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 Krebs-Smith, S. M. Articles by Kantor, L. S. Search for Related Content PubMed PubMed Citation Articles by Krebs-Smith, S. M. Articles by Kantor, L. S.

---

Supplement

Choose a Variety of Fruits and Vegetables Daily: Understanding the Complexities¹

Risk Factor Monitoring and Methods Branch, National Cancer Institute, Bethesda, MD 20892-7344 and ^* Economic Research Service, U.S. Department of Agriculture, Washington, DC

²To whom correspondence should be addressed. E-mail: sk52r{at}nih.gov

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS OF ASSESSMENT SURVEILLANCE DATA STRATEGIES FOR IMPROVEMENT REFERENCES

 
The 2000 edition of Nutrition and Your Health: Dietary Guidelines for Americans is the first to include a recommendation aimed specifically at fruits and vegetables, apart from grains. This paper discusses these changes in the Dietary Guidelines, summarizes the methods of assessment pertaining to fruit and vegetable intakes and their related factors, and reviews the data available on current levels and trends over time. Recent methodological advances in the measurement of both the aggregate U.S. food supply and foods consumed by individuals have allowed for better estimates with which recommendations can be compared. The data on individual intakes suggest the following: Americans are consuming fruits and vegetables at a level near the minimum recommendations; to be in concordance with energy-based recommendations, they would have to consume 2 more servings per day; and dark green and deep yellow vegetables are accounting for a disproportionately small share of the total. Fruit and vegetable consumption appears to be rising, but only slightly, and this increase might be only an artifact of shifts in the population demographics. A number of studies suggest that low income households in poor central cities and sparsely populated rural areas often have less access to food stores and face higher prices for food, including fruits and vegetables, compared with other households. At the aggregate level, supplying enough fruits and vegetables to meet dietary recommendations for all U.S. consumers would require adjustments in U.S. agricultural production, trade, marketing practices and prices of these commodities.

KEY WORDS: • fruits • vegetables • diet surveys • dietary assessment • food supply

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS OF ASSESSMENT SURVEILLANCE DATA STRATEGIES FOR IMPROVEMENT REFERENCES

 
The 2000 edition of Nutrition and Your Health: Dietary Guidelines for Americans is the first to include a recommendation aimed specifically at fruits and vegetables, apart from grains (1 2 3 4 5) . This most recent adjustment reflects an evolving understanding of the role various classes of foods play, not only as suppliers of individual nutrients, but as complex mixtures of nutrients and other dietary components that may be helpful in preventing chronic disease as well as achieving nutrient adequacy.

The first two editions of the Dietary Guidelines (1 ,2) , issued in 1980 and 1985, simply stated, "Eat foods with adequate starch and fiber," suggesting as sources "whole-grain breads and cereals, fruits and vegetables, beans, peas, and nuts." The intended effect of this early guideline was to substitute complex carbohydrate calories for those coming from fats and sugars, which were thought to be excessive, and to increase the amount of fiber, which was thought to be lacking in the U.S. diet (6) . Different classes of fiber had been identified and were understood to have varying biological effects, including those on bowel transit time and cholesterol levels. It was also thought that diets that were habitually low in fiber might increase the risk of developing colon cancer, but convincing data on that issue were not available. The early guidelines were directional rather than quantitative (7) , that is, the amounts of fruits, vegetables and grains necessary to provide "adequate starch and fiber" were not specified.

In 1990, the guideline was changed to "Choose a diet with plenty of vegetables, fruits, and grain products," specifying foods that were being targeted. Replacing fat calories with those from carbohydrate continued to be an aim of this guideline, and the use of foods, rather than supplements, as sources of fiber was stressed. The Dietary Guidelines Advisory Committee (8) inferred that the benefits associated with dietary fiber could be due to some other factor associated with high fiber foods, rather than fiber per se. For the first time, minimum numbers of servings were suggested, i.e., 3 servings of vegetables and 2 servings of fruits. These amounts were based on the recommendations in a food guidance system that had recently been developed and tested by the USDA (9 10 11) ; these subsequently formed the basis of the recommendations in the Food Guide Pyramid (12) (Pyramid hereafter). This quantification of serving sizes and number is consistent with the recommendation to eat 5 servings of fruits and vegetables combined, as promoted by the National Cancer Institute’s Five A Day for Better Health Campaign (13) .

In 1992, the Pyramid was released, portraying grains as the base and fruits and vegetables, as separate food groups, sharing the next level. To reflect this change in emphasis, the wording of the guideline was rearranged slightly in 1995 to "Choose a diet with plenty of grain products, vegetables, and fruits." The text of the guideline highlighted the contribution these foods make to the intake of nutrients and other bioactive compounds, in addition to complex carbohydrates and fiber. As did its predecessor, the 1995 Dietary Guidelines Advisory Committee strongly endorsed foods, as opposed to supplements, as sources of these compounds, stating that "many factors in these foods are likely to contribute to lower risk" of heart disease and cancer (14) . They also pointed out that fruits and vegetables need not be "fresh" (i.e., unprocessed) to be healthful, so that consumers would not feel constrained by the limitations of produce being perishable, expensive or unavailable. The concept of "plenty" was again quantified, this time including a range for number of servings, to be more consistent with the newly adopted Pyramid (12) . The use of serving ranges, as opposed to a single level for all persons, was a major advance for food guidance because it accounted for the fact that intakes of foods from most of the major food groups should rise with rising energy levels (an exception was the recommendation for milk products, which varied instead by life cycle stage). High energy diets that meet only the minimum recommendation for each of the major food groups are disproportionately high in fats and/or sugars, and are likely to be low in relation to recommended nutrient intakes (9 10 11) .

Recognizing that fruits and vegetables are distinct from grains, both nutritionally and in terms of the ways in which they are incorporated into the diet, the 2000 Dietary Guidelines Advisory Committee decided to make separate guidelines for each, with the goals of simplifying the message, focusing attention on each and ultimately increasing intakes (15) . To reflect the fact that the composition of different fruits and vegetables varies widely with regard to nutrient and nonnutrient components, they incorporated an additional concept into the statement, with the suggestion to "Choose a Variety of Fruits and Vegetables Daily." However, they stopped short of specifying exactly how variety should be achieved.

