Testing Variations on Family-Style Feeding To Increase Whole Fruit and Vegetable Consumption among Preschoolers in Child Care

Abstract

Background:

Preschoolers do not consume whole fruit and vegetables (FVs) in recommended quantities. Two strategies to increase FV intake were tested.

Methods:

One Head Start preschool participated. Two variations of family-style feeding were compared to usual practice: (1) Fruits, vegetables, and milk were served before the main meal (first course); and (2) fruits, vegetables, and milk were served before the main meal and meats and grains were removed from the table after the first serving (combination). A within-subject crossover design was used to test each condition for three meals. The amount of food served and consumed was weighed and converted to Child and Adult Care Feeding Program (CACFP) standard serving sizes for analysis.

Results:

Eighty-five children ages 3–5 participated. The sample was 81% Hispanic with diverse racial backgrounds. Thirty percent of the children were overweight. FV consumption was at CACFP recommended levels at baseline and remained consistent across conditions. The average amount served for each meal component was at or above CACFP recommendations for all foods except milk, which was consistently served in small portions. Meat and grains servings were frequently 2–3 times larger than CACFP recommendations. Milk consumption was significantly higher in the Combined intervention for two meals. Children ate significantly less meat during the Combined intervention for one meal.

Conclusions:

The intervention led to significant increases in milk consumption, which was the only underconsumed meal component. These strategies should be tested with children who have lower baseline intake of FVs.

Introduction

Increasing fruit and vegetable (FV) consumption in people over the age of 2 are key Healthy People 2020 objectives.¹ The diets of American preschool-age children are characterized by overconsumption of high-energy, low-nutrient-dense foods and underconsumption of whole FVs.^2,3 Nationally representative data collected in the Feeding Infants and Toddlers study indicate that over 30% of 3-year-olds do not consume any vegetables on a given day; those that do rely heavily on white potatoes (mainly fried), and less than 15% eat a dark green (11%) or yellow/orange (14%) vegetable each day.⁴ National Health and Nutrition Examination Survey (NHANES) findings further document very low vegetable intake among children of all ages.⁵ NHANES data also suggest that although preschool-age children consume enough total fruit, fiber intake could be increased by relying less on fruit juice and more on whole fruits.^5–7

The child care setting provides an important opportunity to influence young children's diets.⁸ Eighty-two percent of preschool-age children receive care outside of home,⁹ and over 40% of young children are in child care for over 35 hours per week.¹⁰ Importantly, many children consume up to two thirds of their meals and snacks while in child care.¹¹ Though there is some evidence that food consumed in child care is nutritionally superior to food consumed outside of care,¹² there is ample room for improvement. Ball and colleagues examined the dietary intakes of children in 20 North Carolina child care centers and found that intakes of whole grains, fruits, and vegetables were low, whereas consumption of saturated fat and added sugar were high.¹³ Another study of 3-year-olds found that they consumed adequate amounts of fruits and meat/meat alternates, but insufficient quantities of grains, vegetables, and dairy, while in child care.¹⁴ A study in New York City similarly found that only 17% of preschoolers ate at least half of the daily recommended intake for vegetables while in child care, which is the amount expected given the number of hours spent there.¹²

A variety of strategies have been proposed to increase children's consumption of healthy foods and decrease consumption of less-nutritious choices. One of the most effective ways to change consumption is to manipulate portion size. Rolls and colleagues found that 5-year-olds in the preschool setting (but not 3-year-olds) consumed more food as portion sizes increased.¹⁵ Similarly, Savage and colleagues increased the size of the entrée and found that children ages 3–5 decreased consumption of other meal components, including FVs.¹⁶ Other preschool meal manipulations have found that doubling an age-appropriate portion of an entrée significantly increased total calorie intake at lunch,¹⁷ and, conversely, reducing the energy density of an entrée by 30% significantly reduced caloric intake for the entire meal.¹⁸ Spill and colleagues found that doubling the portion size of carrots served with lunch resulted in a 47% increase in carrot consumption.¹⁹ Mathias and colleagues reported that vegetable and fruit consumption increased significantly among children (who liked the foods at baseline) when portion sizes were doubled.²⁰

