QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dixon, L. B. Articles by Krebs-Smith, S. M. Search for Related Content PubMed PubMed Citation Articles by Dixon, L. B. Articles by Krebs-Smith, S. M.

---

Supplement

Let the Pyramid Guide Your Food Choices: Capturing the Total Diet Concept¹

Division of Cancer Control and Population Sciences, National Cancer Institute, National Institutes of Health, Bethesda, MD and ^* Nutrition Analyst/Consultant, Washington, DC

²To whom correspondence should be addressed. E-mail: ld120i{at}nih.gov.

	ABSTRACT

TOP ABSTRACT INTRODUCTION ASSESSING THE TOTAL DIET:... STRATEGIES FOR IMPROVEMENT REFERENCES

 
This paper discusses how the guideline "Eat a variety of foods" became "Let the Pyramid guide your food choices," presents background information on the food guidance system upon which the Food Guide Pyramid is based and reviews methods that have been used to assess aspects of the total diet, i.e., the variety, moderation and proportionality, promoted by this guidance. The methods include measures of dietary variety, patterns based on Pyramid food group intakes and scoring methods comprised of multiple dietary components. Highlights of results from these methods include the following. Although approximately one third of the U.S. population eat at least some food from all Pyramid food groups, only 1–3% eat the recommended number of servings from all food groups on a given day. Fruits are the most commonly omitted food group. Vegetables and meat are the groups most commonly met by adults, and dairy the most commonly met by youth. Intakes of specific types of vegetables (i.e., dark green, deep yellow) and of grains (i.e., whole grains) are well below that recommended; intakes of total fat and added sugars exceed current recommendations. Scoring methods show those diets of the majority of the population require improvement, and that diets improve with increases in education and income. This paper also discusses the limitations and strengths of these approaches, and concludes with suggestions to improve current food guidance and methods to assess the total diet.

KEY WORDS: • dietary guideline • dietary pattern • dietary quality • dietary variety • Food Guide Pyramid

	INTRODUCTION

TOP ABSTRACT INTRODUCTION ASSESSING THE TOTAL DIET:... STRATEGIES FOR IMPROVEMENT REFERENCES

 
"Eat a variety of foods" was the first guideline in the Nutrition and Your Health: Dietary Guidelines for Americans, from their inception in 1980 through the fourth edition in 1995 (Table 1 ). Targeting nutrient adequacy, it was meant to convey that no single food supplies all of the nutrients and other food components required for optimum health. In all four editions, to ensure variety, Americans were encouraged to consume foods from among each of several major food groups—breads, cereals, rice and pasta; vegetables; fruits; milk, cheese and yogurt; meat, poultry, fish, dry beans, eggs and nuts. In the 1985 and 1995 editions, Americans were also encouraged to consume different foods within those groups. Variety was quantified in the text of the guideline in 1990 by including recommended ranges for the number of servings from five food groups (grains, vegetables, fruits, milk, and meat or meat alternates). These recommendations were from a food guidance system developed by the USDA (1 ,2) , which, in 1992, was published as the Food Guide Pyramid (referred to in this paper as the Pyramid) (3) . The Pyramid was included in the text of the 1995 Dietary Guidelines. From 1980 through 1995, the guideline statement remained, "Eat a variety of foods."

Although the Pyramid has been featured in the Dietary Guidelines as a way to attain nutrient adequacy, the underlying food guidance system was designed to do much more (1 ,2 ,8) . The system was based on a series of guiding principles that aimed to achieve the following:

provide a framework for a diet to meet the needs of most healthy Americans 2 y old, based on current nutrition research and dietary recommendations

feature foods commonly eaten by Americans, classified into easily recognizable food groups, and allowing maximum flexibility in food choices

cover the range of energy needs among the population by including a range of recommended intakes

provide nutrient adequacy without dependence on dietary supplements or highly fortified foods

moderate food components for which overconsumption was a concern (energy, total fat, saturated fat, cholesterol, added sugars, sodium and alcohol)

account for all foods and beverages consumed as part of meals and snacks. Inherent in this idea is that any food can be part of nutritious diet, although each choice has implications.

be evolutionary. The food guidance system was built on previous food guides but considered current research directions.

The most distinguishing feature of the food guidance system was that, in addition to promoting nutrient adequacy, it provided proportionality and moderation by accounting for all foods consumed—the "total diet"concept.⁴ This was a departure from the "foundation diet" concept used by a previous food guide, "Food for Fitness-A Daily Food Guide " (9) , popularly referred to as the Basic Four or the Four Food Groups (10) . The Basic Four was designed to provide 1200 kcal and 80% of eight nutrients for which Recommended Dietary Allowances (RDA) existed in 1953 (11) . The four food groups, fruit and vegetables, milk, meat, breads and cereals, targeted specific nutrients often low in the diets of Americans. The fruit and vegetable group targeted vitamins A and C, the milk group targeted calcium, and the meat and meat alternates group targeted micronutrients that were difficult to obtain from other food groups. It was assumed that individuals would consume more food than the guide recommended to satisfy energy needs. These less nutrient-dense foods would supply additional nutrients required to bring diets close to the levels of the RDA.

By the 1970s, many scientists recognized that dietary guidance should not only target nutrient adequacy but also provide guidance related to moderation of those dietary components that were being consumed excessively. In 1979, the USDA released the "Hassle-Free Guide," a minor revision of the Basic Four that included a fifth group called "Fats, Sweets, and Alcohol," but did not quantify recommendations for that group (13) . Like the Basic Four, the Hassle-Free Guide suggested amounts of foods to provide a foundation diet.

In contrast, the food guidance system developed in the 1980s and used today suggests amounts of foods for a total diet. To cover the range of energy needs of the population, a range of servings from each major food group is recommended. To achieve nutrient adequacy, certain subgroups within the food groups are emphasized. Although not illustrated in the Pyramid graphic, the text in the Pyramid brochure (3) recommends that dark green leafy vegetables and legumes be included in the diet several times a week, and at least three or more servings from the breads and cereals group be whole grain.

