Most Americans do not eat enough fruits and vegetables ⁽¹⁾ in spite of the demonstrated benefits of consuming adequate amounts⁽²⁾. The Dietary Guidelines for Americans, 2010 encourage Americans to increase intake of fruits and vegetables to add important nutrients that are frequently lacking from their diets, reduce the risk of many leading causes of death and illness, and aid in healthy weight management by substituting them for higher calorie foods⁽²⁾. However, in 2007–2010, among individuals aged 1 year and older, approximately 13% met their age and sex-specific U.S. Department of Agriculture (USDA) Food Patterns vegetable intake recommendations and 25% met fruit recommendations⁽³⁾.

To date, little research has been done using nationally representative data on how people incorporate fruits and vegetables into their day^(4–6), especially among those who consume amounts recommended by the USDA Food Patterns versus those who do not. Given this, we explored how U.S. children (aged 2–18 years) and adults (aged 19 years and over) who consumed recommended amounts of fruits and vegetables on a given day incorporated fruits and vegetables into their diets compared to those who did not using nationally representative data from 2007 – 2010. We identified where (source), when (meal), and in what form(s) individuals consumed their fruits and vegetables among those who consumed recommended amounts on a given day versus those who did not.

We used data from the 2007–2010 National Health and Nutrition Examination Survey (NHANES) to estimate fruit and vegetable intake among individuals aged 2 years and older. NHANES is a nationally representative survey of the non-institutionalized population that is conducted in 2-year phases using a stratified multi-stage probability design. Details on the survey and sampling are described elsewhere⁽⁷⁾. Respondents participated in a household questionnaire and a physical examination, including a 24-hour dietary recall, at the Mobile Examination Center (MEC). Trained interviewers collected 24-hour dietary recalls using the USDA automated multiple-pass method by proxy for those aged 1–5 years, with proxy assistance for those aged 6–11 years, and directly from participants aged 12 years. A total of 20,686 individuals were interviewed and 20,015 completed a physical examination in the MEC. MEC response rates were 75.4% and 77.3% for NHANES 2007–2008 and 2009–2010, respectively. There were 17,571 individuals aged 2 years and older with 1 day of reliable 24-hour recall data who participated in NHANES 2007–10 and were included in analyses. All participants in NHANES provided written informed consent.

For all foods or beverages participants reported, we used USDA’s Food Patterns Equivalent Database (2007–2008 and 2009–2010) to disaggregate each food or beverage item into its individual food components^{(8, 9)}. The total amount of fruits and vegetables for each individual, in cup equivalents, were summed across all foods and beverages that were reported. Cup-equivalents of fruits and vegetables consumed were calculated separately by the meal, source, and form of each food item. Categories of meal, source, and form were derived from survey responses. Of the 20 different types of eating occasions⁽¹⁰⁾, we classified foods and beverages into 5 meal categories: breakfast, lunch, dinner, snack, and other eating occasions such as brunch and drinking occasions. Of the 23 different source options⁽¹⁰⁾, we collapsed foods and beverages into four mutually exclusive categories: 1) Retail (e.g., retail stores, mail order purchases); 2) Restaurants (e.g., restaurants, bars, street vendors, and sport, recreation or entertainment facilities); 3) Cafeterias (e.g., school and other institutional cafeterias, residential dining facilities, vending machines, common snack tray, childcare center, family/adult day care center); and 4) Other (e.g., soup kitchens or food pantries, Meals On Wheels, other community food programs, home grown or gifted, and any other reported source). Lastly, of the 16 different types of food and beverage combinations,⁽¹⁰⁾ we collapsed foods and beverages into four mutually exclusive form categories: 1) raw/alone (e.g. whole or cut up apple or carrots); 2) a salad or sandwich ingredient; 3) an ingredient in a fruit or vegetable mixed dish (e.g. salsa, hummus, tomato sauce on pizza or spaghetti); and 4) all other forms.

Individuals were classified as consuming the recommended amounts based on the total cup-equivalents of fruits and vegetables consumed on the recall day, respectively. Recommended amounts (Table 1) ranged from 1 to 2 cup-equivalents of fruit and 1 to 3 cup-equivalents of vegetables, were age- and sex-specific, and appropriate for individuals with less than 30 minutes of daily physical activity.^{(11, 12)} Ratio of means were calculated to compare the relative proportion of fruits and vegetables consumed at different meals, from different sources, and in different forms among those who consumed recommended amounts versus those who did not. Ratio of means were estimated by dividing each group’s mean amount of fruit and vegetable intake from different meals, sources, and forms by the group’s mean total fruit and vegetable intake. Ratio of mean estimates and p-values for differences in the ratio of means using t-tests for linear contrasts were calculated using PROC RATIO in SAS 9.3.2 and SAS Callable SUDAAN 10.1 to account for the complex sampling design. Analyses were conducted separately for those 2–18 years and those aged 19 years and over. Due to the number of statistical tests performed, a Bonferroni adjustment was used to set the significance level at 0.001.