Additional environmental factors that can influence children's dietary consumption are availability and accessibility of target foods. In a study looking at various foods in school vending machines, Rovner and colleagues found that availability of fruit and/or vegetables and chocolate and/or sweets was positively related to corresponding food intake in young children.²¹ Availability of healthy food in the home is important. Pearson and colleagues found a positive association between home availability and fruit and vegetable consumption in older children.²² Busick and colleagues found a correlation between the amount of produce purchased by the parents of preschoolers and preschooler willingness to try FVs. Greater willingness to try was found among children whose parents had the highest expenditures, suggesting that availability of FVs in the home may influence consumption.²³

Accessibility and visibility influence consumption of highly palatable foods (i.e., candy) in adults.²⁴ Making healthy foods more visible and accessible may be a promising strategy to increase consumption of them. One experimental study with college students found that students ate significantly more apple slices and carrots when they were placed in close proximity to the students and were visible in open bowls.²⁵

Children report that a barrier to consuming FVs is the presence of competing foods that are more palatable.²⁶ One strategy to increase consumption of healthier foods is to decrease their competition on the table. One method to achieve this is to serve the target foods first, before the rest of the meal is accessible. Rolls and colleagues found that low-energy-density preloads, such as salad before the meal, increase vegetable consumption,²⁷ increase satiety, and reduce subsequent energy intake at the meal^28,29 among adults. In one test of this strategy with young children, a first course of fruit led to an increase in fruit consumption; however, when vegetables alone were served first, children did not eat significantly more of them.³⁰

Based on the findings from the literature to date, two possible strategies emerge as potentially effective ways to increase consumption of whole FVs within the context of a child care family-style meal. The first is to serve the FVs as a “first course” before the rest of the meal is served in order to maximize the likelihood that children will eat those foods while waiting for the entrée. The second strategy is to further promote the FVs during the meal by serving one portion of the main dish and then removing it from the table so that it is no longer visible to the children when they consider taking second portions of any of the items. In theory, this should lead children who are still hungry to take more of the FVs instead of the main dish.

This study tested manipulations of the child care family-style feeding environment to promote fresh FVs by making them highly accessible, while also removing the competition from more energy-dense foods. Specifically, the following hypotheses were tested: (1) serving FVs as a first course will increase total consumption of FVs during the meal; and (2) keeping FVs visible on the table and removing the entrée after it is served will result in further increased FV consumption and reduced entrée consumption.

Methods

All procedures for this study were approved by the Yale University Institutional Review Board. A within-subjects crossover design was used. The repeated-measures intervention protocol was powered for 60 participants, an estimate based on previous research conducted with this age group.^18,31,32 One large Head Start preschool participated in the study. As with all Head Start preschools, the center participated in the USDA's Child and Adult Care Food Program (CACFP),³³ a federal nutrition program that provides funds for meals and snacks served in child centers serving low-income families. Informed consent was obtained from the parents of 97% of the preschool children (N = 85); these children were then enrolled in the study. Data collectors were trained to carry out the protocol, which was tested and revised using a similar preschool classroom.

Two feeding strategies were compared to usual practice (baseline): 1) FVs served before the main meal (first course); 2) FVs served before the main meal and only FVs and milk were kept on the table after initial serving (combination). Data were collected during lunch in five classrooms, 3 days per week, over a 3-week period. Three meals were tested in each condition: meal 1: chicken, rice, and beans, orange, broccoli, and milk; meal 2: turkey, tortilla, fruit salad, carrots, and milk; and meal 3: meatballs, sub roll, apples, green beans, and milk. These meals were selected from the center menu for use during the study because the FVs were served separately (not mixed into other dishes), making them easier to measure. Further, the staff reported that these meals were neither favorites nor disliked by most children.

As illustrated in Table 1, a 3-day paradigm was used to limit disruption to the regular meal routine. For 3 weeks, target meals were counterbalanced and served on Monday, Wednesday, and Friday, and nonstudy options were served Tuesday and Thursday.

Table 1.

Data Collection Paradigm

Week	Monday	Tuesday	Wednesday	Thursday	Friday
Week 1	Meal 1: Control	No observation	Meal 2: First course	No observation	Meal 3: First course+visibility
Week 2	Meal 1: First course+visibility	No observation	Meal 2: Control	No observation	Meal 3: First course
Week 3	Meal 1: First course	No observation	Meal 2: First course+visibility	No observation	Meal 3: Control

Measures

Dietary intake

Dietary intake data were collected using standard weighing methods, as described by Fisher and colleagues.³⁴ The weights of all plates, cups, and utensils were measured before the meal. Serving dishes containing meal components were placed on the table for children to self-serve. The teacher began by placing the serving dish on the scale and the observer noted the preweight of the dish. The teacher passed the dish to the first child. After the first child took a serving, the teacher helped the child place the dish back on the scale for postweighing. This procedure was repeated until all children had served themselves each component. When seconds were requested, they were weighed pre-post. When the meal was over, observers weighed all leftover food components separately. Spills were cleaned with a napkin and napkins were weighed.

