Effect of a Breakfast in the Classroom Initiative on Obesity in Urban School-aged Children: A Cluster Randomized Clinical Trial | Nutrition | JAMA Pediatrics | JAMA Network
[Skip to Navigation]
Figure.  Consort Diagram of Student-Level Participation Within 16 Participating Schools
Consort Diagram of Student-Level Participation Within 16 Participating Schools

aStudents who transferred between the study midpoint and the end point (n = 1 in intervention schools and n = 3 in control schools), refused participation, or were chronically absent between baseline and midpoint.

bStudents did not attend a study school during the midpoint data collection but returned to a study school by the study end point.

Table 1.  Baseline Characteristics of 1362 Fourth- Through Sixth-Grade Philadelphia Public School Students Participating in the One Healthy Breakfast Trial
Baseline Characteristics of 1362 Fourth- Through Sixth-Grade Philadelphia Public School Students Participating in the One Healthy Breakfast Trial
Table 2.  Weight Status and BMI Outcomes at Midpoint and End Point Among Participants in the One Healthy Breakfast Triala
Weight Status and BMI Outcomes at Midpoint and End Point Among Participants in the One Healthy Breakfast Triala
1.
Frisvold  DE.  Nutrition and cognitive achievement: an evaluation of the School Breakfast Program.  J Public Econ. 2015;124:91-104. doi:10.1016/j.jpubeco.2014.12.003PubMedGoogle ScholarCrossref
2.
Bartfeld  JS, Ahn  HM.  The School Breakfast Program strengthens household food security among low-income households with elementary school children.  J Nutr. 2011;141(3):470-475. doi:10.3945/jn.110.130823PubMedGoogle ScholarCrossref
3.
School Breakfast Scorecard.  School Year 2016-2017. Washington, DC: Food Research & Action Center; 2018.
4.
Bernstein  LS, McLaughlin  JE, Crepinsek  MK, Daft  LM.  Evaluation of the School Breakfast Program Pilot Project: Final Report. Washington, DC: US Dept of Agriculture Food & Nutrition Service; 2004.
5.
Schwaab  B, Fettweis  C, Caplon  M.  Increasing student breakfast consumption: an established model in Maryland: interview by Jodi R. Godfrey.  Child Obes. 2012;8(2):162-166. doi:10.1089/chi.2012.0082.godfPubMedGoogle ScholarCrossref
6.
Butcher-Powell  LM, Bordi  PL, Borja  M, Cranage  D, Cole  C.  Factors affecting breakfast intake in children.  Topics Clin Nutr. 2003;18(2):130-135. doi:10.1097/00008486-200304000-00009Google ScholarCrossref
7.
Healthy, Hunger-Free Kids Act of 2010, S 3307, 111th Congress (2010). https://www.congress.gov/bill/111th-congress/senate-bill/3307. Accessed May 15, 2018.
8.
 School Breakfast Program Resolution. Washington, DC: US Dept of Agriculture; 2009.
9.
Lawman  HG, Polonsky  HM, Vander Veur  SS,  et al.  Breakfast patterns among low-income, ethnically-diverse 4th-6th grade children in an urban area.  BMC Public Health. 2014;14:604. doi:10.1186/1471-2458-14-604PubMedGoogle ScholarCrossref
10.
Niemeier  HM, Raynor  HA, Lloyd-Richardson  EE, Rogers  ML, Wing  RR.  Fast food consumption and breakfast skipping: predictors of weight gain from adolescence to adulthood in a nationally representative sample.  J Adolesc Health. 2006;39(6):842-849. doi:10.1016/j.jadohealth.2006.07.001PubMedGoogle ScholarCrossref
11.
Albertson  AM, Franko  DL, Thompson  D,  et al.  Longitudinal patterns of breakfast eating in black and white adolescent girls.  Obesity (Silver Spring). 2007;15(9):2282-2292. doi:10.1038/oby.2007.271PubMedGoogle ScholarCrossref
12.
Berkey  CS, Rockett  HR, Gillman  MW, Field  AE, Colditz  GA.  Longitudinal study of skipping breakfast and weight change in adolescents.  Int J Obes Relat Metab Disord. 2003;27(10):1258-1266. doi:10.1038/sj.ijo.0802402PubMedGoogle ScholarCrossref
13.
Szajewska  H, Ruszczynski  M.  Systematic review demonstrating that breakfast consumption influences body weight outcomes in children and adolescents in Europe.  Crit Rev Food Sci Nutr. 2010;50(2):113-119. doi:10.1080/10408390903467514PubMedGoogle ScholarCrossref
14.
Deshmukh-Taskar  PR, Nicklas  TA, O’Neil  CE, Keast  DR, Radcliffe  JD, Cho  S.  The relationship of breakfast skipping and type of breakfast consumption with nutrient intake and weight status in children and adolescents: the National Health and Nutrition Examination Survey 1999-2006.  J Am Diet Assoc. 2010;110(6):869-878. doi:10.1016/j.jada.2010.03.023PubMedGoogle ScholarCrossref
15.
Tin  SP, Ho  SY, Mak  KH, Wan  KL, Lam  TH.  Breakfast skipping and change in body mass index in young children.  Int J Obes (Lond). 2011;35(7):899-906. doi:10.1038/ijo.2011.58PubMedGoogle ScholarCrossref
16.
Utter  J, Scragg  R, Mhurchu  CN, Schaaf  D.  At-home breakfast consumption among New Zealand children: associations with body mass index and related nutrition behaviors.  J Am Diet Assoc. 2007;107(4):570-576. doi:10.1016/j.jada.2007.01.010PubMedGoogle ScholarCrossref
17.