Moderation of fat and lower energy intakes become possible by adjusting the amounts of discretionary fat and added sugars. Discretionary fat includes amounts of fat above that consumed if the lowest fat choices were made in all the food groups (e.g., amount of fat in 2% milk above the amount of fat in skim milk). Added sugars represent all caloric sweeteners, such as table sugar, high fructose corn syrup, and honey, added to foods during processing or preparation, or eaten separately. The system allows for the inclusion of modest amounts of fat and added sugars within an individual’s energy needs. As was the case with the subgroup recommendations, detailed suggestions of amounts of fat and added sugars are included in the Pyramid brochure, but not in the Pyramid graphic.

The food guidance system was designed to provide the recommended amounts of essential nutrients without depending on the use of supplements or highly fortified foods (2) . This was consistent with the recommendations that nutrient needs should be met by the consumption of a variety of foods rather than from supplements (14) . Recently, there has been the recognition of the need for supplementation to meet the nutrient recommendations for some segments of the U.S. population. The first edition of the Dietary Guidelines stated that "you rarely need a vitamin or mineral supplement if you eat a variety of foods" (15) . The most recent edition states "some people need a vitamin-mineral supplement to meet specific needs" (15) . It places special emphasis on foods rich in folate or folic acid supplements to reduce the risk of certain birth defects. However, because foods contain beneficial substances in addition to nutrients, it cautions against depending on dietary supplements to meet usual needs.

To promote the Pyramid only as a means to achieve nutrient adequacy is to ignore some of its most promising aspects. Unfortunately, the Pyramid, although first released as a brochure with detailed information, is most often depicted as a graphic showing quantified recommendations from only the major food groups (Fig. 1 ). As such, it has become something of an icon. It is not obvious, even to a serious observer, that this simple scheme represents an entire guidance system that incorporates suggestions on the consumption of specific types of foods within the major food groups, intakes of fat and added sugars, serving sizes and energy levels. Nonetheless, "Let the Pyramid guide your food choices" is unique among the Dietary Guidelines in that it considers multiple aspects of the diet simultaneously. The themes of variety, moderation and proportionality—in sum, the total diet—have been inherent in this guideline, although to varying degrees over time.

View larger version (78K): [in this window] [in a new window]   Figure 1. The food guide pyramid (3) .

	ASSESSING THE TOTAL DIET: MEASURES AND RESULTS

TOP ABSTRACT INTRODUCTION ASSESSING THE TOTAL DIET:... STRATEGIES FOR IMPROVEMENT REFERENCES

Variety measures

Initial attempts at capturing the total diet were related to the original guideline, "Eat a variety of foods," and involved measuring dietary variety in different ways. Dietary variety, or dietary diversity as it is sometimes called, refers to the inclusion of different items in the diet. Krebs-Smith et al. (6) measured variety among and within food groups from 24-h recall data in the 1977–1978 Nationwide Food Consumption Survey (NFCS). They assessed whether these variety measures were related to variety’s purported benefits of dietary quality, specifically nutrient adequacy and the intakes of energy, fat, sugar, cholesterol and sodium. The authors controlled for the number of foods to measure the effects of variety per se, apart from the additional food that a varied diet often entails. Variety among the major Pyramid food groups (counted as the number of groups present in the diet) explained as much variation in nutrient adequacy as variety within those groups (counted as the number of subgroups or as the number of individual foods within the groups). Neither type of variety was related to intakes of energy, fat, sugar, sodium or cholesterol. This study suggested that variety might best be interpreted as choosing the recommended number of servings from each group. Telling consumers this directly might be an improvement over the ambiguities of "Eat a variety of foods." The finding that there is little nutritional advantage to increasing variety within the food groups is consistent with the fact that certain foods within each group are more nutrient dense than others. This suggests that choosing foods simply because they are different from one another may not be as important as selectively including more of some foods and less of others.

Other studies have examined dietary variety or diversity among food groups. Kant et al. (5) created two different scores from 24-h recall data in the Second National Health and Nutrition Examination Survey (NHANES II) to measure dietary diversity. A food group score gave a maximum score of 5 to adults who reported at least one food, above a minimum gram weight amount, from each of five food groups (dairy, meat, grain, fruits, vegetables). A serving score gave a maximum score of 20 to adults who reported at least two servings each from dairy, meat, fruit, and vegetable groups and four servings from the grain group. Serving sizes were based on the median gram weights of each food reported. Using the food group score of dietary diversity, only one third of the adult population consumed foods from all five food groups on a given day. Using the serving score of dietary diversity, <3% consumed the designated minimum number of servings from all food groups on a given day.

Researchers have also examined whether variety among food groups is linked to health outcomes. Using the 5-point food group score from 24-h recall data in the NHANES I Epidemiologic Follow-Up Study, Kant et al. (16) reported an association, among adults, between the consumption of two or fewer food groups and increased mortality. Lower scores, from similar scoring systems using food-frequency data, have also been associated with increased risk of mortality or chronic disease (17 18 19 20) .

A recent study (7) measured within food group variation using food frequency data from 71 healthy adult men and women. The authors examined whether this type of dietary variety was related to energy intake and body fatness. In this study, variety was defined as "the percentage of different food types within each of 10 food groups, regardless of the frequency with which they were consumed." Results showed increased variation within all food groups to be positively associated with energy intake. Increased variation within certain food groups (e.g., sweets, snacks, condiments, entrees and carbohydrates) was positively associated with body fatness, but increased variation within the vegetable group was negatively associated. Variety ratios of the vegetable group to other food groups were negatively associated with body fatness, even after adjustment for dietary fat.

Dietary variety among or within food groups is related to the total number of foods in the diet (i.e., increases in variety will necessitate increases in food intake), and the quantity of foods is related to nutrient adequacy (i.e., more foods will increase nutrient intakes). Accordingly, many studies of variety have controlled for the number of foods and/or energy intake. However, variety scores do not take into account portion size, beyond a minimum amount, or otherwise quantify the intake of each group. Although variety captures the presence or absence of different food groups, it does not capture them as true "dimensions" of the diet because their quantities are not assessed. Furthermore, the variety score tallies only the number of food groups but does not indicate which groups are included. For example, a score of 3 merely indicates the presence of three food groups, not which groups they are or how much of any group is present. Last, these scores are based on the assumption that eating foods from all food groups is preferable and are biased against diets that do not include certain classes of foods such as meat or milk, which could also be nutritionally adequate.