Although the Dietary Guidelines are designed to advise consumers regarding healthy food choices, the public’s ability to make such choices is dependent on a myriad of factors. Figure 1 , adapted from a schematic provided in the second report to Congress on Nutrition Monitoring in the United States (16) , provides a model of factors influencing food choice. It suggests that there are many points in the sequence from production to consumption that ultimately affect food choice, and highlights where assessments are required. The remainder of this paper summarizes the methods of assessment pertaining to food supply, foods acquired and foods consumed by individuals, especially as they pertain to fruits and vegetables, and reviews the surveillance data available on current levels and trends over time.

View larger version (26K): [in this window] [in a new window]   Figure 1. Conceptual model for understanding factors influencing food choice (16)

	METHODS OF ASSESSMENT

TOP ABSTRACT INTRODUCTION METHODS OF ASSESSMENT SURVEILLANCE DATA STRATEGIES FOR IMPROVEMENT REFERENCES

The Economic Research Service (ERS)³ of the USDA annually calculates the "food supply" in the United States by tracking flows of >250 individual agricultural commodities through domestic marketing channels (17) . For most commodity categories, the food supply is measured as the sum of annual domestic production, beginning inventories and imports, minus exports, industrial nonfood uses, farm uses (seed and feed) and end-of-year inventories. ERS collects these data directly from producers and distributors using techniques that vary by commodity. Per capita consumption is calculated by dividing the food supply by the total U.S. population.

The Food Supply Data (FSD) are the only source of time-series data on food use in the United States, with estimates dating back to 1909 for most commodities. The data can be used to compare the average level of food consumption in the United States with that of other countries, to show year-to-year changes in consumption of major foods, to examine the potential of the domestic food supply to meet the nutritional needs of the population, to perform ecologic analyses of relationships between food and health and to analyze the effects of prices and incomes on food supplies. If waste and other losses in the system are relatively constant over time, as they are for fruits and vegetables, the FSD provide a useful measure of consumption trends.

    Food supply servings. As traditionally reported [in terms of lb/(person · y)], the FSD overstate the amount of food actually eaten by people by capturing substantial quantities of nonedible portions, such as watermelon rinds and peach pits, and edible portions lost to human use through waste and spoilage in the home and marketing system. The series also includes unknown quantities of fruits and vegetables used as ingredients in manufactured foods that are exported, such as pies, soups, spaghetti sauce and frozen meals (18) . As a result of this and other "overcounting," the average energy level associated with the traditional FSD is well above that believed to be actually ingested by the U.S. population (19) .

In 1997, the ERS developed new methods to adjust the FSD for waste and other losses and express the data in terms of Pyramid servings per day (18 ,20) . These methodological advances have provided a way to examine FSD directly in relation to food guidance without wastage and loss concerns.

Food supply servings have been estimated for all of the >120 types of fruits and vegetables and 22 varieties of dry edible beans, lentils and peas in the FSD. For vegetables, servings data are available for fresh, frozen, canned, dehydrated and, in the case of potatoes, for chips and shoestrings; for fruits, servings data are available for fresh, frozen, canned, dried and juices. Supplies of most fruits and vegetables are measured at the farm or primary processing level, such as the canning plant, and reflect the amount of major food commodities entering marketing channels, regardless of their final use (17) . There are little or no data available for many further processed products. For example, although relatively good data exist on tomatoes used for canning, FSD provide no information on end-uses for canned tomato products, such as tomato sauce, ketchup and salsa.

    Limitations of FSD. The lack of production and trade data for some fruits and vegetables means that they are not included in the FSD and that servings estimates may understate actual consumption for some subgroups. The most notable examples are many fresh and frozen dark green leafy vegetables (kale, collards, chard, bok choy, mustard, turnip and beet greens), fresh and frozen summer and winter squash, fresh blueberries and raspberries, as well as the many so-called "exotic" and "specialty" fruits and vegetables, such as chayote, jicamas and dasheens, introduced into mainstream markets in the last decade (19) . However, data from food consumption surveys, in which respondents are asked to describe every food eaten in a day, indicate that the use of these products accounts for only a fraction of total fruit and vegetable intake (21) . In addition, because the FSD capture only fruits and vegetables that move through the commercial production sector, they exclude produce sold through direct marketing channels such as farmers’ markets, consumer-supported agriculture programs, roadside stands or that grown in home gardens. Finally, unlike data collected on foods consumed by individuals, they convey no information about how the particular commodity is eaten, e.g., in a soup, casserole or on a hamburger; how it is prepared, e.g., fresh, boiled or fried; where it is distributed for consumption, e.g., supermarket, hospital, school or restaurant; or the socioeconomic characteristics of the consumer who eats it. Nonetheless, the data are a useful complement to the more direct measures of food consumed by individuals (16) .

Foods acquired

Data on food expenditure patterns by households of different socioeconomic and demographic characteristics can be used to assess existing market conditions, product-distribution patterns and consumer buying habits. Comparisons of food expenditures by different demographic groups are valuable to food marketers and processors, food program administrators, nutrition educators, health professionals, economists and nutrition researchers.

The Consumer Expenditure Survey (CES) is an annual survey of household food expenditures conducted by the Bureau of Labor Statistics of the U.S. Department of Labor (22) . Expenditures for at-home food are recorded in a diary by participants over a 2-wk period. Expenditures are the transaction costs, including excise and sales taxes, of foods purchased at grocery stores and other food stores during the diary-keeping period. Data are collected throughout the year, with increased data collection during the last 6 wk of the year to account for increased buying activity during the holiday season. Information on spending for food purchased at restaurants, carry-out and other types of eating establishments, is collected in a separate interview portion of the survey. Detailed tabulations analysis of the CES data for 133 food categories and 10 household socioeconomic characteristics are available for 1980–1992 from ERS (23) .

Foods consumed by individuals

The assessment of foods consumed by individuals is conducted by surveys in which representative samples are polled about their intakes. Such samples are selected both nationally and within individual states, depending on the population of interest. Because income, demographic and life style information is often gathered along with food intake data in surveys of individuals, interrelationships among all of these factors can be studied. Further information about the major surveys in the National Nutrition Monitoring and Related Research Program can be found in the Directory of Federal and State Nutrition Monitoring and Related Research Activities (24) . The following section describes the two principal methods of dietary assessment at the individual level, i.e., 24-h recalls and food-frequency questionnaires (FFQ), and concludes with the limitations of such assessments and outstanding issues.

