Objective. To compare the overall nutrient intake of American children (ages 2 to 19) who exclusively use skim milk instead of 1%, 2%, or whole milk; lean meats instead of higher-fat meats; or fat-modified products instead of full-fat products.
Study Design. A unique sorting procedure was used to categorize respondents to the 1989–1991 Continuing Survey of Food Intake by Individuals according to reported use or nonuse of certain fat-reduction strategies. Differences in intake of 23 macronutrients and micronutrients, as well as energy, by exclusive users, mixed users, and nonusers of each strategy were statistically analyzed using analysis of variance with Scheffe's test. The number of fat-reduction strategies used by the children as well as demographic characteristics also were analyzed.
Results. Only a small number of children qualified as exclusive users of skim milk (3%), lean meats (2%), and fat-modified products (1%). Energy intakes for all children were below 100% of the recommended dietary allowance. However, children (n = 85) who used skim milk exclusively in place of higher-fat milks closely approximated the current dietary recommendations (30% energy from fat, less than 10% from saturated fat, and less than 300 mg cholesterol) while maintaining adequate micronutrient intake and without significantly impacting energy. Children (n = 52) who used lean meats in place of higher-fat meats achieved the guideline for total fat; however, energy intake was 70% of the recommended dietary allowance and vitamin E was 63%. Children (n = 20) who use only fat-modified versions of cheese, salad dressing, cake, pudding, and yogurt made no significant impact on their energy, fat, or micronutrient intake. Of the 3299 children in the data set, only 3 qualified as users of two fat-reduction strategies and none qualified as users of all three strategies. Exclusive users of skim milk, lean meat, or fat-modified products were more likely to be female, white, and live in families with higher incomes. Those using skim milk or lean meat exclusively also were more likely to be older, whereas those exclusively using fat-modified products were younger. Furthermore, those using skim milk or fat-modified products exclusively were more likely to live in households where the head had more years of education, whereas those exclusively using lean meats were from households headed by those with slightly less years of education.
Conclusions. Despite the inherent limitations of population-based food surveys (including issues of underreporting, lack of biological markers and accurate anthropometric measures, and limited nutrient databases), these results provide insight into the rate of use of certain lower-fat food choices by children and suggest exclusive use can facilitate achievement of contemporary dietary recommendations. The impact of using these fat-reduction strategies on children's overall nutrient intake differs depending on the strategy used. Use of skim milk is an economical single-food strategy that facilitates achievement of contemporary dietary guidelines while maintaining nutrient adequacy. Professional guidance is recommended for children who exclusively use lean meats to assure adequate intake of energy and vitamin E. The impact of fat-modified products needs to be monitored closely as the number of such products increases in the marketplace. Results of this study can be used by health professionals working with children and their parents to highlight the overall efficacy of dietary recommendations while alerting them to potential pitfalls.
Contemporary dietary guidelines recommend that all healthy American children older than age 2 consume a diet that provides an average of 30% energy from total fat, less than 10% of energy from saturated fat, and less than 300 mg of dietary cholesterol per day.1 Current consumption of total fat (∼34%) and saturated fat (∼13%) by American children exceeds these levels.4
Concerns about the safety of lower-fat diets for American children typically focus on increased risk for inadequate micronutrient and/or energy intake. Some experts disagree with current recommendations, suggesting they are unnecessary and potentially dangerous if overzealous parents fail to provide adequate energy and micronutrients to assure optimal growth and development.9,10 However, clinical intervention studies assessing effects of lower-fat diets consistently indicate it is possible for children to maintain or even improve micronutrient intake when following a supervised low-fat diet.6,11
It is unclear whether American children (and/or their care givers) can appropriately implement a lower-fat diet without professional guidance. Several studies that have stratified children based on percent energy from fat report adequate micronutrient intakes,16 whereas others have not.19,20Lifshitz and Moses21 described eight cases of failure to thrive with unsupervised implementation of a low-fat, low-cholesterol diet. Low-fat diets may not have been the sole culprit as these diets also were insufficient in energy and micronutrients. In their review, Kennedy and Goldberg22 concluded a diet with 30% of calories from total fat and 10% from saturated fat provides for adequate growth in children if adequate energy intake is provided.
