Diet, Breakfast, and Academic Performance in Children

Abstract

Over the past few decades, it has become clear that many children in the USA experience chronic mild-to-moderate food deprivation, defined as food insufficiency and/or hunger. A recent survey conducted by the US Department of Agriculture reported the prevalence of hunger/food insufficiency among children in the USA to be 18%, even during a period of remarkable overall financial prosperity for the country as a whole. In recent reports, we have explored the relationship between hunger and academic and psychosocial functioning during childhood and demonstrated that children whose parents reported hunger/food insufficiency were more likely to have repeated a grade in school, received special education and/or mental health counseling, to have gotten lower scores on tests of psychological adjustment, lower grades in math, and to have had more days tardy and absent from school.

Despite the consistent and inverse relationship between hunger and child functioning, the nutritional correlates of this association remain largely unexplored. In one of the few studies to date , about one third of inner city public school children had low caloric intakes and/or low intakes of selected micronutrients. These children were significantly more likely to report hunger and had significantly worse grades in school, higher rates of absenteeism, and more psychological problems in comparison to children with more adequate dietary intakes. A recent study by another group was based on a large and nationally representative sample and documented the link between hunger/food insufficiency and health problems during childhood .

One way to combat problems associated with child hunger would be to augment an existing governmental food program. Previous studies by our group have shown that it is possible to increase participation in the school breakfast program by making it universally free to all students and that students who increase their school breakfast participation show significant gains in academic and psychosocial functioning as well as decreased hunger. However, despite the documented benefits of school breakfast programs in terms of reducing hunger and improving child psychosocial functioning, the nutritional benefits of these programs remain unclear, as does the link between improvements in nutritional status and improvements in academic and behavioral functioning.

Previous studies have shown that failure to eat breakfast or lack of nutritious food for breakfast is common in childhood and may increase the risk of being undernourished. Several studies have suggested a significant relationship between skipping breakfast and low nutrient intakes in children . In the current study, we assessed the relationship between child nutritional risk status, hunger, participation in a school breakfast program, academic performance, and psychosocial functioning.

Methods

Data for the current analyses came from a collaborative study of a universal-free school breakfast program (USBP) in the Boston Public Schools. Children and their parents in three inner city schools were recruited for this study, prior to the implementation of the USBP. The parents of all 227 students in grades 4 through 6 of these schools were invited to participate in the study through letters that were sent home with students, and only 1 child per family was eligible. One hundred forty-nine (66%) students returned permission slips from parents indicating a willingness to participate in the study.

When interviews were scheduled, 50 of these parents were unavailable to be interviewed, leaving a sample of 99 families (66% of the 149 consenting students and 44% of the initial cohort of 227 students). Children were re-interviewed approximately 6 months after the implementation of the USBP. Of the 99 families who had completed the initial interview, 1 (1%) had moved out of town, and 1 (1%) was unable to complete the re-interview, leaving a final sample of 97 subjects with complete pre- and post-USBP parent/student interviews.

The study was approved by the Subcommittee on Human Studies at the Massachusetts General Hospital, Boston (IRB#1999P-006455), and by the research committee of the Boston Public Schools. Participation by parents and children was voluntary and access to school records was made possible through separate consent forms signed by parents and children.

The children’s gender, age, ethnicity, and parental marital status were assessed from a questionnaire administered during the initial interview with parents. A validated 24-hour dietary recall method using standardized graduated food models was adapted for use in interviewing children . A single 24-hour dietary recall was collected on each study participant before and 6 months after the start of the USBP and was analyzed using the Minnesota Nutritional Data System software version 2.3. The dietary recall period reflected the previous 24 h. For the purposes of this study, dietary intake was evaluated as the percent of RDA for age and gender for specific nutrients and energy .

The total daily energy intake of each child was assessed at the initial interview and was compared with the RDA for age and gender . Children were considered to consume a low-energy diet if their total daily energy intake was ≤50% of the RDA. Because several national studies have shown that children in the USA consume substantially less than the recommended daily amounts of vitamins A, B6, B12, C, and folate, iron, zinc, and calcium , the intake of these 8 micronutrients was assessed from the 24-hour recall and compared with the RDAs. Although other investigators have explored using a cut-off of ≤67% of RDA as a standard for low nutrient intake , we adopted a somewhat more stringent standard and considered an intake of a particular nutrient as low if the daily nutrient consumption was ≤50% of the RDA.