Parent survey

Parents were asked to complete a survey on their child's demographic characteristics, including age, gender, and race/ethnicity as well as family participation in federal nutrition assistance programs other than CACFP. Surveys were available in both English and Spanish, and a Spanish-speaking researcher was available to assist participants as needed.

BMI

Head Start routinely collects height and weight data as part of a comprehensive assessment of child health. With parent consent, height and weight data for study participants were obtained from the center administrators and used to calculate BMI using age- and sex-specific criteria.³⁵

Procedures

First-course intervention

Bowls containing the FV meal components were placed on the table along with milk. Children served themselves in the manner described above. Once the fruit, vegetable, and milk had been served to all children, they began eating. After 5 minutes, the meat and grain dishes were served. These items were preweighed, so they needed only to be weighed after the meal (leftovers) or during the serving of second helpings.

First-course+visibility (combination)

The first-course intervention protocol was used. In addition, after first servings of the meat and grain components had been taken, these dishes were removed from the table and placed out of view of the children. Children were permitted to request additional servings of these items.

In each condition, teachers were instructed to behave as they usually would at mealtime by sitting with children, consuming the same food, and engaging the children in conversation. Teachers and children did not discuss the study more than necessary; children were told that researchers were present to learn about preschool meals, and that on the days researchers were present, children would leave their plates on the table at the end of the meal.

Statistical Analysis

Nutrient analysis was completed using ESHA Food Processor SQL Software (ESHA Research, Salem, OR). The amount of food served and consumed was converted from grams to CACFP servings to facilitate comparison across foods and meals. We used linear mixed models to examine the effect of the meal and intervention on the intake of fruits, vegetables, grains, meat, and milk. Models accommodated the unbalanced design resulting from child absences.

Results

The final sample included 85 children (47% male), ranging in age from 3 to 5 years. The sample was 81% Hispanic with diverse racial backgrounds: 36% white; 16% black; 37% mixed race; 5% Asian/Pacific Islander; and 5% American Indian. Thirty percent of children were overweight or obese (BMI >85th percentile). Three quarters of parents reported receiving Women Infant and Children benefits and 50% participated in the Supplemental Nutrition Assistance Program.

Dietary Intake

The mean (standard deviation; SD) energy intake during lunch was 230 (120) calories. Calories consumed varied by meal: the mean (SD) intake was 215 (153) calories for chicken and rice, 212 (85) calories for turkey wraps, and 265 (93) calories for meatball subs. The children consumed significantly more calories when served the meatball meal, compared to the other two meals (F_(663,2) = 13.4; p < 0.0001).

The proportion of standard CACFP servings of each meal component served and consumed was calculated for each condition (see Table 2). Children were often provided portions of meat and grains that were substantially larger than recommended. For example, the amount of meat recommended by the CACFP for this age group at lunch is 1.5 ounces (oz). Children were often offered portion sizes between 2.5 and 3.0 oz of meat at lunch; for the meatball meals, the average serving of meat was 3.1 oz. With respect to grains, the recommended serving is 0.5 oz; however, the average servings of rice and bread were almost 3 times this amount. In contrast, milk was consistently provided in smaller than CACFP recommended quantities.

Table 2.