Quantifying the individual dimensions of the diet

Quantification of the different dimensions of the diet became possible with the development of the Pyramid Servings Database (PSDB) for use with the USDA’s Continuing Survey of Food Intakes by Individuals (CSFII) (21) . This database provides the number of servings of each of the Pyramid’s major food groups and subgroups, and the amounts of discretionary fat and added sugars contained in 100 g of every food mentioned in the survey. It required the development of a recipe file to disaggregate food mixtures into their component ingredients or foods before assigning the components to food groups. Because 75% of the foods reported in the 1989–1991 CSFII were mixtures, this task was especially challenging and required many difficult decisions as to the level of disaggregation (Fig. 2 ) and the choice of appropriate food groups. In addition, food intakes were reported in grams and converted to serving sizes, and new variables were developed to quantify discretionary fat and added sugars. In effect, the PSDB’s exacting system of operationalizing the recommendations of the Pyramid brings to light many of the ideas that were emphasized in the original food guidance system and the brochure, but lost in the widespread use of the graphic icon.

View larger version (27K): [in this window] [in a new window]   Figure 2. Example of a cascaded recipe file showing multiple levels of disaggregation (adapted from Ref. 21 ).

Table 2A

Patterning techniques

Measures of patterns of Pyramid food group intakes have built upon earlier measures of variety. For example, Kant et al. (23) evaluated 24-h recall data in the NHANES II for the presence or absence, according to a minimum gram weight amount, of five food groups (dairy, meat, grain, fruit, vegetable). Each adult’s pattern was then determined on the basis of whether the individual did/did not report consuming at least a minimum amount of food from each of these groups. Two categories for each of five food groups generated 32 possible patterns. The most prevalent patterns among the population were as follows: all food groups present (34%), no fruit (24%), no dairy or fruit (9%), no dairy (8%), and no fruit or vegetable (6%). The pattern with all food groups was associated with mean nutrient intakes above the RDA. This pattern also had the lowest proportion of the population with mean nutrient intakes below the respective RDA. Patterns with fruit and vegetable intakes were also associated with higher concentrations of serum vitamin C.

Krebs-Smith et al. (24) characterized patterns of adults in a similar way, but was able to quantify the number of servings of Pyramid food groups by combining dietary data from the 1989–1991 CSFII with the PSDB. Each person’s servings were compared with the recommended number of servings of each food group that corresponded to their reported energy intakes. Each person’s pattern was then determined on the basis of whether they met/did not meet the recommendation for each of the five major food groups. Of the 32 possible patterns, six represented 44% of the population. The most prevalent pattern (11%) was meeting the recommendations only for vegetables and meat. Only 1% of the adults met the recommendations for all five food groups. Such patterning techniques involve simultaneously assessing multiple food groups and allowing relationships among the various dietary dimensions to emerge. For example, the pattern of meeting all five food groups was associated with intakes of dietary fat, added sugars and micronutrient intakes all in accordance with recommendations. All other patterns were associated with micronutrient and/or fiber intakes below the recommended amounts, intakes of fat and/or sugars above the recommended amounts or some combination. All five patterns that averaged 20 g of fiber met the grain recommendation and one or both of the fruit and vegetable recommendations. Similarly, only five patterns had 30% energy from fat, and all of them met the fruit recommendation.

Similar pattern analyses were conducted on the diets of children and adolescents, ages 2 to 19 y, who participated in the 1989–1991 CSFII (25 ,26) . The most prevalent food pattern was meeting dairy only (12%). The next most prevalent food pattern was not meeting any of the food group recommendations (11%). Almost 40% of children and adolescents had patterns of meeting none or only one of the recommendations. Only 2% met all recommendations; only 10% met four or five of the recommendations. Those who met all food group recommendations had micronutrient intakes above the recommended amounts but also had intakes of fat and added sugars above the recommended amounts. Thus, in contrast to findings among adults, meeting the major food groups recommendations was not necessarily sufficient to ensure a healthful total diet.

Scoring methods

Scoring methods that incorporate aspects of the Pyramid have been developed to assess the total diet. One scoring method is the Healthy Eating Index (HEI) (27) . The HEI is a 10-component index, i.e., five components measure how diets conform to the Pyramid food group servings of grains, vegetables, fruits, milk and meat, and the other five components include intakes of total fat, saturated fat, cholesterol, sodium and a measure of dietary variety (Table 3 ). The scoring system totals 100 points with each of the components having a score that ranges from 0 to 10. Scores of 0 are assigned if no foods are consumed or if nutrient intakes are above the recommended amounts; scores of 10 are assigned if the recommended amounts of foods or nutrients were consumed; intermediate scores are assigned proportionately to the amounts of food or nutrient intakes.

Both scoring methods have been used to assess the diets of Americans and have produced similar results. In 1994–1996, the mean HEI score was 63.6 (30) . In 1994, the mean DQI-R score was 63.4 (29) . In both analyses, 18% of the population had scores <51, indicative of a poor diet, and 12% had scores >80, indicative of a healthy diet. For both indices, less than half of the population met the recommendations for any of the food groups, or for total fat, saturated fat and sodium. More than half of the population met the recommendations for dietary cholesterol. Both scoring methods correlated positively with nutrient intakes, with higher HEI and DQI-R scores associated with higher nutrient intakes in relation to the RDA. Higher HEI and DQI-R scores were also associated with being female, and having higher education or income levels (28 ,30) .

Both scoring methods have also been used to evaluate dietary change in the population. In 1994–1996, higher percentages of individuals met several of the recommendations related to components of the HEI compared with 1989–1991 data (in particular, total and saturated fat) (Table 4 ). Similar findings were observed for components of the DQI. However, change in mean scores across the surveys was very small, suggesting that although Americans improved some aspects of their diets, other aspects of their diets remained unchanged or worsened. For example, the percentage of individuals with adequate intakes of calcium or recommended servings of milk declined (30) . When examined over a longer time period (1965–1991), change in mean DQI scores among both Caucasians and African-Americans, and among low, middle and high income Americans indicated a small improvement in dietary quality (31 ,32) . Among the individual components of the DQI, average intakes of dietary fats and sodium remained above the recommended amounts, whereas average intakes of calcium and servings of fruits, vegetables and grains remained below the recommended amounts in all subgroups of the population.