    Surveys employing 24-hour recalls. Most of the reports of food intake based on nationally representative samples in the United States are derived from data gathered as part of the USDA’s food consumption surveys. The National Health and Nutrition Examination Surveys (NHANES), conducted by the Department of Health and Human Services, also collect food consumption data, but have generally reported only nutrient intakes as opposed to numbers of servings of foods. These surveys employ single or multiple 24-h recalls, which require that the respondent provide a detailed description of every food and beverage consumed on the previous day, thus providing a rich database from which to assess the intricacies of the American diet. However, there are inherent challenges in translating these details from one or a few "snapshots in time" to a view of the usual diet over time; these include developing measures and statistical tools to summarize and assess usual food intakes.

    Measures of intake. Traditionally, food intakes were assessed using 24-h recalls and/or records, grouping like foods together and reporting an average weight of foods consumed from the group. Mixed foods were grouped according to their main ingredient, resulting in reports for such nutritionally heterogeneous groups as "meat mixtures" and "grain mixtures." This method has been employed in USDA food consumption survey reports for >25 y, and thus allows comparability over time (25) . However, because gram weights rather than servings are used and because the components of mixtures are not grouped with their nutritional equivalents, it does not provide a measure of food intakes that is directly comparable to food guidance recommendations (26) .

Patterson et al. (27) drew attention to this concern and made the first attempt to quantify servings of fruits and vegetables relative to recommendations, using the 1976–1980 NHANES II survey. Without the aid of recipe files, which facilitate the disaggregation of food mixtures, these authors estimated the intakes of "servings" of fruits and vegetables by devising upper and lower threshold gram weights of portion size. Examining food group intakes more precisely in relation to recommendations was hampered by the inability to disaggregate food mixtures.

This limitation has been overcome by the Pyramid Servings Database (PSDB) and its associated recipe files, developed by the USDA. A prototype of this database was tested with the 1989–1991 Continuing Survey of Food Intake by Individuals (CSFII) data (26 ,28 ,29) , and the PSDB was made publicly available with the release of the 1994–1996 CSFII data (30) . This database provides, for every item reported in the survey (whether a single food or a mixture), the number of servings from each of the major food groups and subgroups in the Pyramid contained in 100 g of that item. Detailed documentation pertaining to the PSDB can be found on the CD ROM data release (31) . The 1988–1994 NHANES III survey used the same food coding system as the CSFII; thus, the foods reported in that survey can be linked to the PSDB. These linkages are available on the National Cancer Institute’s Division of Cancer Control and Population Sciences, Applied Research Program website: http://www-dccps.ims.nci.nih.gov/ARP/NHANES/index.html.

One of the most controversial aspects of measuring servings of fruits and vegetables has been which forms and uses of these foods to include/exclude in the assessment. At the heart of this debate is the issue of whether to monitor every possible use of fruits and vegetables in the American diet, or only those that are promoted by current dietary guidance. For example, a potato might be readily classified as a vegetable, whereas the assignment of French fries and potato chips seems more ambiguous. At the time the PSDB was being conceived, a consensus workshop was held to determine how to monitor fruit and vegetable intakes for tracking the year 2000 objectives (32) . Although the group felt that, for most purposes, an all-inclusive definition was warranted, they recommended that "miscellaneous" uses, i.e., fruits and vegetables consumed as part of condiments, candy or potato chips, be partitioned and not counted toward meeting the objective. The reluctance to count such uses toward meeting the objective probably was rooted more in concern over causing a public relations disaster, as in the infamous "ketchup is a vegetable" fiasco (33) , than in any real difference in the nutritional contribution between, for example, the tomatoes in ketchup (miscellaneous) and those in tomato sauce (not miscellaneous). However, it is a far different thing to say the tomatoes in ketchup should be assessed as part of total fruit and vegetable intake than that ketchup should be relied on as a staple form of vegetables in the diets of children.

Following the workshop recommendations, total intakes were counted with and without the miscellaneous uses when the 1989–1991 CSFII was employed to provide baseline data for tracking the Healthy People 2000 objective (28 ,29) . A prototype of the PSDB was used, which included not only the servings of each group from each food code (e.g., the number of servings of vegetables in a hamburger sandwich), but also the servings from each ingredient of each food code (e.g., the servings in the ketchup on the sandwich). This added complexity of the database provided additional flexibility as well, and allowed for a relatively simple assessment of the miscellaneous uses.

Such flexibility was not built into the PSDB, which was designed to track servings from the Pyramid, rather than the health objectives per se. Inherent in the Pyramid is a "total diet" concept, which accounts for all foods eaten in one or more of the major groups and/or the tip. Thus, in the PSDB, the potatoes in potato chips are grouped with other potatoes (within the vegetables) and the fat in which they are fried ascribed to the tip of the Pyramid.

The release of the PSDB led to an evolution in the thinking about how the objectives should be measured; consequently, all uses of fruits and vegetables (including miscellaneous) will be included in monitoring the Healthy People 2010 objectives (34) . Although this is a defensible way to track both the recommendations of the Pyramid and the national health objectives, the ability to measure particular uses and partition those that are encouraged vs. those that are not still would be useful. In other words, the PSDB would benefit by building in the capacity to measure miscellaneous uses, as its prototype did. An estimate of fruit and vegetable intake excluding miscellaneous uses was made for a HP2000 Progress Review meeting,⁴ using the 1994–1996 CSFII and the PSDB, but lacked the precision of earlier estimates, due to the limitations of the database.

The emphasis on eating a variety of fruits and vegetables in the current Dietary Guidelines provides an added challenge in assessment. As stated previously, the Dietary Guidelines Advisory Committee did not indicate exactly how this should be achieved, although the background documentation to the Pyramid offers some indications of how it might be achieved (9 10 11) . At the most basic level, fruits and vegetables are classified as separate food groups with separate recommendations, and within each of these, subgroups have been highlighted as follows: citrus, melon, and berries; other fruit; dark green/deep yellow vegetables; starchy vegetables; and other vegetables. The nutritional adequacy of the Pyramid was assessed assuming that the two fruit subgroups would each make up one half of the fruit servings and the three vegetable subgroups would each make up a third of the vegetable servings. Within some of the vegetable subgroups, further subdivisions have been emphasized for inclusion several times a week, i.e., dark green vegetables and dried beans and peas. If increasing the number of different subgroups from within fruits and vegetables is what is intended by "Choose a Variety of Fruits and Vegetables Daily," then this could be measured easily with the current PSDB.