Number of CACFP Servings of Each Meal Component Served and Eaten Across Conditions and Meals

	Oranges		Broccoli		Chicken		Rice		Milk
Meal 1	Served	Eaten	Served	Eaten	Served	Eaten	Served	Eaten	Served	Eaten
Baseline	1.6 ± 0.74	0.96 ± 0.72	1.2 ± 0.76	0.92 ± 0.81	1.4 ± 0.65	1.1 ± 0.69	1.0 ± 0.48	0.85 ± 0.52	0.74 ± 0.36	0.54 ± 0.46
First course	1.9 ± 0.75	1.1 ± 0.65	1.3 ± 0.93	0.97 ± 0.90	1.3 ± 0.64	1.1 ± 0.65	0.99 ± 0.48	0.81 ± 0.48	0.64 ± 0.35	0.50 ± 0.42
Combination	1.5 ± 0.70	0.89 ± 0.58	1.2 ± 0.90	1.0 ± 0.89	1.6 ± 0.69	1.3 ± 0.79	1.2 ± 0.52	0.96 ± 0.59	0.70 ± 0.43	0.57 ± 0.48

	Fruit salad		Carrots		Turkey		Tortillas		Milk
Meal 2	Served	Eaten	Served	Eaten	Served	Eaten	Served	Eaten	Served	Eaten
Baseline	1.1 ± 0.71^b	1.0 ± 0.75^b	1.3 ± 0.74	0.99 ± 0.93	1.7 ± 0.52	1.5 ± 0.66	1.9 ± 0.44	1.6 ± 0.74	0.66 ± 0.26	0.38 ± 0.32^a
First course	0.85 ± 0.39^a	0.80 ± 0.44^a	1.4 ± 0.85	1.1 ± 0.96	1.9 ± 0.43	1.5 ± 0.75	2.0 ± 0.57	1.7 ± 0.75	0.64 ± 0.31	0.46 ± 0.38^ab
Combination	1.2 ± 0.74^b	1.1 ± 0.78^b	1.4 ± 1.3	1.0 ± 1.3	1.9 ± 0.70	1.6 ± 0.72	2.1 ± 0.61	1.6 ± 0.85	0.70 ± 0.49	0.53 ± 0.46^b

	Apples		Green beans		Meatballs		Bun		Milk
Meal 3	Served	Eaten	Served	Eaten	Served	Eaten	Served	Eaten	Served	Eaten
Baseline	1.9 ± 1.1	1.4 ± 1.1	1.0 ± 1.0	0.83 ± 1.0	2.2 ± 0.88^b	1.9 ± 1.1^b	3.2 ± 0.76	2.3 ± 1.3	0.66 ± 0.26	0.40 ± 0.33^a
First course	2.0 ± 0.92	1.4 ± 0.92	1.1 ± 0.84	0.85 ± 0.90	2.2 ± 0.77^b	1.8 ± 0.94^b	3.0 ± 0.43	2.3 ± 1.3	0.76 ± 0.42	0.59 ± 0.45^b
Combination	2.1 ± 1.1	1.6 ± 1.1	1.2 ± 1.0	0.90 ± 1.0	1.8 ± 0.57^a	1.4 ± 0.81^a	3.2 ± 0.64	1.9 ± 1.6	0.77 ± 0.43	0.55 ± 0.43^b

CACFP serving requirements: one quarter cup each of fruit and vegetable (or half cup of fruit or vegetable), three-quarter cup milk, 1.5 ounces meat or meat alternate, and half ounce of grain.

For each food item, the differences in amounts served and eaten across the three conditions were tested for significance. Within each column, means with different superscripts indicate significantly differences between those means at the p < 0.01 level.

CACFP, Child and Adult Care Food Program.

To examine whether consumption of some meal components displaces consumption of others, we tested the correlations of consumption among all meal components. Greater consumption of grain was significantly associated with greater consumption of meat (r = 0.52; p < 0.001) and also associated with higher vegetable consumption (r = 0.14; p < 0.05). There were no inverse relationships observed, indicating that children who eat larger servings of some components do not necessarily consume smaller servings of others.

Intervention Effects

There were significant differences in the amount of food served and consumed across the three meals, so we ran separate models by meal. Each model examined the proportion of CACFP portions for all components served and consumed. Child ID and the intercept were entered as random effects; all other variables were entered as fixed effects. Final models controlled for age. BMI and gender were not significant predictors and so were not included in final models. Owing to the large number of comparisons, significance is only reported if p < 0.01.

The intervention had a significant effect on the calories consumed in the meatball meal; calories were significantly lower in the combined condition (mean = 242 ± 12), compared to the baseline (276 ± 12) and first-course conditions (278 ± 11). Calories consumed for the other two meals were not significantly different by condition.