Measures of moderation and proportionality

Aspects of the total diet, other than variety, include moderation and proportionality. Moderation refers to the avoidance of excesses, especially of those food components (e.g., fats, added sugars, alcohol, sodium) believed to be related to suboptimal health outcomes. A unique component of the DQI-R is the assignment of a maximum dietary moderation score to individuals with intakes of discretionary fat, added sugars, alcohol intakes and sodium intakes below cut-off points determined from Pyramid recommendations. The dietary moderation score contributed significantly to the variation in the overall DQI-R score, providing evidence that dietary moderation is positively associated with diet quality (29) .

Proportionality refers to the amounts of foods consumed in relation to each other. The shape of the Pyramid suggests that more servings of foods should be consumed from the grain group and fewer servings of foods should be consumed from the dairy and meat groups. Although not obvious from the Pyramid graphic, proportionality also refers to recommendations for subgroups within major food groups (e.g., dark green vegetables should be consumed several times per week). Proportionality implies that, to maintain energy intake, increased intake from one food group or subgroup necessarily results in decreased intake from another. The food groups interact, as demonstrated by results of Krebs-Smith et al. (24) who showed that food group patterns meeting the Pyramid recommendation for total fat also met the Pyramid recommendation for fruit.

Issues associated with patterning and scoring methods

The development of the PSDB has made a major contribution to analyses of the U.S. diet, considering the disaggregation of food mixtures, the assignment of foods to their respective groups (including fats and sweets) and the quantification of all food groups (including the tip of the Pyramid) in terms that are consistent with current guidance (e.g., servings). Each of the various dimensions of the diet can be quantified and examined, singly or simultaneously, providing a foundation for assessments of total diet and the direct comparison of diets to recommendations.

However, quantifying the multiple dimensions of diet simultaneously has proven problematic. The seemingly infinite variations in intake along these multiple dimensions necessitate some sort of summary function with which to simplify the interpretation of results. The patterning approach (24 25 26) retains the multiple dimensions on which it is based but, on each dimension, dichotomizes the continuous data into categories of "met"/"not met." Thus, for individuals whose energy-based recommendation for fruit is 2 servings, those who report as few as 2 servings or as many as 10 servings of fruit would be in the "met" group, whereas those who report no servings or as many as 1.9 servings of fruits would be in the "not met" group. Krebs-Smith et al. (24) stated that this approach "obscures the large variability in the amounts of each food group consumed" in an effort to "separate people according to the most rudimentary categories of interest." In addition, patterns are defined on the basis of the five major food groups and do not include recommendations related to fats and sugars. Had fats and sugars been evaluated as separate food groups, and a category of "exceeding the energy-based recommendations" for each of the food groups been created in addition to "met"/"not met," the number of patterns would have increased exponentially. Such a large number of patterns would certainly reduce the precision of estimates associated with those patterns and limit their interpretation. Consequently, fats and sugars were not included in defining the pattern, but only as an outcome.

The scoring methods retain the quantification of each dimension but sum these together, in effect reducing the number of dimensions to one. Although this approach is intuitively appealing and offers a simple way to evaluate diets, it implies that certain components of the diet are independent, equally important and additively related to health. This is probably not the case. For example, two components of these indices, total and saturated fat, are highly correlated; consequently, the score is more heavily weighted toward dietary fat than any other dietary component. The assignment of a range of points to each component does take variability in food and nutrient intakes into account, but like the variety scores and food patterning techniques, intakes of food groups above energy-based recommendations are not distinguished further. Total scores at the extremes of the distribution are straightforward in their interpretation, but those in the middle of the distribution (where the majority of individuals fall) are difficult to interpret because distinctly different dietary patterns could result in the same score. For example, one person could score a 70 by completely meeting 7 of 10 recommendations but completely failing to meet 3 recommendations. Another person could score a 70 by partially meeting each of the 10 recommendations. To clarify total scores, individual components of the index are often evaluated. This raises questions concerning whether total scores add information beyond traditional analyses of individual food and nutrient intakes (33) .

Summary of how U.S. diets fare

Evidence from studies of dietary variety shows that one third of the population eats at least some food from all food groups (5) . Evidence from studies of food group patterns shows that many fewer (1–3%) eat the recommended number of servings from all food groups on a given day (5 ,24 25 26) . Fruits are the most commonly omitted food group, whereas vegetables and meat are the most commonly met by adults and dairy the most commonly met by youth. Among adults, mean number of servings of fruits and dairy fall noticeably below that recommended, whereas among youth, mean number of servings from all food groups but dairy fall below the recommendations (21 ,25 ,26) . Specific choices within vegetables (i.e., dark green, deep yellow) and within grains (i.e., whole grains) are noticeably below amounts recommended by current food guidance. Intakes of total fat, discretionary fat and added sugars continue to exceed current recommendations. Analyses of diets in relation to socioeconomic status have consistently shown that diets improve with increases in education and income (22 ,28 ,30) . These results demonstrate shortcomings in variety, proportionality and moderation in the diets of most Americans.

	STRATEGIES FOR IMPROVEMENT

TOP ABSTRACT INTRODUCTION ASSESSING THE TOTAL DIET:... STRATEGIES FOR IMPROVEMENT REFERENCES

 
Improving food guidance

Although the food guidance system succeeded in its goals of nutrient adequacy, moderation of potential excesses and usability that were set at the time, the research that led to its development is now 20 years old, and much has changed since then in our understanding of nutrition and in the demographics and eating patterns of the American public. However, the food guidance system was designed to accommodate anticipated changes in dietary recommendations over time without the need for substantial revision of the associated food guide (e.g., the Pyramid) (33) . For example, recent changes in nutrient recommendations (34 35 36) and increasing numbers of fortified foods can be accommodated. Although the Dietary Guidelines state that supplements should not replace foods to meet usual nutrient needs, increases in the requirements of some nutrients (e.g., calcium) may require that the recommended number of servings from certain food groups be modified for certain subgroups (e.g., increase in recommended number of servings of milk products for children) (37) . Increased recommendations for other nutrients (e.g., folate) may require messages about the use of supplements by some Americans (e.g., women of child-bearing age) for whom it would be difficult to obtain the recommended amounts by diet alone.