Appropriate implementation of a lower-fat diet is facilitated by translation into food choice information. Dietary recommendations commonly suggest use of lower-fat versions of milk, meats, and added fats and oils.2,23 Dwyer24 suggested that the simplest recommendation for reducing children's total fat and saturated fat intake is to substitute skim milk for whole milk. Commercial fat-modified products such as salad dressing and cheese may further reduce fat intake; however, potential benefits of fat substitutes are speculative,25 and research on the safety and efficacy of fat replacements in children's diets is needed.26
Computer modeling studies using specific lower-fat food choices have identified theoretical strategies that reduce fat intake while maintaining adequate energy and micronutrient levels for young children.27,28 Modeling also demonstrated, however, the negative impact of using multiple fat-reduction strategies. These theoretical menu modifications may not reflect actual consumption patterns of children using lower-fat food choices.
No studies have investigated food choices of American children to assess the impact of specific lower-fat food choices on overall nutrient intake. It is unclear how adoption of common lower-fat food choices by this population affects nutrient intake. Three scenarios are possible: 1) children (and/or their care givers) who choose lower-fat foods could report significantly lower intake of total fat and saturated fat while maintaining adequate energy and micronutrient intakes; 2) children could report lower-fat intake but inadequate energy and micronutrient intakes; or 3) children could compensate for using lower-fat food choices and, therefore, would not report lower-fat intakes with use of lower-fat food choices.
The purpose of this study was to investigate the impact of specific low-fat food choices of free-living American children by comparing the overall nutrient intake of children who used skim milk versus 1%, 2%, or whole milk; lean meats versus higher-fat meats; and fat-modified products versus full-fat products. Overall nutrient intake of children who used any fat-reduction strategy or multiple fat-reduction strategies was compared with children who used none of the strategies. This sorting reflects previous theoretical models and is designed to determine what happens when children consume the lowest-fat products available in each food grouping.
The Continuing Survey of Food Intake by Individuals (CSFII) for the years 1989, 1990, and 1991 were the data sources used. This nationally representative dietary intake survey uses a multistage area probability sampling procedure, a complex survey design described elsewhere.29 Demographic and food data for each individual are available in different record groupings (eg, by specific foods, by aggregate food groups, by specific nutrients, by demographics, and so forth). These records can be linked, in most instances, to the individual, allowing for a variety of analytical manipulations. CSFII data for these years were collected by trained interviewers using one 24-hour recall followed by food records for 2 consecutive days. Diet recalls and records for very young children typically were provided by their mothers. Although some children may provide information about their own intake, the percentage doing so is unknown. Despite the detailed sampling criteria and complex design, there are inherent limitations to collecting data and interpreting findings from nutrition monitoring surveys. For example, precise anthropometric and biological measurements are not available to confirm the accuracy of dietary intake reports. Furthermore, issues of underreporting are a concern. However, nutrition monitoring studies do provide insight into food consumption patterns and are useful in describing trends and answering general dietary intake questions.
The nutrient databases for the CSFII are the United States Department of Agriculture (USDA) Nutrient Data Base for Standard Reference30 and the USDA National Nutrient Data Bank.31 Most database values are obtained from laboratory analysis. Those not available from laboratory analysis were imputed from data for other forms of that food or from data for similar foods.
Only records representing all 3 days of intake were used (approximately two-thirds of the sample). In accordance with CSFII definitions, children were defined as “those greater than 1 and less than 20 years of age.” For each survey year, food codes were used to sort respondents by certain lower-fat and higher-fat food choices, ie, type of milk used (skim vs higher-fat milks), type of meats used (lean vs medium-fat and high-fat),23 and type of cheese, yogurt, salad dressing, cake, and pudding (hereafter referred to as full-fat products or fat-modified products). These food choices represent common dietary recommendations for the general public and have been analyzed previously on a theoretical basis.27
To determine the full impact of using such food choices, a unique and rather restrictive sorting procedure was developed. To be considered a lean meat user, a child must have eaten only lean meats across the 3 days of intake and no medium-fat or higher-fat meats. Thus, any reported intake of ground beef across the 3 days would remove them from the lean meats group. Similarly, to be included in the fat-modified products group, a child must have consumed only fat-modified versions of all five products (or not eaten any of these products) for all 3 days. Sorting CSFII respondents in this manner required the creation of new data subsets from the original CSFII data before analysis. Once sorted, individual food records containing daily nutrient intake for each child were located and analyzed. Detailed methods have been described elsewhere.32
In addition to comparing exclusive users, mixed users (ie, some use of lower-fat and some use of higher-fat foods), and nonusers of each fat-reduction strategy, overall number of strategies used also was analyzed. Single strategy users included: 1) exclusive use of skim milk but not lean meats or fat-modified products; 2) exclusive use of lean meats but not skim milk or fat-modified products; 3) exclusive use of fat-modified products but not skim milk or lean meats. Multiple strategy users included: 1) exclusive use of skim milk and lean meats; 2) exclusive use of skim milk and fat-modified products; 3) exclusive use of lean meats and fat-modified products; or 4) exclusive use of skim milk, lean meats, and fat-modified products. Users of no strategies included those who used 1%, 2%, or whole milk; those who used medium-fat or high-fat meats; and/or those who used full-fat versions of cheese, yogurt, salad dressing, cake, or pudding.