So  HK, Nelson  EA, Li  AM,  et al.  Breakfast frequency inversely associated with BMI and body fatness in Hong Kong Chinese children aged 9-18 years.  Br J Nutr. 2011;106(5):742-751. doi:10.1017/S0007114511000754PubMedGoogle ScholarCrossref
18.
Sandercock  GR, Voss  C, Dye  L.  Associations between habitual school-day breakfast consumption, body mass index, physical activity and cardiorespiratory fitness in English schoolchildren.  Eur J Clin Nutr. 2010;64(10):1086-1092. doi:10.1038/ejcn.2010.145PubMedGoogle ScholarCrossref
19.
Borradaile  KE, Sherman  S, Vander Veur  SS,  et al.  Snacking in children: the role of urban corner stores.  Pediatrics. 2009;124(5):1293-1298. doi:10.1542/peds.2009-0964PubMedGoogle ScholarCrossref
20.
Grynbaum  MM. Hiccup in most important meal. New York Times. April 20, 2012; A1.
21.
Van Wye  G, Seoh  H, Adjoian  T, Dowell  D.  Evaluation of the New York City breakfast in the classroom program.  Am J Public Health. 2013;103(10):e59-e64. doi:10.2105/AJPH.2013.301470PubMedGoogle ScholarCrossref
22.
Corcoran  SP, Elbel  B, Schwartz  AE.  The effect of breakfast in the classroom on obesity and academic performance: evidence from New York City.  J Policy Anal Manage. 2016;35(3):509-532. doi:10.1002/pam.21909PubMedGoogle ScholarCrossref
23.
Gleason  PM, Dodd  AH.  School breakfast program but not school lunch program participation is associated with lower body mass index.  J Am Diet Assoc. 2009;109(2)(suppl):S118-S128. doi:10.1016/j.jada.2008.10.058PubMedGoogle ScholarCrossref
24.
Kuczmarski  RJ, Ogden  CL, Guo  SS,  et al.  2000 CDC Growth Charts for the United States: methods and development.  Vital Health Stat 11. 2002;(246):1-190.PubMedGoogle Scholar
25.
Flegal  KM, Shepherd  JA, Looker  AC,  et al.  Comparisons of percentage body fat, body mass index, waist circumference, and waist-stature ratio in adults.  Am J Clin Nutr. 2009;89(2):500-508. doi:10.3945/ajcn.2008.26847PubMedGoogle ScholarCrossref
26.
Lawman  HG, Mallya  G, Veur  SV,  et al.  Trends in relative weight over 1 year in low-income urban youth.  Obesity (Silver Spring). 2015;23(2):436-442. doi:10.1002/oby.20928PubMedGoogle ScholarCrossref
27.
Lawman  HG, Ogden  CL, Hassink  S, Mallya  G, Vander Veur  S, Foster  GD.  Comparing methods for identifying biologically implausible values in height, weight, and body mass index among youth.  Am J Epidemiol. 2015;182(4):359-365. doi:10.1093/aje/kwv057PubMedGoogle ScholarCrossref
28.
Foster  GD, Sherman  S, Borradaile  KE,  et al.  A policy-based school intervention to prevent overweight and obesity.  Pediatrics. 2008;121(4):e794-e802. doi:10.1542/peds.2007-1365PubMedGoogle ScholarCrossref
29.
Preisser  JS, Lohman  KK, Rathouz  PJ.  Performance of weighted estimating equations for longitudinal binary data with drop-outs missing at random.  Stat Med. 2002;21(20):3035-3054. doi:10.1002/sim.1241PubMedGoogle ScholarCrossref
30.
Jakobsen  JC, Gluud  C, Wetterslev  J, Winkel  P.  When and how should multiple imputation be used for handling missing data in randomised clinical trials—a practical guide with flowcharts.  BMC Med Res Methodol. 2017;17(1):162. doi:10.1186/s12874-017-0442-1PubMedGoogle ScholarCrossref
31.
Anzman-Frasca  S, Djang  HC, Halmo  MM, Dolan  PR, Economos  CD.  Estimating impacts of a breakfast in the classroom program on school outcomes.  JAMA Pediatr. 2015;169(1):71-77. doi:10.1001/jamapediatrics.2014.2042PubMedGoogle ScholarCrossref
32.
Antonogeorgos  G, Panagiotakos  DB, Papadimitriou  A, Priftis  KN, Anthracopoulos  M, Nicolaidou  P.  Breakfast consumption and meal frequency interaction with childhood obesity.  Pediatr Obes. 2012;7(1):65-72. doi:10.1111/j.2047-6310.2011.00006.xPubMedGoogle ScholarCrossref
33.
Kral  TV, Allison  DB, Birch  LL, Stallings  VA, Moore  RH, Faith  MS.  Caloric compensation and eating in the absence of hunger in 5- to 12-y-old weight-discordant siblings.  Am J Clin Nutr. 2012;96(3):574-583. doi:10.3945/ajcn.112.037952PubMedGoogle ScholarCrossref
34.
Carnell  S, Benson  L, Gibson  EL, Mais  LA, Warkentin  S.  Caloric compensation in preschool children: relationships with body mass and differences by food category.  Appetite. 2017;116:82-89. doi:10.1016/j.appet.2017.04.018PubMedGoogle ScholarCrossref
35.
Sherman  S, Grode  G, McCoy  T,  et al.  Corner stores: the perspective of urban youth.  J Acad Nutr Diet. 2015;115(2):242-248. doi:10.1016/j.jand.2014.10.017PubMedGoogle ScholarCrossref
36.
Bauer  KW, Chuisano  S. Intentional Self-Regulation of Eating Among Children and Adolescents. In: Lumeng JC, Fisher JO, eds.  Pediatric Food Preferences and Eating Behaviors. Philadelphia, PA: Elsevier; 2018:255-270. doi:10.1016/B978-0-12-811716-3.00013-0
37.
Foster  GD, Linder  B, Baranowski  T,  et al; HEALTHY Study Group.  A school-based intervention for diabetes risk reduction.  N Engl J Med. 2010;363(5):443-453. doi:10.1056/NEJMoa1001933PubMedGoogle ScholarCrossref
Original Investigation
February 25, 2019