Another way variety might be assessed is by counting the number of different individual fruits and vegetables consumed. The Dietary Guidelines Advisory Committee noted that "different types of fruits and vegetables differ widely in their content of nutrients" (15) . The capacity to measure intakes of each individual fruit and vegetable, using the 1994–1996 and 1998 CSFII, has just been developed through a cooperative venture between USDA and the Environmental Protection Agency (EPA) (35) . This database was designed to assess human exposure to pesticide residues and provides, for each person, intakes of each commodity in terms of g food/(kg body weight · d). Presumably, these could be translated to total grams consumed which, in turn, could be linked to a database that provides grams per serving of each commodity, to determine servings from each commodity for each person. This database is available on CD ROM through the National Technical Information Service (http://www.ntis.gov/fcpc/cpn8898.htm).

    Statistical tools. Once the measures of fruit and vegetable intake have been established, further consideration must be given to how they will be interpreted. Among the greatest concerns to program planners and policy makers are the assessments, at the population level, of mean intakes and proportions above or below a particular cut-off (prevalences). Although 1 or 2 d of intake data do not represent usual intakes at the individual level, the data are robust with respect to obtaining mean intakes of the population. However, prevalences estimated from single (or even 2 or 3) 24-h recalls present a special challenge. Because intraperson variability in food intake is so great, measures covering only a few days have a wider distribution than an assessment of true usual intake would yield, and hence a greater proportion of persons falling above a cut-off at the end of the distribution (or below a cut-off at the beginning); this situation becomes more extreme the further out in the tail the cut-off is located. Consequently, the proportion of the population meeting the recommendation to consume 5 servings of fruits and vegetables, for example, would be very different if it were estimated for a given day (using a single 24-h recall) than if true usual intake could be ascertained. Methods of estimating usual food intake are being developed, even for single 24-hour recall data, but they require external data sources. Until such methods are refined and employed routinely, caution should be used in interpreting prevalences of intakes above or below a particular cut-off, obtained from unadjusted 24-h recalls.

    Surveys that employ FFQ Although 24-h dietary recalls are considered the "gold standard" because of the precision they afford, FFQ are also employed; because they use a longer reference period, they may actually provide a better indication of the propensity to consume a particular food or group of foods (36) . FFQ ask respondents to report the frequency with which they consume a list of foods over a given period (e.g., the past year), in effect, their usual intake. This is a major advantage of an FFQ over a 24-h recall, in addition to being less costly to administer and process. Shortened versions of FFQ, or screeners, are being developed and tested to reduce respondent burden in an otherwise lengthy process (37 ,38) . A comprehensive review of the various instruments available can be found in Thompson and Subar (39) . FFQ have been used to measure fruit and vegetable intake in both the 5 A Day Baseline and Follow-up Surveys, as well as the Behavioral Risk Factor Surveillance System (BRFSS; see below on surveillance for description of these surveys).

Because FFQ are comprised of a finite list of foods, they present fewer challenges regarding measurement of food group intake. Thus, they are both simpler and more limiting than 24-h recalls. For example, assessing variety among fruits and vegetables requires merely counting the number of items reported as ever consumed, but the extent of variety that can be captured is limited by the length of the list. Furthermore, it has been suggested that the number of individual items included on the list may affect the estimates of mean total fruit and vegetable intake (40) .

    Limitations of assessment/outstanding issues. The assessment of fruit and vegetable intakes relative to food guidance has improved greatly in the last decade, but challenges remain. These include the need for greater flexibility in assessing fruit and vegetable intakes, greater comparability across studies and improved ability to assess usual intakes.

Flexibility. Although the food groups and subgroups in the Pyramid form the basis of recommendations in federal food guidance, they were developed nearly two decades ago and may have to be revisited in light of more recent data on diet and health outcomes. For example, the World Cancer Research Fund and the American Institute of Cancer Research (41) have suggested that, for the purposes of reducing cancer risk, potatoes and other starchy vegetables should not be included as vegetables, and bananas and plantains should not be included as fruits. Considering the varying content of nutrients and other phytochemicals among fruits and vegetables, a new taxonomy may be required. Until that time, the ability to count every individual fruit and vegetable, as with the USDA/EPA commodity database, will be required to provide the flexibility to group and regroup foods as warranted.

Comparability. Because studies of fruit and vegetable intake have been done using different instruments and different measures, all for varying purposes, results across studies are not directly comparable. Each study might be thought of as a window through which we can obtain only a sense of what the true picture is; each provides a slightly different perspective, but by examining them all, a more comprehensive view emerges. Calibration studies have been done to compare the results obtained from different assessment methods, providing an important resource for interpreting these various results (42) .

Usual intakes. In a departure from the 2000 Healthy People national health objectives (43) , the 2010 version included separate objectives for fruits and vegetables and, within the vegetable objective, added a specific aim related to dark green/deep yellow vegetables, as follows: 1) Increase the proportion of persons 2 y old who consume at least two daily servings of fruit. 2) Increase the proportion of persons 2 y old who consume at least three daily servings of vegetables, with at least one third being dark green or deep yellow vegetables (34) .

These additional degrees of specificity with respect to food groups that are singled out in food guidance will encourage the monitoring of fruit and vegetable intakes with more specificity, rather than as one large group. However, the use of a prevalence, rather than a mean, as the main indicator implies that estimates of usual intake are being monitored, which is not the case with only a few days of dietary data. Therefore, caveats should be used when reporting, and caution used when interpreting such data until usual intake estimates can be provided.

	SURVEILLANCE DATA

TOP ABSTRACT INTRODUCTION METHODS OF ASSESSMENT SURVEILLANCE DATA STRATEGIES FOR IMPROVEMENT REFERENCES

 
This section reviews the available data pertaining to fruits and vegetables at the food supply, foods acquired and food consumption (individual) level.

Food Supply Data (FSD)

FSD related to fruits and vegetables, with wastage factors accounted for, in terms of servings/(person · d) are shown in Tables 1 and 2. The food supply provided 5.3 servings of fruits and vegetables (including legumes) per person per day in 1998, slightly higher than the minimum recommendation shown in the Dietary Guidelines and the Pyramid. Vegetables accounted for a 74% share of total servings, representing nearly 4 servings/(person · d), whereas fruit represented 1.4 servings.

Fruit and vegetable servings increased an estimated 27% between 1970–1974 and 1998, which is the equivalent of 1.1 servings/d. Examples of a single serving include one medium apple, 1 cup raw salad greens or 1/2 cup of cooked carrots. Increases in fresh fruit (0.22 servings) and fruit juice (0.12 servings) offset a moderate decline in canned/chilled fruit to provide a 0.3 serving increase in fruits overall. Vegetables increased by 0.8 servings, due in large part to a 0.23 serving increase in frozen potatoes (virtually all used for fried potatoes.) Most of this increase in fruit and vegetable servings occurred since 1980, the year in which the Dietary Guidelines were first released.