For the chicken and rice meal, there was a significant effect for fruit consumed, with more fruit consumed during the first-course condition than baseline. Unexpectedly, there was also significantly more fruit consumed in the first-course condition than in the combined condition for this meal. The amount of fruit consumed reflected the finding that significantly more children took a second serving of oranges in the first-course condition (63%) than the combined (47%) and baseline (49%) conditions (F_(2,71) = 3.55; p < 0.05). There were no other significant intervention effects for this meal; however, over one third of the children took a second serving of broccoli in every condition (33% first course; 37% intervention; and 38% baseline).

For the turkey wrap meal, the first-course condition had significantly smaller fruit servings and consumption, compared to the other two conditions, contrary to the hypothesis. Consistent with expectations, there was also a significant intervention effect on milk consumed, with children drinking more milk in the combination condition compared to baseline. Over one quarter of the children took a second serving of carrots with this meal in the first-course condition (28%), compared to the combination (16%) and baseline (17%) conditions, although this difference did not reach statistical significance. Second servings of fruit salad were consistent across conditions (39% first course, 35% combination, and 44% baseline).

For the meatball meal, there were also differences by condition. There was a significant effect on milk consumed, with more milk consumed in the first-course and combination conditions, compared to baseline. There was also a significant effect for protein consumed, with less protein eaten in the combined condition, compared to the baseline and first-course conditions. Over one third of the children took a second serving of apples across all three conditions (39% first course; 35% combination; and 44% baseline), whereas second servings of green beans were less frequent (25% first course; 21% combination, and 20% baseline).

Discussion

The findings did not confirm the hypothesis that offering FVs before the rest of the meal would consistently lead to larger servings and greater consumption of these foods. The first-course strategy led to higher fruit consumption for only one meal and, surprisingly, led to lower fruit consumption in another meal. There was no evidence that the combination method changed fruit consumption, compared to baseline. For vegetables, servings and consumption were remarkably consistent across meals and conditions.

These findings are inconsistent with results of another study in the preschool setting testing this manipulation. Harnack and colleagues found that fruit intake was significantly higher when FVs were served before the remainder of the meal.³⁰ Other studies testing this strategy with adults in a lab setting have also achieved positive results.³⁶

One explanation for our null findings may be a ceiling effect. In this study, baseline FV intake was high and children consumed very close to one (and in many cases more than one) standard CACFP serving of both FVs under all conditions. Impressively, the students seemed to enjoy all of the FV options served across the meals. Consumption levels were right on target for broccoli and carrots and slightly under for green beans. Nearly 1.5 servings of apples were consumed, and servings eaten for fruit salad and oranges ranged from 0.8 to 1.1.

Further evidence of a ceiling effect is the finding that many children took second servings of the FVs across all conditions. Though some items were more popular (close to half of the children took seconds of oranges, compared to only one quarter taking seconds of green beans), it was clear that taking seconds of FVs was normative behavior in this center.

The reason for these excellent baseline rates of produce consumption may be that CACFP participating sites, and Head Start centers in particular (because of their mandate to monitor and address nutrition issues),³⁷ serve more FVs and generally healthier foods overall than non-CACFP participating sites.³⁸ Non-CACFP participating centers need only abide by state child care licensing regulations and laws regarding food served in child care, and in many states there are none.^39,40 Efforts to recruit preschools that do not participate in CACFP or those that are meeting CACFP requirements using poorer-quality or less-nutritious foods may allow for a better test of the environmental modifications made in this study.

Serving milk first and keeping it on the table throughout the meal had a significantly positive impact on milk consumption for two of the three meals. This suggests that one way to increase milk consumption is to pour the first serving before the main entrée is served, and to keep the milk on the table when the main entrée is removed. Notably, milk was the only meal component that was consistently lower than CACFP recommendations. The children were consuming approximately half a cup of milk during lunch throughout the study, which is lower than the three-quarter cup recommendation. The appropriate amount of milk to recommend for CACFP has been debated. The current CACFP standard is three quarters of a cup, but the Institute of Medicine (IOM) recommendations dropped that down to half a cup. The recent USDA proposed rule did not accept the IOM's position and retained the three-quarter cup serving.

Efforts to change the overall health profile of child care meals can focus on both increasing fruit, vegetables, and milk and decreasing excess servings of protein and grain. In the meals under study, protein and grains were consistently served and consumed in substantially larger-than-recommended quantities. The meatballs and bun served together as a meatball sub were 2–3 times the recommended protein and grain servings. The intervention of removing the meatball subs from the table after a first serving was successful in significantly decreasing the amount of the meatball and bun eaten by the children. This suggests that highly palatable dishes, such as the meatball sub, may be best served in single CACFP portions at first, and then removed from the table. Child care providers can then encourage second portions of the FVs, but provide seconds of the main dish only upon request.