Effect of a Breakfast in the Classroom Initiative on Obesity in Urban School-aged Children: A Cluster Randomized Clinical Trial

Author Affiliations
  • 1Center for Obesity Research and Education, College of Public Health, Temple University, Philadelphia, Pennsylvania
  • 2Department of Nutritional Sciences, University of Michigan School of Public Health, Ann Arbor
  • 3Center for Obesity Research and Education, College of Public Health, Temple University, Philadelphia, Pennsylvania
  • 4Department of Social & Behavioral Sciences, College of Public Health, Temple University, Philadelphia, Pennsylvania
  • 5College of Health Sciences and Department of Behavioral Health and Nutrition, University of Delaware, Newark
  • 6The Food Trust, Philadelphia, Pennsylvania
  • 7Department of Statistics, Virginia Polytechnic Institute and State University, Roanoake
  • 8Fox Chase Cancer Center, Temple Health, Philadelphia, Pennsylvania
  • 9WW (formerly Weight Watchers), New York, New York
  • 10Center for Weight and Eating Disorders, University of Pennsylvania, Philadelphia
JAMA Pediatr. 2019;173(4):326-333. doi:10.1001/jamapediatrics.2018.5531
Key Points

Question  What effect does a breakfast in the classroom initiative have on the weight status of low-income, urban students?

Findings  This cluster randomized clinical trial involving 1362 grade school students at 16 Philadelphia public schools found that a breakfast in the classroom initiative did not affect the combined incidence of overweight and obesity compared with offering students breakfast before school in the cafeteria. However, the incidence and prevalence of obesity were significantly higher among students offered the breakfast in the classroom initiative after 2.5 years.

Meaning  Although it increased participation in the federal School Breakfast Program, implementing breakfast in the classroom with complementary breakfast-specific nutrition education may have unfavorable consequences on childhood obesity.

Abstract

Importance  Serving breakfast in the classroom is promoted to increase participation in the federal School Breakfast Program. However, little is known about the effect of breakfast in the classroom on children’s weight status.

Objective  To evaluate the effect of a breakfast in the classroom initiative, which combined breakfast in the classroom with breakfast-specific nutrition education, on overweight and obesity among urban children in low-income communities.

Design, Setting, and Participants  A cluster-randomized clinical trial among 1362 fourth- through sixth-grade students from low-income urban communities across 2.5 years. Sixteen kindergarten through eighth grade Philadelphia public schools with universal breakfast participated. Participants were recruited in September 2013, and the intervention began in January 2014. Data analysis took place from April 1, 2018, to March 31, 2019.

Interventions  Intervention schools received a program that included breakfast in the classroom and breakfast-specific nutrition education. Control schools continued breakfast before school in the cafeteria and standard nutrition education.

Main Outcomes and Measures  The primary outcome was the combined incidence of overweight and obesity. Secondary outcomes included the combined prevalence of overweight and obesity, incidence and prevalence of obesity, changes in body mass index (BMI) z score, and School Breakfast Program participation.

Results  Among the 1362 students, mean (SD) age was 10.8 (0.96) years and 700 (51.4%) were female; 907 (66.6%) were black, 233 (17.1%) were Hispanic, 100 (7.3%) were white, 83 (6.1%) were Asian, and 39 were of multiple or other race/ethnicity. After 2.5 years, students in intervention schools had participated in the School Breakfast Program 53.8% of days, compared with 24.9% of days among students in control schools (β = 0.33; 95% CI, 0.24-0.42). There was no difference between intervention and control schools in the combined incidence of overweight and obesity after 2.5 years (11.7% vs 9.1%; odds ratio [OR] 1.42; 95% CI, 0.82-2.44; P = .21). However, the incidence (11.6% vs 4.4%; OR, 3.27; 95% CI, 1.87-5.73) and prevalence (28.0% vs 21.2%; OR, 1.43; 95% CI, 1.08-1.89) of obesity were higher in intervention schools than in control schools after 2.5 years.

Conclusions and Relevance  A breakfast in the classroom initiative increased participation in the School Breakfast Program and did not affect the combined incidence of overweight and obesity. However, the initiative had an unintended consequence of increasing incident and prevalent obesity. Further research is needed to identify approaches to increase participation in the School Breakfast Program that do not increase obesity among students.

Trial Registration  ClinicalTrials.gov identifier: NCT01924130

Introduction

The federal School Breakfast Program provides children from low-income households access to a nutritious breakfast at school free or at a reduced cost and has been associated with greater student academic achievement1 and may protect families from food insecurity.2 Although on a typical day, 12.2 million children in low-income households participate in the School Breakfast Program, the program is underused, with participation half that of the National School Lunch Program.3

To increase School Breakfast Program participation, the federal government encourages alternative methods of implementing the program, including offering all students breakfast in their classroom at the start of the school day rather than before school in the cafeteria. School districts that implement breakfast in the classroom experience 2-fold to 4-fold increases in School Breakfast Program participation.4-6 School districts are increasingly offered incentives or mandated to implement breakfast in the classroom,3,7,8 and several of the country’s largest urban school districts, including New York, Philadelphia, Pennsylvania, and Detroit, Michigan, have adopted breakfast in the classroom.