Analysis of 3-day mean nutrient intakes included: total fat, saturated fat, monounsaturated fat, polyunsaturated fat, cholesterol, sodium, energy, protein, carbohydrate, vitamin A as retinol equivalents, vitamin C, iron, calcium, zinc, adjusted vitamin B-6,33magnesium, vitamin E, vitamin B-12, thiamin, riboflavin, niacin, folate, potassium, and phosphorus. Although evaluated in other studies, vitamin D, selenium, manganese, copper, and sugars were not evaluated in this study due to their exclusion from the CSFII database.
Because nutrient intake may be influenced by total energy intake, nutrient density (intake per 1000 kcal) is commonly used to minimize the impact of energy on nutrient intake. In this study, nutrient density was assessed by dividing 3-day mean nutrient intakes by 3-day mean energy intakes, then multiplying by 1000.34
Age-related percent recommended dietary allowances (RDAs) (capped at 100) also were assessed for energy, protein, vitamin A, vitamin C, iron, calcium, zinc, vitamin B-6, magnesium, vitamin E, vitamin B-12, thiamin, riboflavin, niacin, folate, and phosphorus. RDAs were capped at 100% to account for the dilution effect of intakes exceeding 100%. To be considered nutritionally adequate, the group means had to achieve at least 67% of the RDA for each vitamin and mineral,35and 100% of the RDA for energy. An overall mean adequacy ratio (MAR) (defined as the total of the 16 capped percent RDAs divided by 16) also was calculated as a measure of overall nutritional adequacy.
To allow for comparison of results with dietary recommendations, percent contribution to energy intake of total fat and saturated fat were assessed, as were cholesterol and sodium intakes. To be considered congruent with guidelines, group means had to achieve 30% or less energy from fat,2 less than 10% energy from saturated fat,2 300 mg or less dietary cholesterol,2 and 2400 mg or less sodium.36
Calculations to determine the exponents that most closely approximate a normal distribution for each nutrient were used (courtesy of Dr Alicia Carriquiry, Statistics Lab at Iowa State University). In accordance with USDA guidelines, weighted data were used for descriptive statistics and unweighted data were used for inferential statistics.37
Analysis of variance with Scheffe's test was used to statistically analyze differences in nutrient intake between pure users, mixed users, and nonusers of each strategy or combined strategies. A Pvalue of .01, which is more conservative than standard practice, was used to determine significant relationships in an attempt to compensate for the design effect of this large, complex survey.
Of the 3299 children in the data set, few qualified as exclusive users of skim milk, lean meats, or fat-modified products (Table1). Of the milk users, 3% (n = 85) used only skim milk. Of the meat users, 2% (n = 52) used only lean meats. Of the users of cheese, yogurt, salad dressing, cake, and pudding, 1% (n = 20) used only fat-modified versions of these products. Given the low usage rates of any strategy, further nutritional analyses for specific age or gender groupings could not be performed.
Children who used only skim milk were more likely to be female, older, white, and live in families with higher income and more years of education for the head of household when compared with children who used higher-fat milks (Table 1). Children who used only lean meats also were more likely to be female, older, white, and live in families with higher income when compared with children who used higher-fat meats; however, the head of household for children who used lean meats reported slightly less years of education than for children who used higher-fat meats. Children who used only fat-modified products were more likely to be female, younger, white, live in families with higher income, and slightly more years of education for the head of household when compared with children who used full-fat products.
Only 3 children who used fat-reduction strategies were exclusively using two of the three fat-reduction strategies, and none reported use of all three. Thus, comparisons involving multiple strategy users were not possible. Combining single strategy users with multiple strategy users is not meaningful because this results in a reanalysis of single strategy use. Therefore, further discussion of strategy use will include milk users, meat users, and product users but will not include comparisons of number of strategies used.
