick breads/donuts, sugars/syrups/jams, fruit drinks, pasta, and white potatoes) and other sources of protein (poultry, ready-to-eat cereal, and pasta) and fat (potato chips/corn chips/popcorn) also contributed >2% each to energy intakes. Fried potatoes and pizza, when not disaggregated, contributed ∼3% each to energy intake and ∼4% each to fat intake.
Soft drinks contribute 3 to 6 percentage points more of both energy and carbohydrate to diets of teenagers than to the diets of 2- to 11-year-olds. Conversely, milk contributes 6 percentage points more energy to diets of 2- to 5-year-olds than to diets of 12- to 18-year-olds. Milk is by far the major contributor to protein intake among 2- to 5-year-olds (25%).
When milk is separated into whole, 1% to 2%, and skim milks (data not shown), estimated contributions to energy are about 7%, 5%, and 0.5%, and to fat, ∼9%, 4%, and 0.1%, respectively (numbers do not add up because of rounding). These findings may overestimate the energy and fat contributions attributed to whole milk, the designated ingredient in recipes. However, food supply data for 1989–1991 indicate that whole milk accounted for 42% to 45% of total fluid milk,21 suggesting that any overestimate is small.
Fiber, Vitamins, and Carotenes
Table 7 through Table 12 show dietary sources of fiber, vitamins, and carotenes. Sources of fiber are dominated by yeast bread, which contributes nearly 14% of total intake for all children, with whole grain and white yeast breads each supplying about half (data not shown). This is attributable to the fact that whole grain yeast breads, although higher in dietary fiber content, are consumed much less frequently than are white yeast breads. Ready-to-eat cereal contributes 7% to 11% of fiber intakes among all age–gender groups, with the lowest contribution to 12- to 18-year-old females. White potatoes and tomatoes, respectively, contribute 6% to 8% and 5% to 7% to fiber intakes, most likely because of frequent intake of french fries, catsup, and tomato sauces (disaggregated from pizza and pasta). The fiber density of dried beans/lentils is highlighted by the fact that they supply only slightly less fiber than white potatoes, although the latter are the most frequently consumed and the former nearly the least frequently consumed vegetables in this country.17
As would be expected, fruits and vegetables are primary contributors to vitamin C, vitamin A, and carotene intakes. Orange/grapefruit juice contributes more than one-quarter of children's vitamin C intake, and that contribution increases with age to about one-third for 12- to 18-year-olds. Carrots contribute ∼15% to all children's vitamin A and nearly half of their carotene intakes. Although folate-dense foods, such as orange/grapefruit juice, and dried beans/lentils, contribute at least 1% of folate intake among children, others, such as spinach/greens, and oranges/tangerines do not because of infrequent consumption.
These data show that fortified foods play a dominant role in micronutrient intake. In fact, because of fortification, ready-to-eat cereal, which has little naturally occurring folate, is the top contributor by far to folate (23% to 33%). Ready-to-eat cereal also is the top contributor for vitamin A intake and the third or fourth highest contributor for vitamin C. Fruit drinks, although only partially made of juice, are the second biggest source of vitamin C for all age–gender groups. They contribute 17% of the vitamin C for 2- to 5-year-olds, a value higher by 3 to 5 percentage points compared with other age–gender groups.
Table 13 through Table 16 show the findings for minerals. Milk supplies >51% of the calcium to all children's diets, although that contribution decreases with age. Cheese also is an important source of calcium, and its contribution increases with age. From the youngest to the oldest age groups, milk is a lesser and cheese a greater contributor.
Ready-to-eat cereal is the major contributor to iron intake among all children (27%), more than three times higher than the contribution of beef (8%), and double that of yeast bread (13%). Cakes/cookies/quick breads/donuts and pasta also are among the top five food sources of iron. These data show the influence of enriched and fortified grain products in providing iron in US diets.
With the exception of 2- to 5-year-olds, for whom milk is the top contributor, beef is the top contributor of zinc intake among all age–gender groups. Ready-to-eat cereal follows milk as the third highest contributor of zinc intakes. Milk is a considerably greater contributor to magnesium intakes (23%) than any other food group examined (≤8%).
This is the first study to provide estimates on the contribution of various foods to intakes of nutrients by children of all ages using a nationally representative sample. Although the survey used was conducted several years ago, these data are the most current findings on this topic. There was a paper published previously that reported data from the 1985–1986 CSFII that was limited to intakes of fats and cholesterol among 2- to 5-year-olds only12; however, that study did not disaggregate food mixtures. Although it is of interest to assess how dietary sources of nutrients may have changed over time, methodologic differences associated with how food mixtures were handled preclude reasonable assessment of trends even among 2- to 5-year-olds.
Ready-to-eat cereal, which is generally fortified with nutrients, has a noticeable impact on contribution to nutrient intakes of children in the United States. Such cereals appear in the top sources for most vitamins and minerals, reflecting both the high level of nutrients added in fortification as well as the high frequency of consumption among all children. Thus, ready-to-eat cereal provides substantial nutrient intake that is not naturally occurring and thereby is acting as a vitamin and/or mineral supplement. Although it may be beneficial that consuming such cereals leads to increased intakes of problem nutrients such as folate or iron, the fact that not enough foods that are naturally dense in folate or iron are being consumed may be problematic in terms of intake of other potentially beneficial nutrients or dietary constituents. The Committee on Recommended Dietary Allowances maintains that recommended nutrient intakes should be consumed as part of a normal diet composed of a variety of food groups rather than by supplementation or fortification.22 However, ready-to-eat cereal is often consumed with milk, which is important in meeting calcium requirements. Nutrition educators can encourage increased consumption of low- or nonfat milk with cereal as well as the addition of fruit to help meet dietary recommendations.