Despite the popularity of breakfast in the classroom, relatively little is known about the effects of breakfast in the classroom on obesity, one of the most pressing health concerns among children. Increasing school breakfast consumption may prevent obesity given the observed associations between regular breakfast consumption and healthier weight status among children.9-18 Alternatively, offering children breakfast in the classroom when some have already eaten at home9 or on the way to school9,19 may promote excess energy intake.20

A small number of studies, all nonrandomized, have examined the potential effect of breakfast in the classroom on childhood obesity. Among New York City students, those provided breakfast in the classroom consumed more energy in the morning and the number of locations where children ate in the morning was greater than among students not offered breakfast in the classroom.21 However, also in New York City, breakfast in the classroom implementation was not associated with changes in obesity prevalence.22

The present study, using a cluster randomized trial design, evaluated the effect of an initiative, One Healthy Breakfast, which combined breakfast in the classroom with breakfast-specific nutrition education, on overweight and obesity among urban low-income children. It was hypothesized that students in schools randomized to implement One Healthy Breakfast would experience a lower incidence of overweight and obesity across 2.5 years compared with students in control schools, where the School Breakfast Program was offered in the cafeteria before the beginning of the school day. This hypothesis was based on the belief that the breakfast in the classroom initiative would increase students’ regular consumption of breakfast at school, which has been associated with healthier weight status.23

Methods
Design and Sample

This study was a 2.5-year cluster randomized clinical trial in 16 kindergarten through eighth grade (K-8) public schools in Philadelphia. Participants were fourth- through sixth-grade students at baseline. Most schools in the School District of Philadelphia, and all the schools in this study, provide universal breakfast, offering breakfast to all students free of charge. The 16 schools were evenly randomized to 1 of 2 conditions: (1) an intervention that included breakfast in the classroom, described below, or (2) breakfast offered in the cafeteria before the beginning of the school day. The primary outcome was the combined incidence of overweight and obesity (ie, the proportion of new cases of overweight and obesity among students without overweight or obesity at baseline). Secondary outcomes included the combined prevalence of overweight and obesity (ie, the proportion of students with overweight or obesity among all students), the incidence and prevalence of obesity, changes in students’ body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) z scores, and School Breakfast Program participation. The study was approved by and conducted in accordance with the ethical standards of the Office of Research and Evaluation at the School District of Philadelphia and the Institutional Review Board at Temple University, Philadelphia. The full trial protocol is given in Supplement 1. Recruitment of students, parental written informed consent, and student assent occurred in September 2013. Neither parents nor students received compensation for their participation.

Schools were eligible if they met the following inclusion criteria: (1) at least 50% of students qualified for free or reduced-priced meals through the National School Lunch Program; (2) no existing breakfast in the classroom program or the willingness to not offer breakfast in the classroom for the duration of the study if randomized to the control condition; and (3) received nutrition education programming as part of the US Department of Agriculture’s (USDA’s) Supplemental Nutrition Assistance Program (SNAP-Ed). The last requirement was necessary because the education component of the breakfast in the classroom initiative complemented existing SNAP-Ed programming.

Among 74 eligible schools, 25 were chosen at random and invited to participate; 9 of the invited schools were unwilling to accept randomization. The remaining 16 schools were similar to other schools in the School District of Philadelphia with respect to student race/ethnicity, student enrollment, and percentage of students qualifying for free or reduced-price school meals. Among the 16 schools, the mean percentage of students eligible for free or reduced-price meals was 89.4% (range, 73.7%-96.2%), and 5324 (65.3%) were black. Before randomization, schools were matched into 8 pairs based on student enrollment, racial/ethnic composition of the student population, and food service type (full-service cafeteria vs satellite cafeteria). Schools were randomized within pairs by using a random number generator in the spring of 2013, and school administrators were notified of their schools’ condition at that time. The enrolled schools received $1000 at the beginning of each school year for participating.

Baseline data were obtained after randomization by trained research staff. Intervention schools began the intervention in January 2, 2014, and it continued for 2.5 school years. Midpoint data were collected after 1.5 years and end point data after 2.5 years. Data were analyzed between April 1, 2018, and March 31, 2019. The CONSORT diagram (Figure) shows the study’s flow at the student level. Parental consent and student assent were obtained from 1463 of 2715 eligible students (53.9%), and 101 students (6.9%) were removed from the analytic sample, yielding a total baseline sample of 1362 students.

There was no attrition at the school level during the study. By the end of the study, 793 students (58.2%) remained in the study; 99% of student-level attrition was owing to student transfer. The eTable in Supplement 2 contains the sociodemographic characteristics of students who were retained at the end of the study vs those lost to follow-up.

Intervention Schools

All students in the target grades (fourth through sixth grades in 2013-2014, fifth through seventh grades in 2014-2015, and sixth through eighth grades in 2015-2016) were offered free breakfast in their classrooms at the start of the school day. The following paragraphs describe the additional components of One Healthy Breakfast that were developed and delivered by The Food Trust, a community-based, nonprofit organization.

Nutrition Education

Students received 18 forty-five–minute nutrition education lessons regarding the importance of breakfast. These lessons were in addition to the students’ regular nutrition education programming.

Social Marketing

Students and staff received a variety of items with the One Healthy Breakfast logo. Teachers and staff received posters for their classrooms and the cafeteria. All students attended an annual kickoff educational assembly.

Corner Store Marketing

Shelf talkers encouraging shoppers to make healthy choices were placed at 14 corner stores located less than 0.5 mile from the intervention schools (approximately 1-2 stores per school). Shelf talkers each included a targeted message and the One Healthy Breakfast mascot and logo.

Parent Outreach

Monthly newsletters (26 across 2.5 years) containing articles about a healthy breakfast, family activities, and the monthly school breakfast menu were sent home. Study staff set up booths at back-to-school night events and report card conferences.