No mean values for any group (including those who used higher-fat milks, higher-fat meats, and full-fat products) achieved 100% of the RDA for energy (Fig 1, Fig2, and Fig 3). Exclusive use of skim milk or fat-modified products had no significant impact on total energy intake or percent RDA for energy. However, children who used only lean meats had a significantly lower mean percent RDA for energy (P < .0001) than those who used mixed meats or higher-fat meats (Fig 2).
Children who used only skim milk had significantly lower intakes of total fat (P < .0001), saturated fat (P < .0001), and cholesterol (P < .0001) when compared with children who used higher-fat milks (Table 2). Using only skim milk had no significant impact on total intake of protein, carbohydrate, vitamins, or minerals. All groups of milk users exceeded 67% of the RDA for vitamin A, calcium, and magnesium and 11 other vitamins and minerals, when nutrients were expressed as percent RDA; although exclusive use of skim milk resulted in statistically significant lower values for vitamin A, calcium, and magnesium when compared with users of mixed milks but not when compared with users of higher-fat milks (Fig 1).
Children who used only lean meats had a significantly lower intake of total fat (P < .0001), saturated fat (P < .0001), and cholesterol (P < .0004) when compared with users of mixed meats and users of higher-fat meats (Table3). Although exceeding two-thirds of the RDA, it seems that total intake of thiamin, potassium, and phosphorus was statistically significantly lower for users of lean meats when compared with users of mixed meats but not when compared with users of higher-fat meats. However, when expressed as nutrient density, intake of each of these nutrients was greater for those who used lean meats (data not shown). Based on percent RDAs, users of only lean meats had statistically significant lower values for protein, thiamin, vitamin E, magnesium, phosphorus, and the MAR when compared with mixed-meat users but not when compared with higher-fat meat users (Fig 2). It should be noted that although percent protein was statistically lower in lean-meat users, it was still more than 67% of the RDA, whereas percent vitamin E for lean-meat users did not reach 67% of the RDA.
Use of only fat-modified versions of cheese, yogurt, salad dressing, cake, or pudding had no significant impact on overall nutrient intake except for a statistically significant lower vitamin A intake when compared with users of mixed products (Table4). Based on percent RDAs, micronutrient intakes were not significantly affected by use of fat-modified products, with all exceeding 67% regardless of whether fat-modified or full-fat products were used (Fig 3).
Children who used skim milk or lean meats achieved dietary recommendations for total fat (Fig 4A). Users of skim milk closely approximated the guideline for saturated fat (Fig4B). All groups achieved 300 mg or less for cholesterol intake regardless of food choices. All groups exceeded the 2400 mg cutoff for sodium intake except for users of only fat-modified products.
Very few children were identified as exclusive users of skim milk, lean meats, or fat-modified products and no child was identified as a multiple fat-reduction strategy user. Demographic characteristics of children who reported use of lower-fat food choices in this study are similar to those reported previously.38 The fact that so few children qualified as multiple strategy users should alleviate concern that a large portion of American parents are being overzealous in their attempt to reduce children's fat intake. The sorting procedure in this study was rather restrictive to evaluate the full impact of using fat-reduction strategies; therefore, food choices of a child who qualified as a multiple strategy user would be very limited. In addition, the sorting procedure used required use or nonuse of fat-modified products for all five types of foods for each of the 3 days when those products were consumed.
Those who express concern about the micronutrient adequacy of a lower-fat diet for children should find these data intriguing. Each of the three proposed scenarios were evident in this study. Adequate micronutrient intake was reported for children who achieved 30% of energy from total fat and 10.7% of energy from saturated fat by using skim milk in place of higher-fat milks. Although energy intake was 76.2% of the RDA, it was not significantly lower for children who used skim milk. Thus, skim milk users were making nutrient dense, lower-fat food choices. These results are congruent with both theoretical27 and observational studies.39 When the diets of 4- to 7-year-old Latino children in New York City were divided into quintiles based on percent energy from saturated fat, frequent consumption of whole milk was the single most important food choice that distinguished children with high- versus low-saturated-fat intakes. In this population, substitution of 1% milk for whole milk reduced saturated-fat consumption to recommended levels for three of the five groups. Because milk is consumed frequently by children, use of skim milk (and probably 1% milk as well) in place of higher-fat milks offers great potential for reductions in fat intake without negatively affecting micronutrient intake.