Foods acquired

Even if aggregate supplies are adequate, the successful adoption of Federal dietary recommendations will require that individual consumers have sufficient access to quantities of fruits and vegetables at affordable prices and in forms that meet standards for quality, taste, palatability and convenience. Between 1982 and 1997, fruits and vegetables led all other food categories in retail price increases, with increases for fresh fruits and vegetables more than double those for processed products (17) .

A number of studies suggest that these issues are especially salient for low income households in poor central cities and sparsely populated rural areas because they often have less access to food stores and face higher prices for food (44 45 46) . Access to food stores can significantly affect the quality and affordability of food available to low income consumers, which in turn affects the ability of such households to fully utilize Federal food assistance programs such as USDA’s Supplemental Nutrition Program for Women, Infants, and Children (WIC) and the Food Stamp Program.

Low income households are less likely to live in suburban locations in which store choices are greater and food prices are lower. Consequently, such households spend a smaller share of their food dollar in supermarkets, which typically offer lower prices, i.e., 10% lower than in other food stores (44) , and a greater range of brands, package sizes and quality. Even when the poor do have equal access to supermarkets, supermarket prices in rural areas, where the poor are more likely to live, are 4% higher than in suburban areas (45) . Supermarkets in low income neighborhoods are generally smaller in size and carry a narrower range of products than those in other neighborhoods (44 ,46) . Low income consumers, who often lack regular access to a vehicle or who live in neighborhoods not well served by public transportation, may be limited to shopping nearby, or to spending more time and money for travel to and from more distant stores (47) .

Even when high quality fruits and vegetables are readily available to low income households at the same prices paid by other consumers, budget constraints may make it difficult to incorporate such items into the diet. Data from the CES indicate that households allocate their at-home food budgets across food groups differently, depending on their income. For example, in 1998, households in the poorest 20% of the nation’s income distribution (those with incomes <$12,367 annually) spent $295 per person on fruits and vegetables compared with $739 spent by those in the highest income bracket. The lowest income households spent 60% less on both fresh and processed fruits and vegetables than the highest income households (22) .

Individual intakes/trends over time

    CSFII. Assessments of fruit and vegetable intake relative to national health objectives and dietary guidance have been reported elsewhere, in terms of mean intakes and prevalences of high and low intakes, using both the 1989–1991 and 1994–1996 CSFII (26 ,28 ,29) . Some of these data are also available in greater detail on the USDA, Agriculture Research Service website (30) . When interpreting the prevalences, as stated before, it should be noted whether the estimates are for a given day or for a multiple-day average. Prevalences based on usual intake distribution estimates are not yet available.

Previously published reports from 1989–1991 are not directly comparable to those for 1994–1996 because of differences in the measures of fruits and vegetable used (i.e., whether miscellaneous fruits and vegetables were included) and in the number of days of data reported. Therefore, Tables 3 4 5 6 7 8 9 are provided to make the appropriate comparisons. In all of these tables, miscellaneous forms (such as from condiments, candy and chips) and fried potatoes are not excluded. Tables 3 4 5 6 7 8 provide the mean number of servings per day from total fruits, total vegetables, and total fruits and vegetables (Tables 3 and 4 ), from vegetable subgroups (Tables 5 and 6 ), and fruit subgroups (Tables 7 and 8 ), for 1989–1991 (Tables 3 , 5 and 7) and 1994–1996 (Tables 4 , 6 and 8) . Table 9 shows the percentage of the population meeting various recommendations, using one vs. the mean of three consecutive days 1989–1991 and one vs. the mean of two nonconsecutive days 1994–1996 of data.

In each time period, potatoes alone dominated vegetable intake, accounting for about a third of vegetables consumed. Consequently, though the Pyramid suggests that vegetable intake be divided evenly among the three vegetable subgroups, i.e., dark green/deep yellow, starchy (including more than potatoes) and other vegetables, Americans are currently eating nearly four times as many servings of starchy as dark green/deep yellow vegetables.

Table 9 shows the difficulty with estimating the percentage of the population meeting various recommendations, by comparing the estimates obtained from 1 vs. 2 d 1994–1996 and 1 vs. 3 d (1989–1991). When the prevalence of the condition approximates 50% of the population, the differences between one- and multiple-day estimates may be only minor. For example, the percentage of persons consuming 5 servings of fruits or vegetables slips from 36.0 to 32.6, when 1 vs. 3 d of data are used (1989–1991), and from 41.4 to 39.8, when 1 vs. 2 d of data are used (1994–1996). However, as noted earlier, if the prevalence of the condition is very low or very high, the difference between the one- vs. multiple-day estimates can be substantial. This is demonstrated by the percentage consuming <1 serving of vegetables, which shifts dramatically from 18.2 to 8.5, and from 18.8 to 10.6, respectively, when 1 vs. 3 d of data and 1 vs. 2 d of data are used. This presents a quandary in the use of these data, i.e., the multiple-day estimates are closer to what the theoretical usual intake is assumed to be, yet the difference in number of days affects the assessment of the trend. Single-day data are comparable, but seem to be meaningless once considered relative to the multiple-day data. The most important inference to be drawn from this table may be the need for establishing usual intake estimates, so that prevalences can be assessed with confidence.

    NHANES III. As mentioned previously, food intakes reported in the NHANES III, conducted between 1988 and 1994, were coded using CSFII codes, which allows for linkage to the PSDB; a system for doing that has been developed. Analyses of NHANES III data using this system provide an assessment of fruit and vegetable intakes that is remarkably similar to that derived from the CSFII. For all individuals 2 y old, mean intakes of total vegetables were 3.4 servings (dark green vegetables, 0.2 servings; deep yellow vegetables, 0.2 servings; dried beans and peas, 0.2 servings; white potatoes, 1.0 servings; other starchy vegetables, 0.2 servings; tomatoes, 0.5 servings; other vegetables, 1.1 servings) and total fruits, 1.5 servings (citrus, melon and berries, 0.7 servings; and other fruits, 0.8 servings) (in-house analyses; unpublished data).