Control Schools

Control schools served breakfast free of charge in the cafeteria before school, and existing SNAP-Ed nutrition education continued in control schools. No purposeful obesity prevention interventions occurred in the control schools during the study.

At baseline, all schools enrolled in the trial offered the same breakfast menus. However, over the course of the trial, the foods offered to students in intervention vs control schools may have differed slightly because intervention schools chose not to offer hot breakfast items (eg, egg sandwiches, pancakes) and offered only cold items (eg, cereal, muffins). Students in control schools were offered both hot and cold items. All breakfasts offered met USDA regulations regarding meal components (ie, fruit or 100% fruit juice, milk, and grain were offered) and across the week, both hot and cold menus averaged between 450 and 500 kcal per meal.

Measures

The following measures were assessed among students at baseline, 1.5 years, and 2.5 years by trained research assistants. Research assistants were not blinded to schools’ intervention conditions because social marketing materials were visible in intervention classrooms.

Height and Weight

Height and weight were measured using standard protocols. Two measurements of height (in centimeters) and weight (in kilograms) were obtained at each time point with portable stadiometers (SECA 217; Seca GmbH) and scales (SECA 869; Seca GmbH). If the 2 measures for height and weight differed by more than 1.0 cm and 0.2 kg, respectively, a third measure was taken and the mean of the 2 within the specified range was calculated. Inflexible hairstyles (eg, braids) were measured and subtracted from the height measurement. Age- and sex-specific BMI percentiles and z scores were calculated based on Centers for Disease Control and Prevention (CDC) 2000 reference data.24 Weight status was defined based on sex-specific BMI-for-age percentiles as follows: underweight (<5th percentile), healthy weight (≥5th to <85th percentile), overweight (≥85th to <95th percentile), and obesity (≥95th percentile).24,25 Students were excluded if their height measurements, weight measurements, or both were biologically implausible (9 students were excluded at baseline).26,27

School Breakfast Program Participation

Students’ School Breakfast Program participation was recorded by teachers for students in intervention schools and by cafeteria staff for students in control schools. Dates that the students attended school and dates that they participated in the School Breakfast Program were extracted from the school district’s database and provided to study researchers. The proportion of days in attendance that students participated in the School Breakfast Program in the fall 2013 semester (baseline), between the spring 2014 and spring 2015 semesters (midpoint), and between the fall 2015 and spring 2016 semesters (end point) was calculated.

Student-Level Demographic Information

Children’s race/ethnicity, sex, birthdate, free or reduced-price meal status, and grade level were obtained from the school district at baseline. Children’s race/ethnicity, based on parental report, was categorized as black, Hispanic, white, Asian, or other/mixed race.

Statistical Analysis

Statistical power was determined for the primary study outcome at the study end point based on preliminary data to estimate incidence, recruitment and retention rates, and design effects. The primary outcome was the combined incidence of overweight and obesity. Previous research suggested a greater than 30% reduction in incidence of overweight and obesity for the intervention group from a baseline rate of approximately 30%.28 Thus, to detect a risk ratio of 0.70 with a baseline incidence of 30%, assuming α = .05, statistical power of 80% or greater, and intraclass correlation coefficient of 0.01 or less, a minimum sample size of 15 schools with approximately 110 students enrolled per school (N = 1760) was needed. To have equal numbers of schools per condition and to protect against the loss of a school, 16 schools were recruited.

Descriptive statistics were used to characterize student demographic characteristics, student weight status, and School Breakfast Program participation rates at baseline, overall, and by intervention condition. Weighted generalized estimating equation (GEE) models with the logit-link function were used to evaluate intervention differences for dichotomous incidence and prevalence outcomes at each time point. Use of weighted estimating equations for longitudinal binary outcomes with dropouts missing at random is recommended by Preisser et al.29 The assumption that dropouts were missing at random and not completely at random is conservative and guided by the potential for differential loss to follow-up. A similar weighted GEE model approach with the gaussian distribution as the working outcome distribution was used to examine intervention differences for BMI z scores and School Breakfast Program participation rates at each time point. All weighted GEE analyses were adjusted for paired stratification of randomization. Students with prevalent overweight or obesity or obesity were excluded from the models estimating incidence of each outcome at each time point. Given the student-level missing data at the study end point (42%), it was concluded that imputing missing outcome values was not an appropriate strategy.30 Therefore, the results presented are based on students retained in the study at either 1.5 or 2.5 years. Statistical significance was taken at the 2-sided P = .05 level and did not account for multiplicity given that the combined incidence of overweight and obesity at the study end point was proposed as the primary outcome a priori. Analyses were conducted using SAS version 9.4 (SAS Institute Inc), and weighted GEE modeling was performed using the geeglm function in R (The R Project for Statistical Computing).

Results
School Breakfast Program Participation

Among the baseline sample of 1362 students, the mean (SD) age was 10.8 (1.0) years, 700 (51.4%) were girls, 907 (66.6%) were black, 233 (17.1%) were Hispanic, 100 (7.3%) were white, 83 (6.1%) were Asian, and 39 were of multiple or other race/ethnicity (Table 1). Among the participating students, 1075 (78.9%) qualified for free or reduced-price meals. At baseline, 241 (17.7%) met criteria for overweight and 291 (21.4%) for obesity.

At baseline, there was no difference between intervention and control schools in students’ participation in the School Breakfast Program (32.1% of days vs 29.7% of days, respectively). Participation increased to 72.0% of days in intervention schools after 1.5 years, whereas in control schools students participated on average 25.9% of days (β = 0.46; 95% CI, 0.38-0.53). By the study end point, differences in participation rates in intervention vs control schools (53.8% of days vs 24.9% of days) remained statistically significant (β = 0.33; 95% CI, 0.24-0.42).