Children who used lean meats reported adequate micronutrient intake, except for vitamin E (64.1%), but statistically significant lower energy intake (70% of RDA), while achieving 30.2% of energy from total fat and 11.7% from saturated fat. It should be noted that no group reported consumption of 100% of the RDA for vitamin E. This is consistent with previous studies of children on lower-fat diets.11,17 High carbohydrate diets consumed by children in the Bogalusa Heart Study also were significantly lower in vitamin E.40 Professional guidance is necessary when children exclusively use lean meats. Although mixed-meat users were further from achieving dietary goals than were lean-meat users, their overall profile was the most nutrient dense.
Children who used only fat-modified versions of cheese, yogurt, salad dressing, cake, and pudding made no significant impact on energy or micronutrient intake and did not achieve dietary recommendations for either total fat or saturated fat. In fact, percent energy from fat for users of fat-modified products was higher than that of mixed-products users; although total fat (g) was lower. Experts suggest use of fat-modified products may be associated with a net decrease in fat intake and net increase in carbohydrate and protein intake, resulting in reductions in percent energy from fat with no significant reductions in 24-hour energy intake.41 In this study it seems children who used fat-modified products may represent those who compensate for lower-fat food choices. However, it should be noted that only 20 children used this strategy; therefore, caution regarding generalization of this finding should be taken.
As stated previously, energy intake did not reach 100% of the RDA for any child, regardless of strategy use or nonuse. This finding is consistent with the overall mean of 88% of the RDA (1781 kcal) obtained by Lin and Guthrie using the same CSFII 1989–1991 data.8 These investigators also showed energy intakes did not reach 100% of the RDA for a number of age/gender groups. It also is consistent with dietary intakes from other studies of children.4,5,44 Despite static or declining energy intake for children since the 1970s, Kennedy and Goldberg22concluded that growth in American children seems adequate. One possible explanation for low-energy intake is underreporting. Underreporting of dietary intake is quite common,45 particularly when 24-hour recalls are used. In the CSFII, parents complete diet records for young children, and parents who report intake for their children tend to underreport.50,51 Another explanation specific to the present study, may be the use of capped energy values during analysis which was done to prevent dilution of the means by those who greatly exceeded 100%. Because this study demonstrates that energy intakes for those children exclusively consuming lean meats are significantly lower than for children consuming some higher-fat meats, it becomes even more critical to closely monitor these children.
Additional limitations of this study are related to limitations of dietary assessment in general. All dietary assessment methods assume participants are able to remember what and how much they ate, have some knowledge of common weights and measures, and have the desire to accurately report such information.52 Even if the information is accurate, participants may modify their eating patterns during the survey period and therefore data may not truly represent respondents' typical food intake. Inaccuracies in databases used to analyze intake also are possible.
Furthermore, there is the possibility of misclassification with use of the food choice-based sorting procedure used in this study. Portion size of individual foods was not considered when categorizing people as users or nonusers. Thus, one report of skim milk and no other reports of milk use across the 3 days would classify a person as a skim milk user, whereas use of skim milk at every meal for 3 days also would classify a person as a skim milk user.
Selecting only those who reported all 3 days' intake reduced the sample size of potential users and nonusers by approximately one third. However, use of one 24-hour recall to estimate usual intake of individuals is not recommended due to high within-person variability.52,53 Neither precise anthropometric measurements nor physiologic markers were available to assess accuracy of dietary reports in relationship to physical growth and development of children who used fat-reduction strategies. Finally, the low number of exclusive users of any strategy precluded analysis by age or gender. Although exclusive users are older, the impact of use may vary depending on the child's age. Thus, utilization of the CSFII allows for analysis of a large nationally representative sample, but requires acknowledgment of methodologic weaknesses inherent to the study design.
None of these limitations, however, diminish the findings of this study that suggest use of certain lower-fat food choices can facilitate achievement of contemporary dietary recommendations. However, the impact of using lower-fat food choices on children's overall nutrient intake differs depending on the strategy used. American children who use skim milk in place of higher-fat milks can closely approximate dietary recommendations while maintaining adequate micronutrient intake without significantly impacting energy intake. Thus, use of skim milk is an economical, single-food strategy that facilitates achievement of contemporary dietary guidelines while maintaining nutrient adequacy. Children who use lean meats in place of higher-fat meats can achieve the dietary guideline for total fat; however, intakes of energy and vitamin E need to be closely monitored. Professional guidance is recommended for children who choose only lean meats. Children who use only fat-modified versions of cheese, salad dressing, cake, pudding, and yogurt may make no significant impact on their energy, fat, or micronutrient intake. However, as the number and types of fat-modified products has greatly expanded since 1991, the impact of their use needs to be closely monitored. Results of this study can be used by health professionals working with children and their parents to highlight the overall efficacy of dietary recommendations while alerting them to potential pitfalls.