Children who consumed the recommended amounts of fruits or vegetables ate or drank about an additional cup each of fruits and vegetables compared to those children who did not consume the recommend amounts (Table 2). Adults who consumed the recommended amounts ate or drank an additional 2.5 cups of fruits and 2.6 cups of vegetables compared to those who did not consume recommended amounts.

In general, the proportion of fruits and vegetables consumed at different meals, from different sources, and in different forms were similar between those who consumed recommended amounts and those who did not. Individuals who reported consuming recommended amounts of fruits and vegetables incorporated them into their diets in similar ways to those who did not. Compared to those who did not report consuming recommended amounts, those who did reported consuming an additional two cup-equivalents of both fruit and vegetables which equated to six times as much fruit and three times as many vegetables.

The findings from this study are consistent with prior work^(4–6). For example, a NHANES 2007–2010 analyses indicated that about 90% of fruit and 70% of vegetables are consumed at home⁽⁴⁾. A survey of 2,000 households in the U.S. indicated that two thirds of vegetables were eaten at dinner and fruits were consumed at all eating occasions but especially breakfast (44%) and lunch (24%)^(5). Similarly, most fruit and vegetables were consumed at home and a large proportion was consumed fresh. A small-scale study among fourth to sixth graders also found higher fruit intake was associated with starting the day with juice or fruit⁽⁶⁾. While the above studies have examined how small populations or subpopulations obtain fruits and vegetables and when they eat them^(4–6), this study differs from others in that we used updated nationally representative data to explore how intakes of fruit and vegetables on a given day differed between those who consumed recommended amounts and those who did not.

Limitations of this study include reliance on the Food Patterns Equivalents Databases to estimate cup equivalents of fruits and vegetables for each food reported. The Food Patterns Equivalents Databases disaggregates foods reported, including composite foods and mixtures, into their nutritionally relevant components and translates these amounts into cup equivalents consistent with federal food guidance. Although, the databases are uniquely designed to examine diets reported via 24-hour recall methodology in NHANES, assumptions are made regarding the consistency of recipes and cooking methods to disaggregate foods into appropriate components. A second limitation of the analysis is that we did not examine whether usual intake of fruits and vegetables exceeded recommendations. One 24-hour dietary recall does not capture day-to-day variation in diets and thus does not represent usual dietary intake but is accurate for estimating differences in mean intakes as we did in these analyses⁽¹³⁾. Third, individuals may not report their food consumption accurately by weight status^{(13, 14)} and possibly between meals⁽¹⁵⁾. NHANES may underreport mean energy intake by 11% compared to total energy expenditure estimated using doubly labeled water⁽¹⁶⁾. Normal weight adults underreported mean energy intake by <3%⁽¹⁶⁾. Fruit intake estimates especially may be higher than reported given that approximately a third of fruit intake occurs as snacks. Next, while most Americans should consume more fruits and vegetables to meet recommendations, increased intake should be balanced within calorie needs. Fruit and vegetable intake recommendations account for calorie needs, however, individual needs may differ from those in the USDA Food Patterns. Finally, we examined patterns among children and adults overall even though fruit and vegetable intake varies within these broad age groups⁽³⁾. In general, we observed no significant differences when we stratified by age in children (see Appendix Table) but further research may be needed to explore patterns by age.

Eating patterns by meal, source, and form for those who consumed recommend amounts of fruits and vegetables on a given day versus those who did not were very similar and mainly differed in the amount consumed. Overall, this indicates that people may not need to shift their current habits substantially in regards to when, from where, and how they eat in order to meet recommendations. Increasing the amount of fruit consumed by 1 cup equivalent equates to eating a banana or a small apple⁽¹¹⁾. Increasing the amount of vegetables consumed by 1 cup equates to eating 12 baby carrots or a large tomato⁽¹²⁾. Small shifts in everyday eating patterns like these could help the majority of the U.S. population who are not consuming enough fruits and vegetables reduce their risk of many leading causes of illness and death, add important nutrients to their diet, and help with weight control⁽²⁾.