3 page paper due in 7 hours

ª 2016 by the Academy of Nutrition and Dietetics. J

RESEARCH

Original Research: Brief

Eating School Lunch Is Associated with Higher Diet Quality among Elementary School Students

Lauren E. Au, PhD, RD; Nila J. Rosen, MPH; Keenan Fenton, MA; Kenneth Hecht, JD; Lorrene D. Ritchie, PhD, RD

ARTICLE INFORMATION

Article history: Submitted 16 September 2015 Accepted 13 April 2016 Available online 21 May 2016

Keywords: Diet quality Healthy Eating Index-2010 (HEI-2010) School breakfast School lunch Children

2212-2672/Copyright ª 2016 by the Academy of Nutrition and Dietetics. http://dx.doi.org/10.1016/j.jand.2016.04.010

ABSTRACT Background Few studies have assessed the dietary quality of children who eat meals from home compared with school meals according to the 2010 Dietary Guidelines for Americans. Objective The objective of this study was to examine diet quality for elementary school students in relation to source of breakfast and lunch (whether school meal or from an outside source). Design An observational study was conducted of students in 43 schools in San Diego, CA, during the 2011-2012 school year. Participants/setting Fourth- and fifth-grade students (N¼3,944) completed a diary- assisted 24-hour food recall. Main outcome measures The Healthy Eating Index-2010 (HEI-2010) scores of children who ate breakfast and lunch at school were compared with the HEI-2010 scores of children who obtained their meals from home and a combination of both school and home. Statistical analysis Analysis of variance, c2 test, and generalized estimating equation models adjusted for age, sex, race/ethnicity, grade, language, and school level clustering were performed. Results School lunch eaters had a higher mean�standard deviation overall diet quality score (HEI-2010¼49.0�11.3) compared with students who ate a lunch obtained from home (46.1�12.2; P¼0.02). There was no difference in overall diet quality score by breakfast groups. Students who ate school breakfast had higher total fruit (P¼0.01) and whole fruit (P¼0.0008) scores compared with students who only ate breakfast obtained from home. Students who ate school foods had higher scores for dairy (P¼0.007 for breakfast and P<0.0001 for lunch) and for empty calories from solid fats and added sugars (P¼0.01 for breakfast and P¼0.007 for lunch). Conclusions Eating school lunch was associated with higher overall diet quality compared with obtaining lunch from home. Future studies are needed that assess the influence of the Healthy Hunger-Free Kids Act on children’s diet quality. J Acad Nutr Diet. 2016;116:1817-1824.

F OODS OBTAINED AT SCHOOL PROVIDE AN IMPOR- tant nutritional contribution to the diets of more than 30 million children in the National School Lunch Program1 and more than 13 million children

in the School Breakfast Program.2 Despite recent growth in these numbers, largely reflecting the recent recession, not all students participate in school meals programs, including many who are eligible for free or reduced-price meals.3 For children who do participate, school meals can account for approximately half of their daily caloric intake, with school lunches contributing up to 31% of daily calories and school breakfasts contributing up to 22%.4

The Healthy Hunger-Free Kids Act (HHFKA) of 20105

was instrumental in improving the nutritional quality of school meals by requiring them to meet the 2010 Dietary Guidelines for Americans (DGA).6 These new school meal regulations included minimum and maximum calorie

allowances, increased fruit and vegetable servings, whole grains, and the elimination of high-fat milk.5 However, although these regulations improved the nutritional con- tent of school breakfast and school lunches, they did not address foods obtained from home. Although there have been a number of studies that have

documented that the nutritional quality of meals from home is lower compared with school meals,7-15 to our knowledge there have been no studies that have assessed whether school meals contribute to the quality of children’s daily diet using the current Healthy Eating Index 2010 (HEI-2010), which is an indicator of conformance to federal dietary guidance.16 Therefore, the objective of this study was to compare the diet quality scores of students in the San Diego, CA, area who ate breakfast and lunch from school compared with those who obtained their meals from home and a combination of both school and home.

OURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1817

RESEARCH

METHODS School and Student Level Participation Data were collected in 2011-2012 as part of a cluster ran- domized, controlled trial17 to evaluate the effectiveness of a school-based educational intervention to promote fruit and vegetable intake and physical activity among fourth- and fifth-grade students in low-resource elementary schools in San Diego and Imperial Counties in California. Schools were excluded from recruitment based on having no fourth- or fifth-grade classes, fewer than 30 students per grade, having received the planned intervention or a similar intervention or other strong wellness activities in the year prior, and char- acteristics that would limit the generalizability of findings (ie, location bordering Mexico or being a juvenile detention school). For inclusion, schools needed to have �50% of the student body qualify for free or reduced-priced school meals. From an initial list of 221 elementary schools, 131 were eligible for participation based on these criteria. The first 45 schools that met eligibility criteria were recruited based on sample size calculations of the parent study.17 Schools were also offered compensation for participating. Each school received $500 and each participating teacher received $200 to compensate them for their time. Subsequently one school discontinued the study due to a campus fire. For the breakfast analysis, one school was excluded because it did not provide school breakfast. The resulting breakfast analytic sample consisted of data collected at baseline from 3,944 children in 43 elementary schools in five school districts. For the lunch analysis, one school was excluded because it did not provide lunch service and another school was excluded due to the lunch period being altered by a shortened school day. The lunch analytic sample consisted of data collected at baseline from 3,219 children in 42 schools in five school districts. All study protocols were reviewed and approved by the institutional review board at the Public Health Institute, an independent nonprofit organization focusing on health promotion, which spearheaded the original study. Parents were sent an information letter about the study and an opt- out consent form. Students were read a verbal assent by research staff before the food diary was completed.

Dietary Intake Student-level data on dietary intake during a school day was collected using diary-assisted 24-hour recall interviews con- ducted by trained interviewers during Tuesday through Friday of the school week. The diary-assisted 24-recall method is a blending of two dietary assessment methods that maximizes the strengths of both the food record and the 24-hour recall.18 This method has shown superior validity in children compared with unassisted recall interviews19 and has been successfully used in a number of studies with elementary schooleaged children.20-22 Before recording intake in their food diary, students received training from research staff about how to keep a 1-day food diary, which asked for details of meals and snacks, source of foods and beverages, food description, and amounts consumed. Each child received a set of measuring cups and spoons and two- dimensional pictures for portion size measurement. Within 2 days of completing the food diary, trained research staff conducted individual 24-hour recall interviews at school with each child using the multiple-pass method23 and the food

1818 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

diary. Food models were used to clarify portion sizes and details on forgotten foods were elicited. Information collected from students on school foods were

supplemented by interviewing school foodservice staff, as described previously.24 Obtaining nutrition information about school foods and beverages from school foodservice staff was used to enhance the accuracy of the diary-assisted method. Foods and beverages were coded using the US Department of Agriculture Food and Nutrient Database for Dietary Studies (version 3.0, 2008). The food and beverage source was queried and recorded by interviewers so that for each item consumed, students were asked whether it was obtained and consumed at school or elsewhere (home, friend’s home, fast-food restaurant, other restaurant, or other). The exposure of interest, location of where the breakfast or lunch meal was obtained, was categorized as school, home, or a combination of both school and home. For the purposes of the analysis, if all breakfast or lunch items reported by students were obtained from school, then they were classified as school breakfast or school lunch, respec- tively; if all breakfast or lunch items were obtained from home, then they were classified as breakfast or lunch ob- tained from home, although items may have been obtained elsewhere (eg, friend’s home, fast-food or pizza restaurant, or other restaurant); and if some breakfast or lunch items came from both school and obtained from home, then they were classified as students who ate breakfast or lunch from school and obtained from home. The outcome of interest, diet quality, was quantified using

the HEI-2010.16 The HEI-2010 is a valid and reliable measure of conformance to the 2010 DGA.25 A total of 12 components (total fruit, which includes 100% juice; whole fruit, which excludes juices; total vegetables; greens and beans; whole grains; dairy; total protein foods; seafood and plant proteins [includes seafood, nuts, seeds, soy products, and beans and peas when the total protein foods standard is otherwise not met]; fatty acids; refined grains; sodium; and empty calories) were scored separately and then summed for a maximum score of 100 signifying the highest quality diet. The standards used for scoring were expressed as a percent of calories or per 1,000 kcal, allowing for diet quality comparisons across a wide range of total energy intakes. Student demographic data (age, sex, race/ethnicity, and language spoken at home) were obtained by self-report on a student questionnaire. Breakfast was defined as items reported as breakfast (food

or caloric beverage) between 4:00 AM and up to half an hour before the school lunch start time, with items reported as lunch items excluded. Instances of seemingly aberrant times (eg, between 1:00 AM and 4:00 AM) were hand checked with the diaries and any obvious errors were corrected. This method for classifying breakfast has been used else- where.26,27 Students who reported not eating breakfast were defined as students who recorded intake of 0 kcal at break- fast. This classification resulted in 3,515 students who re- ported eating breakfast. For lunch, any item reported as lunch and as eaten at school within 60 minutes of the school lunch period was classified as a lunch item. The window for lunch was expanded to accommodate children who may have eaten lunch items after the lunch period and also for inac- curacies in children’s reports of the time of day. Students who reported not eating lunch were defined as students who recorded intake of 0 kcal at lunch. Lunch eaters recorded

November 2016 Volume 116 Number 11

RESEARCH

any intake of calories at lunch. This classification resulted in 3,203 students who reported eating lunch. No foods or bev- erages were double counted as breakfast or lunch items.