We thank Cheryl Achterberg, Fern Willits, Penny Kris-Etherton, and Sharon O'Donnell for their contributions to this project.
- Received November 12, 1996.
- Accepted March 12, 1997.
Reprint requests to (M.S.-G.) Pennsylvania State University, Department of Food Science, 203A Borland Lab, University Park, PA 16802.
- CSFII =
- Continuing Survey of Food Intake for Individuals •
- USDA =
- United States Department of Agriculture •
- RDA =
- recommended dietary allowances •
- MAR =
- mean adequacy ratio
- American Academy of Pediatrics, Committee on Nutrition
- ↵United States Department of Agriculture. Nutrition and Your Health: Dietary Guidelines for Americans. 4th ed. Washington, DC: United States Department of Agriculture, United States Department of Health and Human Services; 1995. Home and Garden Bulletin No. 232
- National Cholesterol Education Program. Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Children and Adolescents. Bethesda, MD: National Institutes of Health; 1991. NIH Publication No. 91–2752
- ↵McDowell MA, Briefel RR, Alaimo K, et al. Energy and Macronutrient Intakes of Persons Ages 2 months and Over in the United States: Third National Health and Nutrition Examination Survey, Phase 1, 1988–1991. Advance data from vital and health statistics, No. 255. Hyattsville, MD: National Center for Health Statistics; 1994
- Nicklas TA,
- Webber LS,
- Srinivasan SR,
- Berenson GS
- Lin B-H,
- Guthrie J
- Olson RE
- McKenzie J,
- Dixon LB,
- Smiciklas-Wright H,
- Mitchell D,
- Shannon B,
- Tershakovec A
- The Writing Group for the DISC Collaborative Research Group
- McPherson RS,
- Nichaman MZ,
- Kohl HW,
- Reed DB,
- Labarthe DR
- Shea S,
- Basch CE,
- Stein AD,
- Contento IR,
- Irigoyen M,
- Zybert P
- Nicklas TA,
- Webber LS,
- Koschak ML,
- Berenson GS
- ↵American Diabetes Association, Inc and American Dietetic Association.Exchange Lists for Meal Planning. Chicago, IL: American Dietetic Association; 1989
- Sigman-Grant MJ,
- Zimmerman SA,
- Kris-Etherton PM
- Magarey A,
- Daniels K,
- Boulton J
- ↵Human Nutrition Information Service. Nationwide Food Consumption Survey, Continuing Survey of Food Intake by Individuals: 1989–1991. Washington, DC: US Department of Agriculture; 1992–1994. Study No. 09010–073
- ↵United States Department of Agriculture, Human Nutrition Information Service. USDA Nutrient Database for Standard Reference, Release 9. Springfield, VA: US Department of Commerce, National Technical Information Service; 1985. Accession No. PB90–502717. Computer tape
- Hepburn, FN
- ↵Peterson S, Sigman-Grant M. Development of a food code level sorting procedure for CSFII data. Fam Econ Nutr Rev. In press
- Guthrie HA,
- Crocetti AF
- ↵Gibney M, Sigman-Grant M, Stanton J, Keast D. Consumption of sugars.Am J Clin Nutr. 1995;62(suppl):178S-194S
- ↵Committee on Diet and Health, Food and Nutrition Board. National Research Council. Diet and Health Implications for Reducing Chronic Disease Risk. Washington, DC: National Academy Press; 1989
- ↵Kott PS, Gray BC, Guenther PM. Guidelines for the Use of Weights When Analyzing and Reporting HNIS Survey Data: October, 1989. Hyattsville, MD: Human Nutrition Information Service, US Department of Agriculture; 1989
- Nicklas TA,
- Myers L,
- Farris RP,
- Srinivasan SR,
- Berenson GS
- Birch LL,
- Johnson SL,
- Jones MB,
- Peters JC
- Freudenheim JL
- ↵Gibson RS. Principles of Nutritional Assessment. New York, NY: Oxford University Press; 1990
- ↵Sempos C, Looker A, Johnson CL, et al. The importance of within-person variability in estimating prevalence. In: McDonald I, ed.Monitoring Dietary Intakes. New York, NY: Springer-Verlag; 1991:99–109
- Copyright © 1997 American Academy of Pediatrics