Weight Status

There was no significant difference in the primary outcome (combined incidence of overweight and obesity) between intervention schools (11.7% [24 of 205 students]) and control schools (9.1% [24 of 263]) after 2.5 years (odds ratio [OR], 1.42; 95% CI, 0.82-2.44; P = .21) (Table 2). The incidence of obesity alone, however, was higher in intervention schools (11.6% [31 of 268 students]) than in control schools (4.4% [15 of 342]) between baseline and the study end point (OR, 3.27; 95% CI, 1.87-5.73). The prevalence of obesity at the end point was also higher in intervention schools (28.0% [98 of 350 students]) than in control schools (21.2% [94 of 443]) at end point (OR, 1.43; 95% CI, 1.08-1.89). There was no difference between intervention and control schools in the combined prevalence of overweight and obesity or BMI z score across the study period. The midpoint assessments revealed that the incidence of obesity was higher in intervention schools (7.1% [24 of 336 students]) than in control schools (4.3% [18 of 422]) after 1.5 years.

Discussion

There are 3 principal findings from this, to our knowledge, first randomized clinical trial of a breakfast in the classroom initiative. First, offering breakfast in the classroom in coordination with breakfast-specific nutrition education is an effective strategy to increase School Breakfast Program participation. These findings are consistent with previous studies.22,31

Second, breakfast in the classroom plus breakfast-specific nutrition education had no effect on the primary outcome—the combined incidence of overweight and obesity. These findings are contrary to our hypothesis and evidence from previous studies suggesting that regular breakfast consumption is associated with lower relative weight.12-14,17,32 However, these findings align the previous observational study of breakfast in the classroom in New York City that found no association between program implementation and prevalent obesity.22 These findings suggest that providing a breakfast in the classroom initiative does not prevent overweight and obesity.

Third, the breakfast in the classroom initiative conferred unintended and untoward consequences on children’s weight status. The effects of the intervention on the incidence of obesity were pronounced: an approximately 3-fold increase. Although the precise reasons for this increase in obesity are unknown, in New York City breakfast in the classroom was associated with students consuming 95 more kilocalories per morning.21 Previous studies have shown that students in Philadelphia have numerous opportunities to eat in the morning before school,9,19 which may make breakfast in the classroom an additional breakfast, rather than students’ only breakfast. Because One Healthy Breakfast included more than just breakfast in the classroom, other intervention components may have contributed to the observed increases in obesity prevalence and incidence among students. However, given the increases in breakfast participation resulting from the intervention, it is unlikely that the increases in obesity are not at least in part a result of breakfast in the classroom as a meal delivery method.

The finding of increased incidence and prevalence of obesity, but not overweight and obesity combined, suggests that children in the overweight category are at particularly high risk of developing obesity when offered a breakfast in the classroom initiative. Children with higher BMIs are less able to compensate for energy consumed and are more likely to consume greater amounts and eat in the absence of hunger when presented food.33,34 Therefore, when offered breakfast in the classroom, children with overweight may have particular difficulty self-regulating intake.

Limitations

Study limitations include unknown generalizability to nonurban locations or populations of differing sociodemographic characteristics. The intervention effect may be unique to older elementary school–aged and middle school–aged students who report increasing freedom in their morning food purchases.35 It is unknown whether the intervention would have had a similar effect on younger children, who may not have as much access to convenience foods in the morning, or older children, who may be better able to intentionally limit their morning intake to the breakfast meal provided at school.36 Student-level attrition in the study was high although similar to previous school-based trials in low-income school districts,28,37 and virtually all (99%) of attrition was a result of school transfer and not differential by student-level characteristics, minimizing concerns over selection bias. Although there was no school-level attrition, which protects against large reductions in statistical power to identify intervention effects, the study did not have statistical power to detect as small of effect sizes as anticipated.

Conclusions

Breakfast in the classroom in combination with breakfast-specific nutrition education is an effective method to increase School Breakfast Program participation. However, it was not an effective strategy to prevent overweight and obesity among low-income, urban, older elementary and middle school–aged children, and, in fact, increased incident and prevalent obesity. It is essential to identify alternative implementation approaches that increase School Breakfast Program participation without promoting obesity. Further research is necessary to understand whether breakfast in the classroom poses a similar risk for obesity among differing populations of students and whether alternative implementation models can increase School Breakfast Program participation without unfavorable consequences on children’s weight status.

Back to top
Article Information

Accepted for Publication: December 6, 2018.

Published Online: February 25, 2019. doi:10.1001/jamapediatrics.2018.5531

Correction: This article was corrected on May 28, 2019, to fix errors in the Results section and Table 2 due to errors in the statistical code used.

Corresponding Author: Gary D. Foster, PhD, Center for Obesity Research and Education, College of Public Health, Temple University, 3223 N Broad St, Ste 175, Philadelphia, PA 19046 (gfoster@temple.edu).

Author Contributions: Ms Polonsky and Dr Bauer are co–first authors. Dr Hanlon had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Polonsky, Fisher, Davey, Sherman, Abel, Hanlon, Foster.

Acquisition, analysis, or interpretation of data: Polonsky, Bauer, Fisher, Hanlon, Ruth, Dale, Foster.

Drafting of the manuscript: Polonsky, Bauer, Davey, Hanlon.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Polonsky, Bauer, Davey, Hanlon, Ruth.

Obtained funding: Davey, Foster.

Administrative, technical, or material support: Polonsky, Bauer, Fisher, Hanlon, Foster.

Supervision: Foster.