Conclusions

The child care setting is an ideal setting in which to intervene with strategies to improve the diets of young children. This study provides tentative support for the combination of serving a first course of foods to be promoted and removing other foods from the table in cases where children are consistently underconsuming some foods and overconsuming others. Further research using these strategies with meals of poorer nutrition quality and with children who do not adequately consume FVs will clarify the effectiveness of these strategies.

Footnotes

Acknowledgments

This project was funded by the Robert Wood Johnson Foundation Healthy Eating Research program. The authors are grateful to Elizabeth Claydon, MPH, for her assistance with managing data collection and to Robin Masheb, PhD, for consulting on the project. The authors also thank the Yale University students who assisted with data collection. The authors have no corporate affiliations.

Author Disclosure Statement

No competing financial interests exist.

References

Office of Disease Prevention and Health Promotion. Healthy People 2020—Nutrition and weight status objectives. US Department of Health and Human Services: Washington, DC, 2010.

Guenther

, Dodd

, Reedy

, et al. Most Americans eat much less than recommended amounts of fruits and vegetables. J Am Diet Assoc, 2006; 106:1371–1379.

Kranz

, Hartman

, Siega-Riz

, et al. A diet quality index for American preschoolers based on current dietary intake recommendations and an indicator of energy balance. J Am Diet Assoc, 2006; 106:1594–1604.

Fox

, Condon

, Briefel

, et al. Food consumption patterns of young preschoolers: are they starting off on the right path?. J Am Diet Assoc, 2010; 110:S52–S59.

Kim

, Moore

, Galuska

, et al.; Division of Nutrition, Physical Activity, and Obesity, National Center for Chronic Disease Prevention and Health Promotion, CDC. Vital signs: Fruit and vegetable intake among children—United States, 2003–2010. MMWR Morb Mortal Wkly Rep, 2014; 63:671–676.

Lorson

, Melgar-Quinonez

, Taylor

. Correlates of fruit and vegetable intakes in US children. J Am Diet Assoc, 2009; 109:474–478.

Butte

, Fox

, Briefel

, et al. Nutrient intakes of US infants, toddlers, and preschoolers meet or exceed dietary reference intakes. J Am Diet Assoc, 2010; 110:S27–S37.

Institute of Medicine.. 2011. Early Childhood Obesity Prevention Policies. The National Academies Press: Washington, DC.

Capizzano

, Adams

. The hours that children under five spend in child care: Variation across states. The Urban Institute: Washington, DC, 2000.

10.

Capizzano

, Main

. Many young children spend long hours in child care. The Urban Institute: Washington, DC, 2005.

11.

Fox

, Glantz

, Geitz

, et al. Early childhood and child care study. USDA: Alexandria, VA, 1997.

12.

Erinosho

, Dixon

, Young

, et al. Nutrition practices and children's dietary intakes at 40 child-care centers in New York City. J Am Diet Assoc, 2011; 111:1391–1397.

13.

Ball

, Benjamin

, Ward

. Dietary intakes in North Carolina child-care centers: Are children meeting current recommendations?. J Am Diet Assoc, 2008; 108:718–721.

14.

Padget

, Briley

. Dietary intakes at child-care centers in central Texas fail to meet Food Guide Pyramid recommendations. J Am Diet Assoc, 2005; 105:790–793.

15.

Rolls

, Engell

, Birch

. Serving portion size influences 5-year-old but not 3-year-old children's food intakes. J Am Diet Assoc, 2000; 100:232–234.

16.

Savage

, Fisher

, Marini

, et al. Serving smaller age-appropriate entree portions to children aged 3–5 y increases fruit and vegetable intake and reduces energy density and energy intake at lunch. Am J Clin Nutr, 2012; 95:335–341.

17.

Orlet Fisher

, Rolls

, Birch

. Children's bite size and intake of an entree are greater with large portions than with age-appropriate or self-selected portions. Am J Clin Nutr, 2003; 77:1164–1170.

18.

Leahy

, Birch

, Rolls

. Reducing the energy density of an entree decreases children's energy intake at lunch. J Am Diet Assoc, 2008; 108:41–48.