Statistical Analysis Associations of student characteristics with breakfast groups (school, school and home, or home) were assessed by c2 test of independence or analysis of variance (ANOVA). Student characteristics assessed included age; sex; race/ethnicity (Hispanic, non-Hispanic white, non-Hispanic black, Asian, and multiracial or other); Spanish-speaker (language usually spoken at home); and grade level. Demographic character- istics that were significantly different between breakfast or lunch groups were controlled for in the main analyses. As- sociations between breakfast groups and total HEI-2010 and HEI-2010 component scores were assessed using generalized estimating equation (GEE) models to adjust for age, race/ ethnicity, grade, language, and school level clustering. Asso- ciations between lunch groups and total HEI-2010 and HEI- 2010 component scores were assessed using GEE models to adjust for sex, race/ethnicity, grade, language, and school level clustering. ANOVA was carried out to determine whether the source of breakfast or lunch was associated with diet quality. For those outcomes found to have significant differences by ANOVA, a post hoc multiple comparison test using a Bonferroni approach at a 5% procedure-wise error rate was used. Associations of percent of total calories from school meals with HEI-2010 scores were also assessed using GEE models adjusted for all the demographic covariates.

Table 1. Comparison of demographic characteristics by breakfast controlled school-based intervention trial conducted in 2011-201

Breakfast Grou

Breakfast from School (n[965)

Breakfast from Sc and Home (n[559)

Characteristics of sampled fourth- and fifth-grade students (n¼ ��������������������mean�sta

Age (y) 9.7�0.7 9.8�0.7 ���������������������������

Male 49.3 47.5

Race/ethnicity

Hispanic 56.6 53.7

Non-Hispanic white 7.1 7.7

Non-Hispanic black 10.9 12.3

Asian 8.5 9.1

Multiracial/otherd 17.0 17.2

Spanish speaker 52.4 49.5

Fifth grade 44.5 51.7

aIncludes 429 students who reported not eating breakfast. bBecause of missing values, the total is not the same for all variables: age (n¼3,928), sex (3,91 cFor unadjusted comparisons of breakfast groups, analysis of variance was used for age and c dMultiracial/other for race includes American Indian/Alaskan Native, Native Hawaiian/Pacific Isla *Significant at P<0.05.

November 2016 Volume 116 Number 11 JO

Because items obtained from a fast-food restaurant were grouped in the same category as foods from home and they could potentially influence diet quality negatively, an addi- tional examination of the main analyses excluding students who consumed any fast-food items from the analyses was conducted. Data were analyzed using SAS version 9.4.28

A P value of <0.05 was considered statistically significant.

RESULTS In the breakfast sample, the average age of students was 9.8 years. Roughly half of students were boys (49.3%) and in the fifth grade (50.5%). Half of students were Hispanic (49.2%) and 41.7% of students self-identified as Spanish-speakers. Of the 3,944 elementary students included in the breakfast sample, 3,515 (89.1%) consumed breakfast, whereas 3,203 of the 3,219 students in the lunch sample (99.5%) consumed lunch, according to the defined breakfast and lunch periods. The student characteristics by breakfast groups are presented in Table 1. Compared with students who ate breakfast from home, students who ate breakfast from school were younger (P¼0.02), more likely to be Hispanic (P<0.0001), Spanish speakers (P<0.0001), and fourth graders (P¼0.0002). The student characteristics by lunch groups are presented in Table 2. Compared with students who ate lunch from home, students who ate lunch from school were more likely to be boys (P<0.0001), Hispanic (P<0.0001), Spanish speakers (P<0.0001), and fourth-graders (P¼0.004). Total HEI-2010 scores were not significantly different be-

tween the breakfast groups (Table 3). Students who

groups of students at baseline as part of a cluster randomized 2 in California (n¼3,944)a

p

Total (n[3,944)b P valuec

hool Breakfast from Home (n[1,991)

3,944)

ndard deviation ��������������������!

9.8�0.7 9.8�0.7 0.02* �� % �����������������������������!

49.2 49.3 0.75

<0.0001*

44.5 49.2

16.5 12.5

7.2 9.1

8.2 8.5

23.6 20.7

35.3 41.7 <0.0001*

52.4 50.5 0.0002*

6), race/ethnicity (3,944), language (3,917), and grade (3,944). 2 test was used for all other variables. nder, or multiracial.

URNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1819

Table 2. Comparison of demographic characteristics by lunch groups of students at baseline as part of a cluster randomized controlled school-based intervention trial conducted in 2011-2012 in California (n¼3,219)a

Lunch Group

Total (n[3,219)b P valuec

Lunch from School (n[2,286)

Lunch from School and Home (n[255)

Lunch from Home (n[662)

Characteristics of sampled fourth- and fifth-grade students (n¼3,219) ��������������������mean�standard deviation��������������������!

Age (y) 9.7�0.7 9.8�0.7 9.8�0.7 9.7�0.7 0.05 ����������������������������%����������������������������!

Male 51.2 41.7 43.1 48.8 <0.0001*

Race/ethnicity <0.0001*

Hispanic 50.6 43.9 40.3 47.9

Non-Hispanic white 10.6 12.2 21.8 13.1

Non-Hispanic black 8.8 14.5 7.1 8.9

Asian 9.1 10.2 6.8 8.7

Multiracial/otherd 21.0 19.2 24.0 21.5

Spanish speaker 43.0 37.3 29.4 39.6 <0.0001*

Fifth grade 48.0 52.6 55.0 49.9 0.004*

aIncludes 16 students who reported not eating lunch. bBecause of missing values, the total is not the same for all variables: age (n¼3,213), sex (n¼3,210), race/ethnicity (n¼3,219), language (n¼3,219), and grade (n¼3,219). cFor unadjusted comparisons of lunch groups, analysis of variance was used for age and c2 test was used for all other variables. Because of missing values, the total is not the same for all variables. dMultiracial/other for race includes American Indian/Alaskan Native, Native Hawaiian/Pacific Islander, or multiracial. *Significant at P<0.05.

RESEARCH

consumed school breakfast had higher scores for total fruit (P¼0.01) and for whole fruit (P¼0.0008) than students who ate breakfast at home. Higher dairy scores (P¼0.007) and empty calorie scores (P¼0.01) were observed for school breakfast eaters compared with students who ate breakfast from home, indicating higher dairy consumption and lower empty calorie consumption for students who ate school breakfast. Seafood and plant proteins (P¼0.001) and sodium (P¼0.02) scores were lower for the school breakfast eaters compared with the other groups, indicating lower seafood, nuts, seeds, beans, and peas consumption and higher sodium consumption for students who ate school breakfast. School lunch eaters had higher overall diet quality

compared with students who ate lunch from home (P¼0.02) (Table 4). Dairy (P<0.0001) and empty calorie scores (P¼0.007) were higher for school lunch eaters compared with the other groups, indicating diets higher in dairy and lower in empty calories in school lunch eaters. Seafood and plant proteins (P<0.0001) and fatty acids (P¼0.02) scores were lower for the school lunch eaters compared with the lunch- from-home group, indicating lower consumption of sea- food, nuts, seeds, beans, and peas and lower consumption ratio of unsaturated to saturated fat for students who ate school lunch. Refined grain (P¼0.03) scores were higher for school lunch eaters compared with lunch-from-home eaters, indicating a lower intake of refined grains in school lunch eaters. Overall, percent of total calories from school breakfast and

school lunch was positively associated with total diet quality

1820 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

score (b�standard error [SE]¼0.08�0.01; P<0.0001). This positive relationship was also seen separately for school lunch (b�SE¼0.10�0.02; P<0.0001), but not for school breakfast (b�SE¼0.03�0.02; P¼0.17) (data not shown). Excluding students who obtained fast-food items from home from the analyses did not affect the results.

DISCUSSION In this study involving nearly 4,000 elementary school stu- dents in southern California, school lunch consumption was associated with higher overall diet quality and both school breakfast and lunch contributed to higher dairy and empty calorie quality scores, indicating a healthier intake. However, on average, all students achieved about half of the maximum HEI-2010 diet quality score, a value based on 12 key recom- mendations in the 2010 DGA, showing the need for nutrition- related improvement for all students, regardless of where their meals come from. Students who consumed school breakfast had higher

scores for total fruit, whole fruit, dairy, and empty calories, and lower seafood and plant proteins and sodium scores than students who ate breakfast at home. These findings are consistent with a 2013 study of children aged 6 to 17 years from the 2003-2006 National Health and Nutrition Exami- nation Survey sample, which suggested that dietary quality may be higher among School Breakfast Program participants than nonparticipants.29 School breakfast consumption was not associated with higher overall diet quality. The lack of a

November 2016 Volume 116 Number 11

Table 3. Characteristics of students’ diet qualities by breakfast groups of students at baseline as part of a cluster randomized controlled school-based intervention trial conducted in 2011-2012 in California (n¼3,515)a

Dietary characteristic

Breakfast from School (n[965)