Conflict of Interest Disclosures: Mss Polonsky, Abel, Ruth, and Dale and Drs Bauer, Fisher, Davey, Sherman, and Foster reported receiving funding from the US Department of Agriculture (USDA). Dr Fisher reported being a paid consultant on an unrelated project for Nestle. Dr Sherman reported receiving grants from Temple University and the USDA Supplemental Nutrition Assistance Program (SNAP)–Ed. Dr Abel reports receiving grants from Temple University (CORE) and USDA SNAP-Ed. Dr Foster reported being an employee and shareholder of WW (formerly Weight Watchers). No other disclosures were reported.

Funding/Support: This research was supported by grant AFRI 2012-68001-19616 from the US Department of Agriculture (Dr Foster).

Role of the Funder/Sponsor: The funding organization 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.

Data Sharing Statement: See Supplement 3.

Additional Contributions: The authors thank the School District of Philadelphia, particularly Wayne Grasela, BS, and Amy Virus, MS, from the Division of Food Services, for their support in completing the study. Neither individual was financially compensated for their contribution.

References
1.
Frisvold  DE.  Nutrition and cognitive achievement: an evaluation of the School Breakfast Program.  J Public Econ. 2015;124:91-104. doi:10.1016/j.jpubeco.2014.12.003PubMedGoogle ScholarCrossref
2.
Bartfeld  JS, Ahn  HM.  The School Breakfast Program strengthens household food security among low-income households with elementary school children.  J Nutr. 2011;141(3):470-475. doi:10.3945/jn.110.130823PubMedGoogle ScholarCrossref
3.
School Breakfast Scorecard.  School Year 2016-2017. Washington, DC: Food Research & Action Center; 2018.
4.
Bernstein  LS, McLaughlin  JE, Crepinsek  MK, Daft  LM.  Evaluation of the School Breakfast Program Pilot Project: Final Report. Washington, DC: US Dept of Agriculture Food & Nutrition Service; 2004.
5.
Schwaab  B, Fettweis  C, Caplon  M.  Increasing student breakfast consumption: an established model in Maryland: interview by Jodi R. Godfrey.  Child Obes. 2012;8(2):162-166. doi:10.1089/chi.2012.0082.godfPubMedGoogle ScholarCrossref
6.
Butcher-Powell  LM, Bordi  PL, Borja  M, Cranage  D, Cole  C.  Factors affecting breakfast intake in children.  Topics Clin Nutr. 2003;18(2):130-135. doi:10.1097/00008486-200304000-00009Google ScholarCrossref
7.
Healthy, Hunger-Free Kids Act of 2010, S 3307, 111th Congress (2010). https://www.congress.gov/bill/111th-congress/senate-bill/3307. Accessed May 15, 2018.
8.
 School Breakfast Program Resolution. Washington, DC: US Dept of Agriculture; 2009.
9.
Lawman  HG, Polonsky  HM, Vander Veur  SS,  et al.  Breakfast patterns among low-income, ethnically-diverse 4th-6th grade children in an urban area.  BMC Public Health. 2014;14:604. doi:10.1186/1471-2458-14-604PubMedGoogle ScholarCrossref
10.
Niemeier  HM, Raynor  HA, Lloyd-Richardson  EE, Rogers  ML, Wing  RR.  Fast food consumption and breakfast skipping: predictors of weight gain from adolescence to adulthood in a nationally representative sample.  J Adolesc Health. 2006;39(6):842-849. doi:10.1016/j.jadohealth.2006.07.001PubMedGoogle ScholarCrossref
11.
Albertson  AM, Franko  DL, Thompson  D,  et al.  Longitudinal patterns of breakfast eating in black and white adolescent girls.  Obesity (Silver Spring). 2007;15(9):2282-2292. doi:10.1038/oby.2007.271PubMedGoogle ScholarCrossref
12.
Berkey  CS, Rockett  HR, Gillman  MW, Field  AE, Colditz  GA.  Longitudinal study of skipping breakfast and weight change in adolescents.  Int J Obes Relat Metab Disord. 2003;27(10):1258-1266. doi:10.1038/sj.ijo.0802402PubMedGoogle ScholarCrossref
13.
Szajewska  H, Ruszczynski  M.  Systematic review demonstrating that breakfast consumption influences body weight outcomes in children and adolescents in Europe.  Crit Rev Food Sci Nutr. 2010;50(2):113-119. doi:10.1080/10408390903467514PubMedGoogle ScholarCrossref
14.
Deshmukh-Taskar  PR, Nicklas  TA, O’Neil  CE, Keast  DR, Radcliffe  JD, Cho  S.  The relationship of breakfast skipping and type of breakfast consumption with nutrient intake and weight status in children and adolescents: the National Health and Nutrition Examination Survey 1999-2006.  J Am Diet Assoc. 2010;110(6):869-878. doi:10.1016/j.jada.2010.03.023PubMedGoogle ScholarCrossref
15.
Tin  SP, Ho  SY, Mak  KH, Wan  KL, Lam  TH.  Breakfast skipping and change in body mass index in young children.  Int J Obes (Lond). 2011;35(7):899-906. doi:10.1038/ijo.2011.58PubMedGoogle ScholarCrossref
16.
Utter  J, Scragg  R, Mhurchu  CN, Schaaf  D.  At-home breakfast consumption among New Zealand children: associations with body mass index and related nutrition behaviors.  J Am Diet Assoc. 2007;107(4):570-576. doi:10.1016/j.jada.2007.01.010PubMedGoogle ScholarCrossref
17.