    5 A Day surveys. The National Cancer Institute’s 5 A Day for Better Health Program conducted a baseline survey in 1991, when the program was initiated, and a follow-up survey in 1997 (49, Stables et al., unpublished data). The 1991 survey included 33 items, and estimates of total servings were adjusted on the basis of responses to summary questions (49) . In a more recent analysis of the 1991 and 1997 surveys done by Stables et al. (unpublished data), the 1997 survey was analyzed using a 7-item fruit and vegetable screener, and the 1991 survey was reanalyzed using the same 7 fruit and vegetable screener questions. Consistent with the promotional messages of the 5 A Day Program, certain fruits and vegetables were excluded from consideration, i.e., fried potatoes, fruits and vegetables that are considered "miscellaneous" according to the aforementioned criteria, and high fat fruits and vegetables, such as olives and avocados. Careful attention was paid to keeping the survey methods as comparable as possible, which lends credence to trend analyses. Results from the follow-up survey showed a statistically significant increase of nearly one fourth of a serving per day for the total population of adults; when the data were adjusted for demographic shifts between the two surveys, however, the increase was not significant, only about one eighth of a serving. These results suggest that, although the population as a whole may be eating fruits and vegetables more frequently than previously, the increase may be explained in large part by demographic shifts rather than by any real change in behavior among population subgroups. Indeed, the proportions of elderly, nonsmokers, and persons with higher education and/or higher income are growing in the population, and these groups tend to consume fruits and vegetables more frequently than others.

    State surveys. The Centers for Disease Control and Prevention conduct the BRFSS, an on-going telephone survey of adults conducted by health departments in all states (50 51 52) . Fruit and vegetable intakes are assessed via six food-frequency questions. Trends in fruit and vegetable consumption among adults in 16 states were reported by Li et al. (51) . They found "encouraging" progress in intakes between 1990 and 1994 because the prevalence of consuming fruits and vegetables 5 times/d increased from 19 to 22%; a plateau occurred between 1994 and 1996 because intake inched up only to 23%. Trends varied by physical activity level and weight status, with intake increasing among those of normal weight who were active, staying the same among the normal weight inactive and decreasing among the obese. Forthcoming investigations will provide estimates on the percentage of persons from all 50 states consuming fruits and vegetables 5 times/d and the mean frequency of consumption, based on data from the 1998 survey, using individual questions and selected constructs.

Methodological differences between FFQ and the 24-h recall indicate that the results from the 5 A Day surveys and the BRFSS are not directly comparable to those obtained from CSFII and NHANES. In addition, the operational definition of fruits and vegetables is different, in that these FFQ specifically exclude fried potatoes and they assess how often foods are eaten, rather than how many servings are consumed. Thus, they are not direct measures of the concept being conveyed by the Dietary Guidelines (that of servings as quantities rather than occasions) although they are better than 24-h recalls at assessing the propensity to consume fruits and vegetables.

	STRATEGIES FOR IMPROVEMENT

TOP ABSTRACT INTRODUCTION METHODS OF ASSESSMENT SURVEILLANCE DATA STRATEGIES FOR IMPROVEMENT REFERENCES

 
Individual-level behavior changes

Data on individual intakes suggest the following: on average, Americans are consuming fruits and vegetables at a level near the minimum recommendations; to be in concordance with energy-based recommendations, they would have to consume 2 servings/d more; and dark green and deep yellow vegetables are accounting for a disproportionately small share of the total. Although the optimal mix of fruits and vegetables that would ensure nutrient adequacy and reduce the risk of chronic disease is not known, the Pyramid recommendations provide for a degree of variety (obtaining servings proportionately from among the dark green/deep yellow, starchy and other vegetables) that would most certainly be an improvement over the current imbalance. Increasing intakes of fruit and of dark green/deep yellow vegetables, each by a single serving per day, would go a long way toward correcting this situation. Because most individuals cannot consume additional energy without adversely affecting their weight status (that is, most are not prepared to perform the compensatory activity), consumers should be alerted that these additional fruits and vegetables should replace sources of added sugars and discretionary fat, whose intakes are currently higher than recommended. Fruit and vegetable consumption appears to be rising, but only slightly, and this increase might be only an artifact of shifts in the population demographics.

System-level changes

At the aggregate level, supplying enough fruits and vegetables to meet dietary recommendations for all U.S. consumers would require adjustments in U.S. agricultural production, trade, marketing practices and prices of these commodities (53 ,54) . For some sectors, particularly citrus fruits, the required change could be significant compared with recent changes.

Nationwide, the market for fruits and vegetables has expanded in recent years, with many consumers benefiting from an ever-widening array of fruit and vegetable choices in their local supermarkets. Supermarket produce departments typically carry >400 produce items today, up from 250 in the late 1980s and 150 a decade earlier. Also, the number of ethnic, gourmet and so-called "natural" food stores, which highlight fresh produce, continues to rise (19) .

Fresh-cut fruits and vegetables, such as short-cut carrots, prepackaged salads, restaurant and supermarket salad bars, locally grown items, and exotic and specialty produce, as well as hundreds of new varieties and processed fruit and vegetable products have been introduced or expanded since the early 1980s, helping to fuel increased consumption (19) . The introduction of prepackaged, precut and other value-added products helped boost average consumption of fresh broccoli by 76% between 1993 and 1999 and average consumption of fresh carrots by 25% (55) .

Sharp increases in off-season imports of fruits from Chile and other southern hemisphere countries mean that consumers now have, virtually year-round, a choice of fruits such as grapes, peaches, nectarines and other fresh fruits that were once available only in summer months. For example, the import share of fresh grapes available in the food supply increased from 14% in 1980 to 44% in 1998. A sharp rise in imports of frozen French fries from Canada and frozen broccoli from Mexico in the 1990s has increased the import share of frozen vegetable consumption from 1 to 13% of domestic supplies since 1980.

Finally, many consumers increasingly have access to fresh, locally grown produce. Although vegetable production from home gardens declined sharply during the last century from 131 pounds (59.5 kg) per person in 1919 to <11 pounds (5 kg) in 1998, increased use of farmers’ markets and other direct marketing outlets in recent years has increased the availability of fresh, locally grown produce. According to USDA’s Agricultural Marketing Service, the number of farmers’ markets has nearly doubled in the United States since 1994 to over 2700 in 1998 (56) .

Despite these positive recent trends in fruit and vegetable availability, growing concern about insufficient food store access and low levels of fruit and vegetable consumption among low income households has prompted a number of public and private efforts to increase access to affordable, high quality fruits and vegetables in areas in which conventional retail sources may be limited or unavailable. These efforts include supporting the location of farmers’ markets and other direct marketing outlets in low income neighborhoods, community gardening programs, and gleaning and food rescue programs that increase the quantity of fresh fruits and vegetables to food pantries, soup kitchens, and other charitable food providers (57) . Federal efforts include the WIC Farmers’ Market Nutrition Program that allows participants to redeem vouchers for fresh fruits and vegetables at farmers’ markets and the USDA Farm-to-School Initiative that is helping farmers sell fresh fruits and vegetable to local schools participating in the National School Lunch Program (58) .