The ability to disaggregate mixtures is a methodologic advance that allows the data to present a truer picture of contributions of each food group listed than would otherwise be possible. These data, therefore, are useful to nutritionists, physicians, and health care providers in determining precisely which food commodities or groups are the most important sources of nutrient intake in the United States. However, these more discrete food groupings may be less useful for establishing food groups for inclusion in dietary assessment instruments using predefined food categories or for monitoring intake of a specific food mixture. For example, one might be specifically interested in looking at the contributions of food mixtures such as pizza or fried potatoes. For this reason, we have presented data for some of these most frequently consumed food mixtures.
Dietary guidance continues to recommend that children limit fat intakes to ≤30% of energy.23 Consuming less fat often requires limited intakes of added fats or of foods prepared with fat. These data show that such fats are well represented in the top 10 contributors by margarine; cakes/cookies/quick breads/donuts; potato chips/corn chips/popcorn; salad dressings/mayonnaise; oils; and other fats. These data also show the strong presence of other low nutrient-dense foods such as cakes/cookies/quick breads/donuts; soft drinks/soda; sugar/syrups/jams; potato chips/corn chips/popcorn; and ice cream/sherbet/frozen yogurt in contributing to intakes of energy, fats, fatty acids, and carbohydrates. Continued high intakes of these foods are not in keeping with dietary guidance and compromise intakes of more nutritious foods. Children are influenced both by adults and by the marketplace, neither of which always present a positive nutrition role model. These same analyses for adults2 show similar findings with respect to the contribution of foods to many macro- and micronutrients. On the brighter side, these data show the continued contributions of nutritious staples such as yeast bread, meat, milk, cheese, white potatoes, ready-to-eat cereal, pasta, and rice/cooked grains toward energy and macronutrient intake in the United States. Nutritionists, health educators, and pediatricians should continue to emphasize the importance of increased intakes of grains, fruits, and vegetables in the diets of US children.
Caution is advised in using these data to rank food sources of nutrients among US children. Decisions on how to group foods have a major influence on the ranking. For example, in these data, most of the frequently consumed fruits and vegetables are their own separate food groups. Had we decided to combine all citrus fruit, or all vegetables, the relative contributions to intake of many nutrients would have increased accordingly.
As the food supply changes, these data will need to be continually updated. Of interest will be whether the increased use of low-, reduced-, or nonfat foods will have an impact on dietary sources of energy and fats. Some of the lower fat foods were just becoming popular during the late 1980s and early 1990s. Therefore, their frequency of consumption was so low that they had no impact on these findings even when lower fat varieties of foods were separated from their higher fat counterparts. In addition, as fortification of grains with folate proceeds, breads, pastas, and baked goods will be greater contributors to folate intakes. Analyses of more current data will show the impact of such foods on sources of nutrients for both children and adults.
- Received December 11, 1997.
- Accepted March 20, 1998.
Reprint requests to (A.F.S.) National Cancer Institute, Applied Research Branch, 6130 Executive Blvd, MSC 7344, EPN 313, Bethesda, MD 20892-7344.
- Subar AF,
- Block G,
- James LD
- Kant AK,
- Block G
- Murphy SP,
- Subar AF,
- Block G
- Thompson FE,
- Dennison BA
- ↵Food and Nutrient Intake by Individuals in the United States, 1 Day, 1989–91. Continuing Survey of Food Intakes by Individuals, 1989–91. Washington, DC: US Dept of Agriculture, Agricultural Research Service; NFS Report 91-2; 1995
- ↵Federation of American Societies for Experimental Biology, Life Sciences Research Office. Prepared for the Interagency Board on Nutrition Monitoring and Related Research. Third Report on Nutrition Monitoring in the United States. Vol 1. Washington, DC: US GPO; 1995
- ↵USDA Nutrient Data Base for Individual Intake Surveys. Release 7. Washington, DC: US Dept of Agriculture, Human Nutrition Information Service; computer tape. Accession number PB94-504552GEI; 1992
- ↵Nutrition and Your Health: Dietary Guidelines for Americans. 4th ed. Washington, DC: US Depts of Agriculture and Health and Human Services; Home and garden bulletin no. 232; 1995
- ↵The Food Guide Pyramid. Washington, DC: US Dept of Agriculture; Home and garden bulletin no 252; 1992
- ↵Putnam JJ, Allshouse JE. Food consumption, prices, and expenditures, 1970–1990. Washington, DC: US Dept of Agriculture, Economic Research Service; Statistical Bulletin No 840; 1992
- ↵Food and Nutrition Board. Recommended Dietary Allowances. 10th ed. Washington, DC: National Academy Press; 1989
- ↵Composition of Foods—Raw, Processed, Prepared. Washington, DC: US Dept of Agriculture; Agriculture handbook no 8; 1989 supplement; 1990
- Copyright © 1998 American Academy of Pediatrics