Various food group designations, however, may deserve reconsideration. In particular, specific types of fat might receive greater or lesser emphasis, and a new typology for fruits and vegetables may be required. For example, a recent report notes convincing evidence that diets high in vegetables, particularly dark green leafy, cruciferous and alium vegetables, and/or high in fruits, particularly citrus fruits, are protective for some types of cancer (38) . This report recommends five or more servings a day of a variety of vegetables and fruits, excluding starchy vegetables and fruits (e.g., cassava, sweet potato, potato, yam, plantain, banana). This report also recommends more than seven servings a day of a variety of other plant foods including cereals (grains), pulses (legumes), roots, tubers and plantains.

More recent data on typical portion sizes of commonly consumed foods may be helpful in quantifying the desirable number and size of servings from each group. For example, the Pyramid graphic recommends 6–11 servings of grain but does not specify how much counts as a serving. Examination of the Pyramid brochure reveals that each serving is the equivalent of only 1 slice of bread or 1/2 cup of rice or pasta, amounts that are about half the size of typical portions, if not smaller (39) . This was the case even when the food guidance system was developed (40) , but the developers chose to keep the suggested serving sizes consistent with previous food guides and with food labels of that time period (1 ,2) . This decision allowed the recommended number of servings to be greater, thereby giving grains more attention in order to increase consumption. Consequently, the grains group formed the base of the Pyramid. This message requires clarification, given that grain consumption has increased without noticeable improvement in whole-grain intake (41) .

It has been argued that the Pyramid recommendations are not the only way to achieve a nutritionally adequate diet, and that alternative schemes must be developed for vegetarians and persons who abstain from milk and milk products. In fact, the food guidance system was never meant to be limiting. The developers recognized that different guidance would be necessary for infants and very young children, vegetarians, particular ethnic groups and others with distinctly different food preferences or dietary needs (1 ,2) . Given the increasing diversity of the U.S. population and the accompanying changes in the marketplace, the system must be adapted to different cultures and incorporate more types of foods commonly consumed by members of the rapidly growing ethnic groups in the U.S. (42) . Many alternative "pyramids" have been put forth (43 44 45 46) . However, with the exception of the Puerto Rican Pyramid (46) , none has been tested rigorously for validity and reliability against a set of stated goals and objectives based on current nutrition science. Superficial changes in the graphics with examples of ethnic or vegetarian foods will not suffice, particularly because nutrient adequacy is not verified. At the present time, national dietary data are lacking for many cultural groups and should be collected for alternative pyramids to be formally developed, tested and disseminated at the national level.

Indispensable pieces of healthful dietary patterns must be identified, and commonalities drawn across different schemes to produce universal nutrition messages. However, it is important to recognize that food guidance development is an evolutionary process. Accommodation must be made for nutrient attainment of the population, nutrient availability, diet/health recommendations and what is acceptable to people (11) . The total diet concept, which has barely been realized, given the Pyramid graphic’s emphasis on only the major food groups, continues to have merit and should be retained and strengthened. Many of the inadequacies and imbalances in the current American diet relate to issues that were integral in the development of the food guidance system but lost in the translation to the Pyramid graphic, i.e., insufficient intakes of whole grains and dark green/deep yellow vegetables and an overabundance of added sugars and, to a lesser degree, total fat. Graphic representations of any subsequent guidance systems should prominently display these ideas.

In addition, changes in the food supply are necessary for all Americans to meet Pyramid recommendations. McNamara et al. (47) estimated the gaps between recommended intakes and food intakes in 1994 in order to determine changes in agriculture policy that might be needed for all consumers to meet recommended intakes. These gaps were substantial for fruits, certain subgroups of vegetables, and added sugars, and to a lesser degree, discretionary fats. For all Americans to immediately meet the corresponding Pyramid recommendations, the following changes in the food supply were estimated. The supply of fruit would have to increase by approximately two thirds. The supplies of dark green vegetables, deep yellow vegetables, and dry beans, peas and lentils would each have to triple. The supply of white potatoes and other starchy vegetables would have to decrease by about one half. The supply of added sugars would have to decline by at least one half (amounting to 21 billion pounds). The supply of added fats would have to decline by 16% (amounting to 3 billion pounds). Projecting to the year 2020, based on Census estimates for population shifts and growth, larger gaps would occur in dark green and deep yellow vegetables and in legumes. Such changes in the food supply, in combination with improved food guidance, could certainly increase the likelihood that more consumers would "Let the Pyramid guide their food choices." The challenge is how to make these changes in the food supply.

Improving assessment

The PSDB has proved to be an extremely worthwhile tool for evaluating the diet relative to current recommendations and could aid future assessments of dietary variety, food patterns and scores. Though developed originally for use with the CSFII, it can also be used to examine data from the NHANES III with a food code linking system designed for this purpose (available at http://www-dccps.ims.nci.nih.gov/ARP/). The incorporation of tools such as the PSDB into dietary assessment software programs would broaden their capabilities tremendously. However, even the PSDB could be expanded to examine intricacies of the diet beyond those incorporated into the food guidance system. For example, it would be useful to have information on more levels of disaggregation (such as individual food commodities as well as food groups) and on more attributes of foods (such as botanical classifications for fruits and vegetables). A multilevel system with numerous attributes would increase the flexibility of the food grouping system and facilitate future analyses of the total diet accordingly.

The variety, food patterning and scoring methods provide worthwhile starting points with which to assess the total diet of Americans. Each method utilizes a different approach for condensing numerous continuous variables into a single assessment. Variety scores assess the presence/absence of food intakes, then assign a score for the number of different foods/groups; food patterns measure food group intake continuously, then truncates intake into categories; total dietary scores measure intakes of food and other dietary components continuously, then add them together. To improve upon these methods, further consideration is warranted regarding the underlying goal of the assessment.