Breakfast from School and Home (n[559)

Breakfast from Home (n[1,991) P valueb

����������������������mean�standard deviation����������������������! Healthy Eating Index-2010 total scorec

49.6�11.0 49.8�10.8 47.4�11.9 0.20 Healthy Eating Index-2010 component scorec

Adequacy

Total fruit 3.2�1.9x 3.4�1.8x 2.8�2.0y 0.01* Whole fruit 3.4�2.1x 3.5�2.1x 2.6�2.3y 0.0008* Total vegetables 2.1�1.7 1.9�1.6 2.1�1.6 0.06 Greens and beans 0.4�1.2 0.4�1.1 0.4�1.2 0.40 Whole brains 2.4�2.6 2.5�2.6 2.6�2.9 0.24 Dairy 7.8�2.9x 7.8�2.8x 7.3�3.1y 0.007* Total protein foods 3.6�1.5 3.7�1.5 3.6�1.6 0.69 Seafood and plant proteins 1.0�1.8x 1.4�2.0y 1.4�2.0y 0.001* Fatty acids 4.3�3.4 4.1�3.2 4.0�3.3 0.35 Moderation

Refined grains 3.9�3.5 4.3�3.4 4.2�3.6 0.09 Sodium 4.0�3.3x 4.3�3.1y 4.2�3.3y 0.02* Empty calories 13.4�5.4x 12.7�5.2y 12.1�5.4y 0.01* aExcludes students who reported not eating breakfast (n¼429). bP values based on analysis of variance using generalized estimating equation models adjusting for age, race/ethnicity, grade, language usually spoken at home, and cluster design (type 3 or Wald). Means sharing a common superscript (x, y, z) are not significantly different from each other using a Bonferroni approach at a 5% procedure-wise error rate. Different superscripts indicate statistical differences between groups using a post hoc Bonferroni-Holm multiple comparison test. cMaximum total Healthy Eating Index-2010 total score is 100; component scores can range from 0 to 5 (for total fruit, whole fruit, total vegetables, green and beans, total protein foods, and seafood and plant proteins) or 0 to 10 (for whole grains, dairy, fatty acids, refined grains, and sodium), and 0 to 20 for empty calories. A higher number signifies a healthier intake. *Significant at P<0.05.

RESEARCH

statistically significant difference in diet quality between the breakfast groups may be attributable to the fact that break- fast makes a relatively smaller caloric contribution to the entire 24-hour period than other meals. School lunch eaters had higher overall diet quality

compared with students who ate lunch from home. In our study, school lunch eaters consumed diets that were higher in dairy-rich foods, which may be attributable to school lunches containing more dairy products, such as providing skim or low-fat milk, yogurts, and low-fat cheeses. This was observed in a 2012 study of more than 2,000 second-grade students from a large suburban area where home lunches were significantly less likely to contain dairy products compared with school lunches.7 School lunch eaters also consumed diets that were lower in empty calories from solid fats and added sugars and lower in refined grains than students who ate lunches from home. This finding may be attributable to school lunches containing less foods with refined grains, solid fats, and added sugars, which is consistent with a 2014 study of 337 elementary and in- termediate school students from Texas that showed that lunches obtained from home contained more sweetened beverages, snack chips, desserts, and refined grains than

November 2016 Volume 116 Number 11 JO

school lunches.8 School lunch eaters had lower seafood and plant protein and fatty acid component scores compared with students who obtained their lunch from home, which may indicate that school meals should incorporate more nutritious, child-friendly food items rich in lean proteins and poly- and monounsaturated fatty acids, such as tuna salad, nuts, seeds, and beans. The higher HEI-2010 scores for stu- dents who ate school lunch may be due to the fact that more students who ate school lunch also ate school breakfast compared with students who did not eat either school breakfast or school lunch, which is also supported by our study findings that higher percent of calories from school meals resulted in a higher diet quality index score. Prior studies have not shown a clear indication of the

direction of association between students who consume school meals and overall diet quality. Two studies suggest that overall diet quality was equivalent for school lunch participants vs nonparticipants.14,30 A study conducted in pre-kindergarten and kindergarten children in four schools in rural Virginia found that the sodium content of school lunches was higher than for lunches from home.10

However, other studies have shown that school meals tend to be of higher diet quality than food obtained from

URNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1821

Table 4. Characteristics of students’ diet qualities by lunch groups of students at baseline as part of a cluster randomized controlled school-based intervention trial conducted in 2011-2012 in California (n¼3,203)a

Diet characteristic

Lunch from School (n[2,286)

Lunch from School and Home (n[255)

Lunch from Home (n[662) P valueb

��������������mean�standard deviation��������������! Healthy Eating Index-2010 total scorec