19.

Spill

, Birch

, Roe

, et al. Eating vegetables first: The use of portion size to increase vegetable intake in preschool children. Am J Clin Nutr, 2010; 91:1237–1243.

20.

Mathias

, Rolls

, Birch

, et al. Serving larger portions of fruits and vegetables together at dinner promotes intake of both foods among young children. J Am Diet Assoc, 2011; 112:266–270.

21.

Rovner

, Nansel

, Wang

. Food sold in school vending machines is associated with overall student dietary intake. J Adolesc Health, 2011; 48:13–19.

22.

Pearson

, Biddle

, Gorely

, et al. Family correlates of fruit and vegetable consumption in children and adolescents: A systematic review. Public Health Nutr, 2008; 12:267–283.

23.

Busick

, Brooks

, Pernecky

, et al. Parent food purchases as a measure of exposure and preschool-aged children's willingness to identify and taste fruit and vegetables. Appetite, 2008; 51:468–473.

24.

Wansink

Environmental factors that increase the food intake and consumption volume of unknowing consumers. Annu Rev Nutr, 2004; 24:455–479.

25.

Privitera

, Creary

. Proximity and visibility of fruits and vegetables influence intake in a kitchen setting among college students. Environ Behav, 2013; 45:876–886.

26.

Nicklas

, Jahns

, Bogle

, et al. Barriers and facilitators for consumer adherence to the Dietary Guidelines for Americans: The HEALTH study. Acad Nutr Diet, 2013; 113:1317–1331.

27.

Rolls

, Roe

, Meengs

. Salad and satiety: Energy density and portion size of a first-course salad affect energy intake at lunch. J Am Diet Assoc, 2004; 104:1570–1576.

28.

Flood-Obbagy

, Rolls

. The effect of fruit in different forms on energy intake and satiety at a meal. Appetite, 2009; 52:416–422.

29.

Flood

, Rolls

. Soup preloads in a variety of forms reduce meal energy intake. Appetite, 2007; 49:626–634.

30.

Harnack

, Oakes

, French

, et al. Results from an experimental trial at a Head Start center to evaluate two meal service approaches to increase fruit and vegetable intake of preschool aged children. Int J Behav Nut Phys Act, 2012; 9:51.

31.

Leahy

, Birch

, Fisher

, et al. Reductions in entree energy density increase children's vegetable intake and reduce energy intake. Obesity, 2008; 16:1559–1565.

32.

Fisher

, Liu

, Birch

, et al. Effects of portion size and energy density on young children's intake at a meal. Am J Clin Nutr, 2007; 86:174–179.

33.

USDA, Food and Nutrition Service. Child and Adult Care Food Program. Available at www.fns.usda.gov/cnd/care/ProgramBasics/Meals/Meal_Patterns.htm Last accessed August 10, 2011

34.

Fisher

, Arreola

, Birch

, et al. Portion size effects on daily energy intake in low-income Hispanic and African American children and their mothers. Am J Clin Nutr, 2007; 86:1709–1716.

35.

CDC, Division of Nutrition, Physical Activity and Obesity. BMI percentile calculator for child and teens, English version. Available at http://apps.nccd.cdc.gov/dnpabmi Last accessed May 6, 2015.

36.

Rolls

, Roe

, Meengs

. Reductions in portion size and energy density of foods are additive and lead to sustained decreases in energy intake. Am J Clin Nutr, 2006; 83:11–17.

37.

US Deparmtent of Health and Human Services, Administration of Children and Families. Head Start program performance standards and other regulations. Child nutrition. 45-CFR-1304.23. Available at http://eclkc.ohs.acf.hhs.gov/hslc/standards/hspps/1304/1304.23%20Child%20nutrition.htm Last accessed May 6, 2015.

38.

Ritchie

, Boyle

, Chandran

, et al. Participation in the child and adult care food program is associated with more nutritious foods and beverages in child care. Child Obes, 2012; 8:224–229.

39.

Kaphingst

, Story

. Child care as an untapped setting for obesity prevention: State child care licensing regulations related to nutrition, physical activity, and media use for preschool-aged children in the United States. Prev Chronic Dis, 2009; 6:A11.

40.

Benjamin

, Cradock

, Walker

, et al. Obesity prevention in child care: A review of U.S. state regulations. BMC Public Health, 2008; 8:188.