Figure 1 provided a depiction of the interplay among agricultural, economic and social forces and the supply, acquisition and consumption of foods. Although much can be learned from the existing assessment measures discussed in this paper (data on food supply, foods acquired and food consumption), many more data are required to develop an understanding of the intervening factors and their relationships to the food production/food consumption sequence.

	ACKNOWLEDGMENTS

	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: BRFSS, Behavioral Risk Factor Surveillance System; CES, Consumer Expenditure Survey; CSFII, Continuing Survey of Food Intake by Individuals; EPA, Environmental Protection Agency; ERS, Economic Research Service; FFQ, food-frequency questionnaires; FSD, Food Supply Data; NHANES, National Health and Nutrition Examination Surveys; PSDB, Pyramid Servings Database; WIC, Supplemental Nutrition Program for Women, Infants, and Children.

⁴ Healthy People: National Health Promotion and Disease Prevention Objectives. Progress Review: Nutrition. February 13, 1998. Internal document.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS OF ASSESSMENT SURVEILLANCE DATA STRATEGIES FOR IMPROVEMENT REFERENCES

 

1. U.S. Department of Agriculture, U.S. Department of Health and Human Services (1980) Nutrition and Your Health: Dietary Guidelines for Americans, 1st ed. Home and Garden Bull. 232, U.S. Government Printing Office, Washington, DC.

2. U.S. Department of Agriculture, U.S. Department of Health and Human Services (1985) Nutrition and Your Health: Dietary Guidelines for Americans, 2nd ed. Home and Garden Bull. 232, U.S. Government Printing Office, Washington, DC.

3. U.S. Department of Agriculture, U.S. Department of Health and Human Services (1990) Nutrition and Your Health: Dietary Guidelines for Americans, 3rd ed. Home and Garden Bull. 232, U.S. Government Printing Office, Washington, DC.

4. U.S. Department of Agriculture, U.S. Department of Health and Human Services (1995) Nutrition and Your Health: Dietary Guidelines for Americans, 4th ed. Home and Garden Bull. 232, U.S. Government Printing Office, Washington, DC.

5. U.S. Department of Agriculture, U.S. Department of Health and Human Services (2000) Nutrition and Your Health: Dietary Guidelines for Americans, 5th ed. Home and Garden Bull. 232, U.S. Government Printing Office, Washington, DC.

6. U.S. Department of Agriculture, Human Nutrition Information Service, Dietary Guidelines Advisory Committee (1985) Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans, to the Secretary of Health and Human Services and the Secretary of Agriculture.

7. Davis, C. & Saltos, E. (1999) Dietary recommendations and how they have changed over time In: America’s Eating Habits:Changes and Consequences (Frazao, E.,ed.) U.S. Department of Agriculture, Economic Research Service, Food and Rural Economics Division. Agriculture Information Bulletin no. 750.

8. U.S. Department of Agriculture, Human Nutrition Information Service, Dietary Guidelines Advisory Committee (1990) Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans, 1990, to the Secretary of Health and Human Services and the Secretary of Agriculture.

9. 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 1985 USDA Human Nutrition Information Service. Administrative Report no. 377.

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

11. U.S. Department of Agriculture, Human Nutrition Information Service (1993) USDA’s Food Guide: Background and Development. Miscellaneous Publication no. 1514.

12. U.S. Department of Agriculture, Human Nutrition Information Service (1992) The Food Guide Pyramid. Home and Garden Bull. 252, U.S. Government Printing Office, Washington, DC.

13. Heimendinger J., Van Duyn M. A., Chapelsky D., Foerster S., Stables G The national 5 A Day for Better Health Program: a large-scale nutrition intervention. J. Public Health Manag. Pract. 1996;2:27-35[Medline]

14. U.S. Department of Agriculture, Agricultural Research Service, Dietary Guidelines Advisory Committee (1995) Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans, 1995, to the Secretary of Health and Human Services and the Secretary of Agriculture.

15. U.S. Department of Agriculture, Agricultural Research Service, Dietary Guidelines Advisory Committee (2000) Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans, 2000, to the Secretary of Health and Human Services and the Secretary of Agriculture.

16. Life Sciences Research Office (1989) Nutrition Monitoring in the United States: An Update Report on Nutrition Monitoring in the United States. Prepared for the U.S. Department of Agriculture and the U.S. Department of Health and Human Services. DHHS Publication no. (PHS)89–1255. U.S. Government Printing Office, Washington, DC.

17. Putnam J. J., Allshouse J. Food Consumption, Prices, and Expenditure 1999:1970-1997 U.S. Department of Agriculture Economic Research Service, SB-965.

18. Kantor, L. S. (1998) A Dietary Assessment of the U.S. Food Supply: Comparing Per Capita Food Consumption with Food Guide Pyramid Serving Recommendations. Food and Rural Economics Division, Economics Research Service, U.S. Department of Agriculture, Agricultural Economic Report no. 772.

19. Putnam J. J. U.S. food supply providing more food and calories. Food Rev 1999;22:2-12

20. Kantor L. S., Lipton K., Manchester A., Oliveira V. Estimating and Addressing America’s Food Losses. Food Rev 1997;20:1-12

21. Krebs-Smith, S. M., Guenther, P. M., Cook, A., Thompson, F. E., Cucinelli, J. & Udler, J (1997) Foods Commonly Eaten in the United States: Quantities Consumed Per Eating Occasion and in a Day, 1989–91. U.S. Department of Agriculture NFS Report no. 91–3.

22. U.S. Department of Labor, Bureau of Labor Statistics (2000) Consumer Expenditure Surveys. [Online]. Available under "Tables (1984 through 1998)" <http://stats.bls.gov/csxhome.htm>.

23. Smallwood, D. M., Blisard, N., Blaylock, J. R., & Lutz, S. (1994) Food Spending in American Households, 1980-92. U.S. Department of Agriculture, Economic Research Service, SB-888.

24. Interagency Board for Nutrition Monitoring and Related Research, Bialostosky, K., ed. (2000) Nutrition Monitoring in the United States:The Directory of Federal and State Nutrition Monitoring and Related Research Activities. National Center for Health Statistics, Hyattsville, MD <http//www.cdc.gov/nchs/> DHHS Publication no. (PHS) 00–1255.