If the goal is descriptive, such as an overall assessment of diets in relation to current guidance, scoring multiple components associated with specific dietary recommendations makes sense. However, glyphs may provide more useful information than a total diet score. A glyph is a way of visually depicting the dimensions of several variables simultaneously by constructing a common object, such as a face, and graphing each parameter to a particular feature, such as nose length or mouth shape. For example, The Interactive Healthy Eating Index, an on-line version of the HEI, allows someone to enter the foods he/she has eaten in a day (48) . In addition to receiving a "score" for the overall quality of their diet and a "score" for each of the components in relation to the dietary recommendations, this tool also depicts a person’s diet according to building blocks of the pyramid. This pyramid glyph shows how each component of that pyramid measures up visually to the recommended amounts.

If the goal is to determine which areas of the diet are most critical, an assessment technique that discriminates, through an iterative process, which dietary factors (such as food groups) will best distinguish between "healthful" and "nonhealthful" diets would be preferred. Grain intake may contribute to the diet’s healthfulness, but grains are so ubiquitous in American diets that their measurement does not sort diets into qualitatively different groups. On the other hand, if fruit intake is the food group most likely to be absent, then fruits would discriminate among diets better than other food groups. Subgroup intakes (e.g., whole grains, dark green/deep yellow vegetables) are likely to be more discerning than the main food groups. Use of dietary supplements may further discriminate among individuals. Discussion regarding what are appropriate outcomes (e.g., nutrient adequacy, mortality, body fatness) and ways to validate these methods are necessary (e.g., use of biomarkers).

In addition to food and nutrient intakes, the concepts of variety, moderation and proportionality are important to consider. Although measures of variety and moderation have been developed and used in methods that assess the total diet, further empirical evaluation of these measures is required. Ways to assess proportionality also must be explored and tested empirically. Consistency in these measures would aid in the comparison of results across studies.

Other analytical methods, not discussed in this paper, have been used to assess the total diet. Methods such as cluster analysis, principal component/factor analysis and structural equation modeling take advantage of the correlations between dietary components to characterize diets and identify patterns. For example, cluster analysis identifies clusters of individuals with characteristic dietary patterns. Factor analysis identifies groups of dietary variables (e.g., foods listed on food records or food-frequency questionnaires), related to each other but relatively independent of other dietary variables, and calculates factor scores for each pattern for each individual. The application of other methodologies including discriminant analysis, categorization and regression tree analysis, hierarchical regression and signal detection methodology to dietary patterns is intriguing because of their potential to discriminate among interrelated variables. It is important to keep in mind that many elements of these methods are subjective (49 50 51) . In addition, dietary patterns that result from such methods can be difficult to interpret and translate into usable dietary guidance.

Unfortunately, all methods that assess the total diet are plagued by measurement error and bias that affect dietary assessment in general (52) . National food consumption surveys are limited by the number of days of dietary data collected; thus, they cannot provide estimates of usual intakes by individuals. Analyses based on single 24-h recalls further increase the likelihood that food and nutrient intakes of individuals, especially those in the extremes of the distributions, are misclassified, resulting in overestimation of those above or below a particular cut-off value (e.g., servings of food groups). Underreporting of energy is particularly prevalent in 24-h recalls, which may further affect whether individuals are classified as meeting or not meeting recommendations.

Efforts to counter these limitations are ongoing. For example, methods have been developed to adjust dietary data to produce nutrient estimates that more closely represent "usual intake," provided that at least 2 d of dietary data have been collected on a subset of individuals (53) . Use of these methods is strongly recommended when assessing the percentage of individuals whose nutrient intakes meet the dietary recommendations (e.g., percentage of individuals with saturated fat <10% of energy) (54) . Methods to adjust dietary data to produce food estimates that more closely represent "usual intake" are being explored (55) . To increase awareness and use of these methods, software programs that incorporate these methods (e.g., C-SIDE developed by researchers at Iowa State University) must be disseminated throughout the research community with the release of national survey data. Use of doubly labeled water to validate energy intake (56) and analytical techniques that compute energy intake in relation to basal metabolic rate using age- and gender-based equations (57 ,58) are strategies to assess the degree of underreporting.

In conclusion, the diets of most Americans are in need of improvement. The issues surrounding dietary guidance are complex. To best assist consumers, messages must be simple and direct. This is the intent of the newly worded guideline, "Let the Pyramid guide your food choices." However, clarification and promotion of key themes of the total diet, i.e., variety, moderation and proportionality, inherent in the food guidance system that underlies the Pyramid, are required for the guideline to have the desired effect. Improvements in methods that capture the total diet concept are warranted to best assess the diets of Americans.

	FOOTNOTES

 
¹ Published as a supplement to The Journal of Nutrition. The publication of this supplement was sponsored by the National Cancer Institute, National Institutes of Health, Bethesda, MD. The guest editor for this publication was Susan M. Krebs-Smith, NCI, NIH, Bethesda, MD.

³ Abbreviations used: CSFII, Continuing Survey of Food Intakes by Individuals; DGAC, Dietary Guidance Advisory Committee; DQI, Diet Quality Index; DQI-R, Diet Quality Index revised; HEI, Healthy Eating Index; NHANES, National Health and Nutrition Examination Survey; NFCS, Nationwide Food Consumption Survey; PSDB; Pyramid Servings Database; RDA, Recommended Dietary Allowance.

⁴ Distinct from the Food and Drug Administration’s Total Diet Study that estimates both nutrient and contaminant intakes of population subgroups (12) .

	REFERENCES

TOP ABSTRACT INTRODUCTION ASSESSING THE TOTAL DIET:... STRATEGIES FOR IMPROVEMENT REFERENCES

 

1. Cronin F. J., Shaw A. M., Krebs-Smith S. M., Marsland P. M., Light L. Developing the food guidance system for "Better Eating for Better Health," a nutrition course for adults. U.S. Department of Agriculture, Human Nutrition Information Service. Administrative Report no. 377 1985;

2. Cronin F. J., Shaw A. M., Krebs-Smith S. M., Marsland P. M., Light L. Developing the food guidance system to implement the Dietary Guidelines. J. Nutr. Educ. 1987;19:281-302

3. U.S. Department of Agriculture (1992) The Food Guide Pyramid. Home and Garden Bull. no. 252, Washington, DC.

4. U.S. Department of Agriculture, Dietary Guidelines Advisory Committee (1985, 1990, 1995, 2000) Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans (1985,1990, 1995, 2000), to the Secretary of Agriculture and the Secretary of Health and Human Services. Washington, DC.