So  HK, Nelson  EA, Li  AM,  et al.  Breakfast frequency inversely associated with BMI and body fatness in Hong Kong Chinese children aged 9-18 years.  Br J Nutr. 2011;106(5):742-751. doi:10.1017/S0007114511000754PubMedGoogle ScholarCrossref
18.
Sandercock  GR, Voss  C, Dye  L.  Associations between habitual school-day breakfast consumption, body mass index, physical activity and cardiorespiratory fitness in English schoolchildren.  Eur J Clin Nutr. 2010;64(10):1086-1092. doi:10.1038/ejcn.2010.145PubMedGoogle ScholarCrossref
19.
Borradaile  KE, Sherman  S, Vander Veur  SS,  et al.  Snacking in children: the role of urban corner stores.  Pediatrics. 2009;124(5):1293-1298. doi:10.1542/peds.2009-0964PubMedGoogle ScholarCrossref
20.
Grynbaum  MM. Hiccup in most important meal. New York Times. April 20, 2012; A1.
21.
Van Wye  G, Seoh  H, Adjoian  T, Dowell  D.  Evaluation of the New York City breakfast in the classroom program.  Am J Public Health. 2013;103(10):e59-e64. doi:10.2105/AJPH.2013.301470PubMedGoogle ScholarCrossref
22.
Corcoran  SP, Elbel  B, Schwartz  AE.  The effect of breakfast in the classroom on obesity and academic performance: evidence from New York City.  J Policy Anal Manage. 2016;35(3):509-532. doi:10.1002/pam.21909PubMedGoogle ScholarCrossref
23.
Gleason  PM, Dodd  AH.  School breakfast program but not school lunch program participation is associated with lower body mass index.  J Am Diet Assoc. 2009;109(2)(suppl):S118-S128. doi:10.1016/j.jada.2008.10.058PubMedGoogle ScholarCrossref
24.
Kuczmarski  RJ, Ogden  CL, Guo  SS,  et al.  2000 CDC Growth Charts for the United States: methods and development.  Vital Health Stat 11. 2002;(246):1-190.PubMedGoogle Scholar
25.
Flegal  KM, Shepherd  JA, Looker  AC,  et al.  Comparisons of percentage body fat, body mass index, waist circumference, and waist-stature ratio in adults.  Am J Clin Nutr. 2009;89(2):500-508. doi:10.3945/ajcn.2008.26847PubMedGoogle ScholarCrossref
26.
Lawman  HG, Mallya  G, Veur  SV,  et al.  Trends in relative weight over 1 year in low-income urban youth.  Obesity (Silver Spring). 2015;23(2):436-442. doi:10.1002/oby.20928PubMedGoogle ScholarCrossref
27.
Lawman  HG, Ogden  CL, Hassink  S, Mallya  G, Vander Veur  S, Foster  GD.  Comparing methods for identifying biologically implausible values in height, weight, and body mass index among youth.  Am J Epidemiol. 2015;182(4):359-365. doi:10.1093/aje/kwv057PubMedGoogle ScholarCrossref
28.
Foster  GD, Sherman  S, Borradaile  KE,  et al.  A policy-based school intervention to prevent overweight and obesity.  Pediatrics. 2008;121(4):e794-e802. doi:10.1542/peds.2007-1365PubMedGoogle ScholarCrossref
29.
Preisser  JS, Lohman  KK, Rathouz  PJ.  Performance of weighted estimating equations for longitudinal binary data with drop-outs missing at random.  Stat Med. 2002;21(20):3035-3054. doi:10.1002/sim.1241PubMedGoogle ScholarCrossref
30.
Jakobsen  JC, Gluud  C, Wetterslev  J, Winkel  P.  When and how should multiple imputation be used for handling missing data in randomised clinical trials—a practical guide with flowcharts.  BMC Med Res Methodol. 2017;17(1):162. doi:10.1186/s12874-017-0442-1PubMedGoogle ScholarCrossref
31.
Anzman-Frasca  S, Djang  HC, Halmo  MM, Dolan  PR, Economos  CD.  Estimating impacts of a breakfast in the classroom program on school outcomes.  JAMA Pediatr. 2015;169(1):71-77. doi:10.1001/jamapediatrics.2014.2042PubMedGoogle ScholarCrossref
32.
Antonogeorgos  G, Panagiotakos  DB, Papadimitriou  A, Priftis  KN, Anthracopoulos  M, Nicolaidou  P.  Breakfast consumption and meal frequency interaction with childhood obesity.  Pediatr Obes. 2012;7(1):65-72. doi:10.1111/j.2047-6310.2011.00006.xPubMedGoogle ScholarCrossref
33.
Kral  TV, Allison  DB, Birch  LL, Stallings  VA, Moore  RH, Faith  MS.  Caloric compensation and eating in the absence of hunger in 5- to 12-y-old weight-discordant siblings.  Am J Clin Nutr. 2012;96(3):574-583. doi:10.3945/ajcn.112.037952PubMedGoogle ScholarCrossref
34.
Carnell  S, Benson  L, Gibson  EL, Mais  LA, Warkentin  S.  Caloric compensation in preschool children: relationships with body mass and differences by food category.  Appetite. 2017;116:82-89. doi:10.1016/j.appet.2017.04.018PubMedGoogle ScholarCrossref
35.
Sherman  S, Grode  G, McCoy  T,  et al.  Corner stores: the perspective of urban youth.  J Acad Nutr Diet. 2015;115(2):242-248. doi:10.1016/j.jand.2014.10.017PubMedGoogle ScholarCrossref
36.
Bauer  KW, Chuisano  S. Intentional Self-Regulation of Eating Among Children and Adolescents. In: Lumeng JC, Fisher JO, eds.  Pediatric Food Preferences and Eating Behaviors. Philadelphia, PA: Elsevier; 2018:255-270. doi:10.1016/B978-0-12-811716-3.00013-0
37.
Foster  GD, Linder  B, Baranowski  T,  et al; HEALTHY Study Group.  A school-based intervention for diabetes risk reduction.  N Engl J Med. 2010;363(5):443-453. doi:10.1056/NEJMoa1001933PubMedGoogle ScholarCrossref
×