25. 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.

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

27. Patterson B. H., Block G., Rosenberger W. F., Pee D., Kahle L. L. Fruit and vegetables in the American diet: data from the NHANES II survey. Am. J. Public Health 1990;80:1443-1449[Abstract/Free Full Text]

28. Krebs-Smith S. M., Cook D. A., Subar A. F., Cleveland L., Friday J. U.S. adults’ fruit and vegetable intakes, 1989 to 1991: a revised baseline for the Healthy People 2000 objective. Am. J. Public Health 1995;85:1623-1629[Abstract/Free Full Text]

29. Krebs-Smith S. M., Cook D. A., Subar A. F., Cleveland L., Friday J., Kahle L. L. Fruit and vegetable intakes of children and adolescents. Arch. Pediatr. Adolesc. Med. 1996;150:81-86[Abstract]

30. U.S. Department of Agriculture, Agricultural Research Service (1998) Pyramid Servings Data: Results from USDA’s 1994–96 Continuing Survey of Food Intakes by Individuals. <http://www.barc.usda.gov/bhnrc/foodsurvey/Products9496.html#anchor164107>.

31. U.S. Department of Agriculture, Agricultural Research Service. (1998) 1994–96 Continuing Survey of Food Intakes by Individuals and Diet and Health Knowledge Survey. CD-ROM. National Technical Information Service accession number PB98–500457.

32. Wright, J. D., Ervin, B. & Briefel, R. R., eds. (1994) Concensus Workshop on Dietary Assessment:Nutrition Monitoring and Tracking the Year 2000 Objectives. U.S. Department of Health and Human Services, Hyattsville, MD.

33. Thornton, M. & Schram, M. (1981) U.S. holds the ketchup in schools:hold the pickles, hold the relish, hold the new school lunch regs. The Washington Post, 9/26/1981, A1.

34. U.S. Department of Health and Human Services (2000) Healthy People 2010 (Conference edition, in two volumes). Washington, DC.

35. Environmental Protection Agency, Office of Pesticide Programs Food Commodity Intake Database 2000 CD-ROM. National Technical Information Service accession number PB2000–500101.

36. Dodd, K. W., Krebs-Smith, S. M., Subar, A. F. & Kipnis, V. (2000) Estimating usual food intake using 24-hour recalls augmented with a food frequency questionnaire. The Fourth International Conference on Dietary Assessment Methods, September 17–20. Tucson, AZ

37. Thompson F. E., Kipnis V., Subar A. F., Krebs-Smith S. M., Potischman N., Schatzkin A Performance of two short instruments and a food frequency questionnaire to estimate daily number of servings of fruit and vegetables. Am. J. Clin. Nutr. 2000;71:1503-1510[Abstract/Free Full Text]

38. Block G., Gillespie C., Rosenbaum E. H., Jenson C. A rapid food screener to assess fat and fruit and vegetable intake. Am. J. Prev. Med. 2000;18:284-288[Medline]

39. Thompson F. E., Subar A. F. Dietary assessment methodology. Coulston A. M. Rock C. L. Monsen E. R. eds. Nutrition in the Prevention and Treatment of Disease 2001 Academic Press New York, NY (in press).

40. Krebs-Smith S. M., Heimendinger J., Subar A. F., Patterson B. H., Pivonka E. Using food frequency questionnaires to estimate fruit and vegetable intake: association between the number of questions and total intakes. J. Nutr. Educ. 1995;27:80-85

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

42. Dietary Assessment Calibration/Validation Register. [online] <http://www-dacv.ims.nci.nih.gov/> (Visited 10/13/2000).

43. U.S. Department of Health and Human Services (1990) Healthy People 2000. National Health Promotion and Disease Prevention Objectives. DHHS publication PHS 91–50212, Washington, DC.

44. Kaufman, P. K., MacDonald, J. M., Lutz, S. M. & Smallwood, D. M. (1997) Do the Poor Pay More for Food? Item Selection and Price Differences Affect Low-Income Household Food Costs. U.S. Department of Agriculture, Economic Research Service, AER-759.

45. Kaufman P. K. Rural poor have less access to supermarkets, large grocery stores. Rural Dev. Perspect. 1999;13:19-25

46. Mantovani, R. E., Daft, L., Macaluso, T., Welsh, J. & Hoffman, K. (1997) Authorized Food Retailer Characteristics Study: Technical Report IV (Authorized Food Retailers’ Characteristics and Access Study), U.S. Department of Agriculture, Food and Consumer Service, Office of Analysis and Evaluation.

47. Ohls, J. C., Ponza, M., Moreno, L., Zambrowski, A. & Cohen, R. (1999) Food Stamp Participants Access to Food Retailers: Final Report, U.S. Department of Agriculture, Food and Nutrition Service.

48. Kumanyika S. K., Krebs-Smith S. M. Preventive nutrition issues in ethnic and socioeconomic groups in the United States. Bendich A. Deckelbaum R. J. eds. Preventive Nutrition, Volume II: Primary and Secondary Prevention 2000 Humana Press Totowa, NJ (in press).

49. Subar A. F., Heimendinger J., Patterson B. H., Krebs-Smith S. M., Pivonka E., Kessler R. Fruit and vegetable intake in the U.S.: the baseline survey of the 5 A Day for Better Health Program. Am. J. Health Promot. 1995;9:352-360[Medline]

50. Byers T., Serdula M., Kuester S., Mendlein J., Ballew C., McPherson R. S. Dietary surveillance for states and communities. Am. J. Clin. Nutr. 1997;65(suppl.):1210S-1214S[Abstract/Free Full Text]

51. Li R., Serdula M., Bland S., Mokdad A., Bowman B., Nelson D. Trends in fruit and vegetable consumption among adults in 16 U.S. states: behavioral risk factor surveillance system, 1990–96. Am. J. Public Health 2000;90:777-781[Abstract/Free Full Text]

52. Serdula M. K., Coates R. J., Byers T., Simoes E., Mokdad A. H., Subar A. F. Fruit and vegetable intake among adults in 16 states: results of a brief telephone survey. Am. J. Public Health 1995;85:236-239[Abstract/Free Full Text]

53. Young, C. E. & Kantor, L. S. (1999) Moving Toward the Food Guide Pyramid: Implications for U.S. Agriculture, Economic Research Service, U.S. Department of Agriculture, AER-779.