49.0�11.3x 47.5�10.7xy 46.1�12.2y 0.02* Healthy Eating Index-2010 component scoresc

Adequacy

Total fruit 3.0�2.0 3.0�1.9 2.9�2.0 0.99 Whole fruit 3.0�2.2 3.0�2.2 2.6�2.3 0.22 Total vegetables 2.1�1.7 2.0�1.6 2.0�1.6 0.43 Greens and beans 0.5�1.2 0.4�1.0 0.4�1.1 0.15 Whole grains 2.5�2.7 2.5�2.8 2.5�2.7 0.89 Dairy 7.9�2.8x 6.8�3.3y 5.9�3.4z <0.0001* Total protein foods 3.7�1.5 3.7�1.6 3.7�1.5 0.76 Seafood and plant proteins 1.1�1.8x 1.5�2.0y 2.1�2.3z <0.0001* Fatty acids 4.1�3.3x 4.3�3.4xy 4.5�3.5y 0.02* Moderation

Refined grains 4.2�3.5x 4.1�3.5xy 3.7�3.5y 0.03* Sodium 3.9�3.2 4.3�3.2 4.3�3.4 0.08 Empty calories 12.9�5.4x 12.0�5.1y 11.4�5.2yz 0.007* aExcludes students who reported not eating lunch (n¼16). bP values based on analysis of variance using generalized estimated equation models adjusting for sex, race/ethnicity, grade, language spoken at home, and cluster design (type 3 or Wald). Means sharing a common superscript (x, y, z) are not significantly different from each other using a Bonferroni approach at a 5% procedure-wise error rate. Different superscripts indicate statistical differences between groups using a post hoc Bonferroni-Holm multiple comparison test. cMaximum total Healthy Eating Index-2010 total score is 100; component scores can range from 0 to 5 (for total fruit, whole fruit, total vegetables, green and beans, total protein foods, and seafood and plant proteins) or 0 to 10 (for whole grains, dairy, fatty acids, refined grains, and sodium), and 0 to 20 for empty calories. A higher number signifies a healthier intake. *Significant at P<0.05.

RESEARCH

home.7-9,13,14,18,31-34 A study conducted in 12 elementary and intermediate schools in Texas found that when compared with school lunches, those obtained from home contained fewer servings of fruits, vegetables, whole grains, and milk.9

In that study,9 packed lunches also contained more des- serts, chips, and sweetened nondairy drinks, consistent with our findings. To our knowledge, ours is the largest study in terms of

number of students and schools to date to examine the cur- rent diet quality of schoolchildren in relation to their source of breakfast and lunch (whether school meal or from an outside source). Additional study strengths include use of a validated diary-assisted 24-hour recall method to assess di- etary intakes, and examination of diet quality using the most current federal dietary guidance index (ie, HEI-2010) avail- able. The diary-assisted 24-recall method18 has the potential for providing more accurate information by recording foods as they are consumed. Further, obtaining nutrition-related information about school foods and beverages from school foodservice staff enhanced the accuracy of the diary-assisted method. It is noteworthy that this study was conducted before

implementation of the 2010 HHFKA, which mandated

1822 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

improvements to school meals. Beginning in 2012, gradual improvements first to school lunch and then to school breakfast have been mandated nationally to provide students with more fruits, vegetables, and whole grains and less saturated fat, sugar, and salt.5 This is especially important given our study, which took place before HHFKA, showed there were no differences in vegetable or whole grains scores for school breakfast and school lunch, as well as no differ- ences in fruit scores for school lunch, compared with meals obtained from home. It will be important to compare meals from school vs from elsewhere in terms of healthfulness and influence of school meals on overall diet quality after HHFKA standards have been fully implemented. This study had several limitations. It is observational in

nature and was not designed specifically to compare the diet quality of breakfasts and lunches from school with those obtained from home. All data were collected in a single geographic region of Southern California, which limits generalizability of the findings. There were no data available on whether the school meal consumed was part of a reim- bursable meal, or whether it was part of à la carte items, which may be less healthy. Furthermore, given that individ- ual data on the socioeconomic status of each student were

November 2016 Volume 116 Number 11

RESEARCH

not available, it was not possible to adjust for socioeconomic or free or reduced-price lunch status in the analysis. How- ever, because students from lower-income households are more likely than higher-income students to eat school meals (by virtue of qualifying for free or reduced-price school meals) and because lower income is associated with lower dietary quality,35,36 it is possible that differences would have been more pronounced had we been able to adjust for so- cioeconomic status.