Children who consumed two or more nutrients at ≤50% of the RDA and/or who had an energy intake of ≤50% RDA were aggregated into a low nutrient intake group for the purpose of examining relationships between nutrient intake and other outcome variables. Children who did not meet these criteria for low nutrient intake were considered to have adequate nutrient intakes for the purposes of the study.

School breakfast participation was coded from official school food service records at the schools. For each student, food service staff provided data on daily participation for 1 week prior to the start of the USBP, and another week, after the USBP had been running for approximately 6 months. Data on each student’s daily school attendance was also collected for the same 1-week periods to permit the calculation of a school breakfast participation rate (number of days eating school breakfast divided by number of days present; range from 0 to 100%). Children were considered to participate ‘often’ if they ate breakfast 80% or more of days present, to participate ‘sometimes’ if they ate between 20 and 79% of days present, and to participate ‘rarely’ if they ate breakfast in school <20% of the time.

Parents completed an 8-item hunger/food insufficiency questionnaire developed by the Community Childhood Hunger Identification Project (CCHIP) at the time of the interviews. Children were classified as ‘hungry’ if the parent responded positively to 5 or more of the 8 questions concerning hunger in the past year. Children were classified as ‘at risk for hunger’ if the parent responded positively to one or as many as 4 of the 8 food insufficiency questions. If the parent did not respond positively to any of the 8 food insufficiency questions, the household and child were classified as ‘not hungry’. Parents were only interviewed in the fall, so on the Parent CCHIP form, only pretest data are available.

Children were asked to fill out a 5-item version of the CCHIP survey, the Child Hunger Index Child Report (CHI-C), which has been shown to be a valid and reliable instrument to identify hunger by child report . Children who responded positively to 2 or more of the CHI-C questions were classified as ‘hungry’. When the child responded positively to one of the questions, the child was classified as ‘at risk for hunger’ on the CHI-C. Those who did not respond positively to any of the items were classified as ‘not hungry’.

The Pediatric Symptom Checklist (PSC) consists of 35 items that are reported ‘never’, ‘sometimes’, or ‘often’ present and scored 0, 1 or 2, respectively. A total score is obtained by adding the scores for each of the items. For the PSC, higher scores indicate worse functioning. Also used in the study was the youth self-report version of the Pediatric Symptom Checklist (PSC-Y) , which is scored the same way as the parent-completed form.

Official school records were reviewed to obtain student grades for Math, Reading, Science, and Social Studies and their absence and tardiness rates for the full 1998–1999 (pre-USBP) and 1999–2000 (post-USBP) school years. Letter grades were converted into numeric values based on a 4.0 grading system (A = 4.0, B = 3.0, C = 2.0, D = 1.0, F = 0.0), and summed to generate an overall grade point average.

Because data on prior year academic performance for middle school students are kept in their original elementary schools, and it was not feasible to travel to each middle school student’s elementary school to obtain his/her grades and attendance for the year prior to the first year of the school breakfast program, the year two and change data reported in this study for these variables concern only the 79 students who were in elementary school the year before the study began.

The goal of our data analyses was to examine the association between dietary intake and hunger, school breakfast participation, psychosocial functioning, and academic performance scores. Categorical group differences were examined using Pearson χ2 analysis, and continuous score differences among dietary intake groups were examined using one-way analysis of variance.

Results

The socioeconomic and ethnic characteristics of the children were similar in all three schools (>70% of the children were eligible for free or reduced-price meals, and >70% of the children were of African-American or Hispanic descent). In the combined sample, approximately half of the children were male (41%; 40/97). Eighty-two percent (80/97) were from elementary grades (3 through 5), and 18% (17/97) were from middle school (grade 6). Fifty-five percent (53/97) of the children were from single-parent families.

Twenty-nine percent (28/97) of the children had two or more nutrient intakes that were ≤50% of the RDA. Thirteen percent (13/97) of the children consumed ≤50% of the RDA for energy. Combining the two groups, 30% (29/97) of these children consumed two or more of the micronutrients and/or a total energy intake of ≤50% of the RDA. Children in the low nutrient intake group did not differ significantly from their peers with respect to gender, grade, or parental marital status.

Data on low nutrient intake and child functioning prior to the start of the USBP are shown in table 1. Prior to the introduction of the USBP, 61% (59/97) of the children were classified as eating school breakfast rarely, 16% (15/97) as eating school breakfast sometimes, and 24% (23/97) ate school breakfast often. Children who ate school breakfast rarely were significantly more likely (41%) to be nutritionally at risk than were children who ate school breakfast sometimes (7%) or often (17%; χ2 = 5.8, d.f. = 2, p < 0.05).