5. Kant A. K., Schatzkin A., Ziegler R. G., Nestle M. Dietary diversity in the U.S. population, NHANES II, 1976–1980.. J. Am. Diet. Assoc. 1991;91:1526-1531[Medline]

6. Krebs-Smith S. M., Smiciklas-Wright H., Guthrie H. A., Krebs-Smith J. The effects of variety in food choices on dietary variety. J. Am. Diet. Assoc. 1987;87:897-903[Medline]

7. McCrory M. A., Fuss P. J., McCallum J. E., Yao M., Vinken A. G., Hays N. P., Roberts S. B. Dietary variety within food groups: association with energy intake and body fatness in men and women. Am. J. Clin. Nutr. 1999;69:440-447[Abstract/Free Full Text]

8. Cronin F. J. Reflections on food guides and guidance systems. Nutr. Today 1998;33:186-188

9. U.S. Department of Agriculture, Agricultural Research Service (1977) Food for Fitness—A daily food guide, 1958. 3rd rev. ed. USDA Leaflet no. 424, Washington, DC.

10. U.S. Department of Agriculture, Agricultural Research Service (1957) Essentials of an adequate diet Facts for nutrition programs ( by Page, L & Phipard. E. F.) Home Economics Research Report no. 3. Washington, DC.

11. Light L., Cronin F. J. Food guidance revisited. J. Nutr. Educ. 1981;13:470-475

12. Pennington J. A., Schoen S. A. Total diet study: estimated dietary intakes of nutritional elements, 1982–1991. Int. J. Vitam. Nutr. Res. 1996;66:350-362[Medline]

13. U.S. Department of Agriculture, Science and Education Administration (1979) Food. Home and Garden Bull. no. 228. Washington, DC.

14. National Research Council, Food and Nutrition Board Recommended Dietary Allowances 10th ed. 1989 National Academy of Sciences Washington, DC.

15. U.S. Department of Agriculture and U.S. Department of Health and Human Services (1980, 1985, 1990, 1995, 2000) Nutrition and Your Health: Dietary Guidelines for Americans. Home and Garden Bull. no. 232, Washington, DC.

16. Kant A. K., Schatzkin A., Harris T. B., Ziegler R. G., Block G. Dietary diversity and subsequent mortality in the First National Health and Nutrition Examination Survey Epidemiologic Follow-up Study. Am. J. Clin. Nutr. 1993;57:434-440[Abstract/Free Full Text]

17. Fernandez E., D’Avanzo B., Negri E., La Vecchia C., Franceschi S. Diet diversity and colorectal cancer. Prev. Med. 2000;31:11-14[Medline]

18. Franceschi S., Favero A., La Vecchia C., Negri E., Dal Maso L., Salvini S., Decarli A., Giacosa A. Influence of food groups and food diversity on breast cancer risk in Italy. Int. J. Cancer 1995;63:785-789[Medline]

19. Kant A. K., Schatzkin A., Graubard B. I., Schairer C. A prospective study of diet quality and mortality in women. J. Am. Med. Assoc. 2000;283:2109-2115[Abstract/Free Full Text]

20. La Vecchia C., Munoz S. E., Braga C., Fernandez E., Decarli A. Diet diversity and gastric cancer. Int. J. Cancer 1997;72:255-257[Medline]

21. Cleveland L. E., Cook D. A., Krebs-Smith S. M., Friday J. Method for assessing food intakes in terms of servings based on food guidance. Am. J. Clin. Nutr. 1997;65(suppl):1254S-1263S[Abstract/Free Full Text]

22. Kumanyika S., Krebs-Smith S. M. Preventive nutrition issues in ethnic and socioeconomic groups in the United States. Bendich A. Deckelbaum R.J. eds. Primary and Secondary Preventive Nutrition 2000 Humana Press Totowa, NJ.

23. Kant A. K., Schatzkin A., Block G., Ziegler R. G., Nestle M. Food group intake patterns and associated nutrient profiles of the U.S. population. J. Am. Diet. Assoc. 1991;91:1532-1537[Medline]

24. Krebs-Smith S. M., Cleveland L. E., Ballard-Barbash R., Cook D. A., Kahle L. L. Characterizing food intake patterns of American adults. Am. J. Clin. Nutr. 1997;65(suppl):1264S-1268S[Abstract/Free Full Text]

25. Munoz K. A., Krebs-Smith S. M., Ballard-Barbash R., Cleveland L. E. Food intakes of U.S. children and adolescents compared with recommendations. Pediatrics 1997;100:323-329[Abstract/Free Full Text]

26. Munoz K. A., Krebs-Smith S. M., Ballard-Barbash R., Cleveland L. E. Errors in food intake article. Pediatrics 1998;101:952-953[Free Full Text]

27. Kennedy E. T., Ohls J., Carlson S., Fleming K. The Healthy Eating Index: design and applications. J. Am. Diet. Assoc. 1995;95:1103-1108[Medline]

28. Patterson R. E., Haines P. S., Popkin B. M. Diet Quality Index: capturing a multidimensional behavior. J. Am. Diet. Assoc. 1994;94:57-64[Medline]

29. Haines P. S., Siega-Riz A. M., Popkin B. M. The Diet Quality Index Revised: a measurement instrument for populations. J. Am. Diet. Assoc. 1999;99:697-704[Medline]

30. Bowman, S. A., Lino, M., Gerrior, S. A. & Basiotis, P. P. (1998) The Healthy Eating Index: 1994–96. USDA, Center for Nutrition Policy and Promotion. CNPP-5. http://www.usda.gov/cnpp (accessed October 2000).

31. Popkin B. M., Siega-Riz A. M., Haines P. S. A comparison of dietary trends among racial and socioeconomic groups in the United States. N. Engl. J. Med. 1996;335:716-720[Abstract/Free Full Text]

32. Popkin B. M., Siega-Riz A. M., Haines P. S. Correction and revision of conclusions—dietary trends in the United States. N. Engl. J. Med. 1997;337:1846-1848[Free Full Text]

33. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion. Background information on the Dietary Guidelines for Americans. http://www.usda.gov/cnpp/Pubs/DG2000/Index.htm (accessed May 2000).