CONCLUSIONS Eating school meals, particularly school lunch, may contribute to overall higher diet quality. Findings from this study showed that students who ate breakfast and lunch from school consumed diets higher in dairy-rich foods while limiting calories from solid fats and added sugars, compared with students eating meals from home. However, eating school breakfast alone was not related to overall diet quality and eating school meals was not associated with differences in intakes of vegetables and whole grains compared with eating meals from home. In addition, given that these data were collected before the implementation of the HHFKA regula- tions, future studies in schools using these new guidelines are needed to assess the influence on children’s diet quality.

References 1. US Department of Agriculture. National School Lunch Program:

Participation and lunches served. 2016. http://www.fns.usda.gov/ sites/default/files/pd/slsummar.pdf. Accessed January 20, 2016.

2. US Department of Agriculture. School Breakfast Program participa- tion and meals served. 2016. http://www.fns.usda.gov/sites/default/ files/pd/sbsummar.pdf. Accessed January 20, 2016.

3. Leos-Urbel J, Schwartz AE, Weinstein M, Corcoran S. Not just for poor kids: The impact of universal free school breakfast on meal partici- pation and student outcomes. Econ Educ Rev. 2013;36(0):88-107.

4. Gordon AR, Cohen R, Crepinsek MK, Fox MK, Hall J, Zeidman E. The third School Nutrition Dietary Assessment Study: Background and study design. J Am Diet Assoc. 2009;109(2 suppl):S20-S30.

5. US Department of Agriculture. School meals: Healthy Hunger-Free Kids Act. 2014. http://www.fns.usda.gov/school-meals/healthy- hunger-free-kids-act. Accessed April 2, 2015.

6. US Departments of Agriculture and Health and Human Services. Dietary Guidelines for Americans, 2010. 7th ed. 2010. http://www. cnpp.usda.gov/dietaryguidelines.htm. Accessed December 19, 2014.

7. Johnston CA, Moreno JP, El-Mubasher A, Woehler D. School lunches and lunches brought from home: A comparative analysis. Child Obes. 2012;8(4):364-368.

8. Hur I, Burgess-Champoux T, Reicks M. Higher quality intake from school lunch meals compared with bagged lunches. ICAN. 2011;3(2): 70-75.

9. Caruso ML, Cullen KW. Quality and cost of student lunches brought from home. JAMA Pediatr. 2015;169(1):86-90.

10. Farris AR, Misyak S, Duffey KJ, et al. Nutritional comparison of packed and school lunches in pre-kindergarten and kindergarten children following the implementation of the 2012-2013 National School Lunch Program standards. J Nutr Educ Behav. 2014;46(6):621-626.

11. Briefel RR, Wilson A, Gleason PM. Consumption of low-nutrient, energy-dense foods and beverages at school, home, and other lo- cations among school lunch participants and nonparticipants. J Am Diet Assoc. 2009;109(2 suppl):S79-S90.

12. Cole N, Fox M. Diet quality of American school-age children by school lunch participation status: Data from the National Health and Nutri- tion Examination Survey, 1999-2004. 2008. http://www.fns.usda.gov/ sites/default/files/NHANES-NSLP.pdf. Accessed March 21, 2016.

13. Condon E, Drilea S, Lichtenstein C, Mabli J, Madden E, Niland K. Diet quality of American school children by National School Lunch participation status: Data from the National Health and Nutrition Examination Survey, 2005-2010. 2015. http://www.fns.usda.gov/diet-

November 2016 Volume 116 Number 11 JO

quality-american-school-children-national-school-lunch-program- participation-status-data. Accessed March 21, 2016.

14. Campbell BL, Nayga RM, Park JL, Silva A. Does the National School Lunch Program improve children’s dietary outcomes? Am J Agric Econ. 2011;93(4):1099-1130.

15. Bhattacharya J, Currie J, Haider SJ. Breakfast of champions? The School Breakfast Program and the nutrition of children and families. J Hum Resource. 2006;41(3):445-466.

16. Guenther PM, Casavale KO, Reedy J, et al. Update of the Healthy Eating Index: HEI-2010. J Acad Nutr Diet. 2013;113(4):569-580.

17. Keihner A, Rosen N, Wakimoto P, et al. Impact of California Chil- dren’s Power Play! campaign on fruit and vegetable intake and physical activity among fourth- and fifth-grade students [published online ahead of print November 15, 2015]. Am J Health Promot. http://dx.doi.org/10.4278/ajhp.141125-ARB-592.

18. McPherson RS, Hoelscher DM, Alexander M, Scanlon KS, Serdula MK. Dietary assessment methods among school-aged children: Validity and reliability. Prev Med. 2000;31(suppl 2):S11-S33.

19. Crawford PB, Gosliner W, Kayman H. The ethical basis for promoting nutritional health in public schools in the United States. Prev Chronic Dis. 2011;8(5):A98.

20. Lytle LA, Murray DM, Perry CL, Eldridge AL. Validating fourth-grade students’ self-report of dietary intake. J Am Diet Assoc. 1998;98(5): 570-572.

21. Lytle LA, Nichaman MZ, Obarzanek E, et al. Validation of 24-hour recalls assisted by food records in third-grade children. The CATCH Collaborative Group. J Am Diet Assoc. 1993;93(12):1431-1436.

22. Weber JL, Lytle L, Gittelsohn J, et al. Validity of self-reported dietary intake at school meals by American Indian children: The pathways study. J Am Diet Assoc. 2004;104(5):746-752.

23. Conway JM, Ingwersen LA, Moshfegh AJ. Accuracy of dietary recall using the USDA five-step multiple-pass method in men: An obser- vational validation study. J Am Diet Assoc. 2004;104(4):595-603.

24. Bartlett S, Olsho L, Klerman J, Patland K, Blocklin M, Connor P. Evaluation of the fresh fruit and vegetable program (FFVP): Final evaluation report. 2013. http://www.fns.usda.gov/evaluation-fresh- fruit-and-vegetable-program. Accessed March 21, 2016.

25. Guenther PM, Kirkpatrick SI, Reedy J, et al. The Healthy Eating Index- 2010 is a valid and reliable measure of diet quality according to the 2010 Dietary Guidelines for Americans. J Nutr. 2014;144(3):399-407.

26. Fenton K, Rosen NJ, Wakimoto P, Patterson T, Goldstein LH, Ritchie LD. Eat lunch first or play first? Inconsistent associations with fruit and vegetable consumption in elementary school. J Acad Nutr Diet. 2015;115(4):585-592.

27. Ritchie LD, Rosen NJ, Fenton K, Au LE, Goldstein LH, Shimada T. School breakfast policy is associated with dietary intake of fourth- and fifth-grade students. J Acad Nutr Diet. 2016;116(3):449-457.

28. SAS Institute, Inc. Released 2013. SAS version 9.4. Cary, NC: SAS Institute Inc.

29. Hanson KL, Olson CM. School meals participation and weekday dietary quality were associated after controlling for weekend eating among U. S. school children aged 6 to 17 years. J Nutr. 2013;143(5):714-721.

30. Bhattacharya J, Currie J, Haider S. Poverty, food insecurity, and nutritional outcomes in children and adults. J Health Econ. 2004;23(4):839-862.

31. Longacre MR, Drake KM, Titus LJ, et al. School food reduces house- hold income disparities in adolescents’ frequency of fruit and vegetable intake. Prev Med. 2014;69:202-207.

32. Hubbard KL, Must A, Eliasziw M, Folta SC, Goldberg J. What’s in children’s backpacks: Foods brought from home. J Acad Nutr Diet. 2014;114(9):1424-1431.

33. Rees GA, Richards CJ, Gregory J. Food and nutrient intakes of primary school children: A comparison of school meals and packed lunches. J Hum Nutr Diet. 2008;21(5):420-427.

34. Gleason PM, Suitor CW. Eating at school: How the National School Lunch Program affects children’s diets. Am J Agr Econ. 2003;85(4): 1047-1061.

35. Leung CW, Blumenthal SJ, Hoffnagle EE, et al. Associations of food stamp participation with dietary quality and obesity in children. Pediatrics. 2013;131(3):463-472.

36. Leung CW, Epel ES, Ritchie LD, Crawford PB, Laraia BA. Food inse- curity is inversely associated with diet quality of lower-income adults. J Acad Nutr Diet. 2014;114(12):1943-1953.e1942.

URNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1823

RESEARCH

AUTHOR INFORMATION L. E. Au is an assistant researcher, K. Hecht is director of policy, and L. D. Ritchie is director and a cooperative extension specialist, Nutrition Policy Institute, Division of Agriculture and Natural Resources, University of California, Berkeley. N. J. Rosen is a senior associate, Informing Change, Berkeley, CA. K. Fenton is a biostatistician, Seattle Genetics, Bothell, WA.

Address correspondence to: Lauren E. Au, PhD, RD, Nutrition Policy Institute, Division of Agriculture and Natural Resources, University of California, 2115 Milvia St, Suite 301, Berkeley, CA 94704. E-mail: [email protected]

STATEMENT OF POTENTIAL CONFLICT OF INTEREST No potential conflict of interest was reported by the authors.

FUNDING/SUPPORT The original data collection for the study was funded by the Public Health Institute (grant no. 1017431). The Public Health Institute had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

ACKNOWLEDGEMENTS The authors thank the schools, school staff, and students who participated in the study.

1824 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS November 2016 Volume 116 Number 11