34. Institute of Medicine, Food and Nutrition Board Dietary Reference Intakes for calcium, phosphorus, magnesium, vitamin D, and fluoride 1997 National Academy Press Washington, DC.

35. Institute of Medicine, Food and Nutrition Board Dietary Reference Intakes for thiamin, riboflavin, niacin, vitamin B6, folate, vitamin B12, pantothenic acid, biotin, and choline 1998 National Academy Press Washington, DC.

36. Institute of Medicine, Food and Nutrition Board Dietary Reference Intakes for vitamin C, vitamin E, selenium, and carotenoids 2000 National Academy Press Washington, DC.

37. Shaw A. M., Escobar A. J., Davis C. A. Reassessing the Food Guide Pyramid: decision-making framework. J. Nutr. Educ. 2000;32:111-118

38. World Cancer Research Fund and American Institute for Cancer Research Food, Nutrition, and the Prevention of Cancer: A Global Perspective 1997 Washington DC.

39. Krebs-Smith, S. M., Guenther, P. M., Cook, A., Thompson, F. E., Cucinelli, J. & Udler, J. (1997) Foods commonly eaten in the United States. U.S. Dept. of Agriculture NFS Report no. 91–3.

40. Krebs-Smith S. M., Smiciklas-Wright H. Typical serving sizes: implications for food guidance. J. Am. Diet. Assoc. 1985;85:1139-1141[Medline]

41. Kantor, L. S. (1998) A dietary assessment of the U.S. food supply: comparing per capita food consumption with Food Guide Pyramid serving recommendations. U.S. Department of Agriculture, Agricultural Economic Report no. 772.

42. Achterberg C., McDonnell E., Bagby R. How to put the Food Guide Pyramid into practice. J. Am. Diet. Assoc. 1994;94:1030-1035[Medline]

43. Dwyer J., Marino L. An expedition into the pyramids. Nutr. Today 1998;33:180-249

44. Russell R. M., Rasmussen H., Lichtenstein A. H. Modified Food Guide Pyramid for people over seventy years of age. J. Nutr. 1999;129:751-753[Free Full Text]

45. Willett W. C., Sacks F., Trichopoulou A., Drescher G., Ferro-Luzzi A., Helsing E., Trichopoulos D. Mediterranean diet pyramid: a cultural model for healthy eating. Am. J. Clin. Nutr. 1995;61:1402S-1406S[Abstract]

46. Macpherson-Sanchez A. E. A food guide pyramid for Puerto Rico. Nutr. Today 1998;33:198-209

47. McNamara P. E., Ranney C. K., Kantor L. S., Krebs-Smith S. M. The gap between food intakes and the Pyramid recommendations: measurement and food system ramifications. Food Policy 1999;24:117-133

48. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion. The Interactive Healthy Eating Index. http://www.usda.gov/cnpp (accessed May 2000).

49. Martinez M. E., Marshall J. R., Sechrest L. Invited commentary: factor analysis and the search for objectivity. Am. J. Epidemiol. 1998;148:17-19[Free Full Text]

50. Slattery M. L., Boucher K. M. The senior authors’ response: factor analysis as a tool for evaluating eating patterns. Am. J. Epidemiol 1998;148:20-21[Free Full Text]

51. Wirfalt A.K.E., Jeffery R. W. Using cluster analysis to examine dietary patterns: nutrient intakes, gender, and weight status differ across food pattern clusters. J. Am. Diet. Assoc. 1997;97:272-279[Medline]

52. Thompson F. E., Byers T. Dietary assessment resource manual. J. Nutr. 1994;124(suppl.):2245S-2317S

53. Nusser S. M., Carriquiry A. L., Dodd K. W., Fuller W. A. A semiparametric transformation approach to estimating usual daily intake distributions. J. Am. Stat. Assoc. 1996;91:1440-1449

54. Institute of Medicine, Food and Nutrition Board Dietary Reference Intakes: Applications in Dietary Assessment 2000 National Academy Press Washington, DC.

55. Dodd, K., Krebs-Smith, S. M., Subar, A. F. & Kipnis, V. (2000) Estimating usual food intake using 24-hour recalls augmented with a food frequency questionnaire. Presented at The Fourth International Conference on Dietary Assessment Methods, Tucson, AZ.

56. Schoeller D. A. Recent advances from application of doubly labeled water to measurement of human energy expenditure. J. Nutr. 1999;129:1765-1768[Abstract/Free Full Text]

57. Goldberg G. R., Black A. E., Jebb S. A., Cole T. J., Murgatroyd P. R., Coward W. A., Prentice A. M. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur. J. Clin. Nutr. 1991;45:569-581[Medline]

58. Schofield W. N. Predicting basal metabolic rate, new standards and review of previous work. Hum. Nutr. Clin. Nutr. 1985;39C:5-41

59. American National Red Cross Better Eating for Better Health: Instructors Guide and Participants Packet 1984 ANRC Washington, DC.

60. U.S. Department of Agriculture, Agricultural Research Service. (1998) Food and nutrient intakes by individuals in the United States, by sex and age, 1994–96, Nationwide Food Surveys Report no. 96–2.

This article has been cited by other articles:

				L. B. Dixon, A. F. Subar, U. Peters, J. L. Weissfeld, R. S. Bresalier, A. Risch, A. Schatzkin, and R. B. Hayes Adherence to the USDA Food Guide, DASH Eating Plan, and Mediterranean Dietary Pattern Reduces Risk of Colorectal Adenoma J. Nutr., November 1, 2007; 137(11): 2443 - 2450. [Abstract] [Full Text] [PDF]

				J. J. Fogli-Cawley, J. T. Dwyer, E. Saltzman, M. L. McCullough, L. M. Troy, and P. F. Jacques The 2005 Dietary Guidelines for Americans Adherence Index: Development and Application J. Nutr., November 1, 2006; 136(11): 2908 - 2915. [Abstract] [Full Text] [PDF]