Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Effects of the school breakfast programme

Diet

Before the initiation of the school breakfast programme, there were no differences (p > .10) between groups in the intake of energy, protein, and iron (table 5). After implementation of the programme, significant differences were seen in favour of the experimental group. In particular, the mean values for experimental and control groups, respectively, were 2,182 versus 1,731 kcal (p <.001) for energy, 56.1 versus 43.6 g (p < .001) for protein, and 21.6 versus 12.5 mg (p < .001) for iron. The school breakfast made the single most significant contribution to these observed differences (p < .001).

Attendance

No significant differences in attendance rates at baseline September 1993 were found between the experimental (90.90%) and control (89.28%) groups. When the breakfast was implemented, October 20 through November 8, attendance increased by 0.58 points (not significant) in the experimental schools, but it declined by 2.92 points (p <.05) among controls. As can be seen in figure 1, the inter-group difference was statistically significant (p < .05).

In the ensuing period, when both groups were receiving the breakfast, attendance rates increased significantly (p < .05), reaching 92.80% and 88.25% in the experimental and control groups, respectively. In addition, inter-group differences also reached significance (p < .05).

Psychoeducational effects

The unadjusted means and standard deviations of the studied outcomes, as well as the probability values of the t tests performed on the adjusted test scores (leastsquares method), are shown in table 6. No significant differences between groups were found. Accordingly, we proceeded to analyse the data at the individual level to test for interaction effects.

The ANCOVA model at the individual level included the confounders described and selected on the basis of their correlations with the outcomes. The child's age was excluded, as it was associated with three educational variables of interest: repetition of a grade (r = .77; p = .001), height-for-age Z score (r= - .38; p=.001), and age at entering school (r = .44; p = .001).

As shown in table 7, the vocabulary model was the only significant one, with an F value of 1.97. This model shows a positive, although not significant, effect of the treatment, confirming the t test. On the other hand, the interaction weight-R x treatment was positive and significant (parameter=.37; F= 4.97; p < .05). Among those in the treatment group, the greater their weight, the better their vocabulary test scores (fig. 2).

TABLE 5. Mean (SD) energy (kcal), protein (g), and iron (mg) consumption, estimated by 24-hour recall by study group and evaluation period (n » 58/group)

^a. This period includes the consumption of snacks in school, and at time 2, the school breakfast in the Rx group. Study tests were administered at approximately 11.00 a.m.
^b. Rx vs control, p < .001.

FIG. 1. School attendance before the school breakfast programme (September), when the Rx group entered the programme (October), and when the programme was fully implemented ( November)

TABLE 6. Mean (SD) of outcome variables by study group

^a. Not significant for a t test of the difference between least-square means (LSM) of troth groups at time 2 The LSM are adjusted for test score at time 0, and the school is the unit of analysis.

Discussion

The three objectives most frequently cited for implementing school feeding programmes are nutritional benefits, increases in enrollment and attendance, and enhanced learning performance. This study addressed the last two of these objectives and offered an approximation to the first. Briefly, the data show that on average, the school breakfast programme resulted in increased dietary intake and improved attendance. It also enhanced performance on a vocabulary test among a subset of the subjects.

Although the study design did not include an assessment of nutritional impact, it is to be expected that the increased nutrient intake may have a salutary health effect in the long term. Considering the high nutritional quality of the breakfast, the case of iron is of particular significance. According to our data, 100% of the child's RDA for iron is provided, about 9.5 mg/day. Under a more conservative approach, 6.7 mg/day of iron is provided, accounting for a potential 30% overestimation of our dietary assessment [16], a figure that is about the median basal requirement of 6.3 mg/day [13]. Over the long term, the school breakfast programme may well help reduce the high prevalence of iron-deficiency anaemia, thus directly contributing to the educability of these children [20].

The data confirm once again that school feeding programmes are an incentive for poor families to send their children to school. Of importance in this study is that this effect was observed over a period of 30 days. Evidence of similar effects over longer periods ( > 3 months) has been reported for both developing [7] and developed countries [4]. Because of its effects on attendance, it is possible to infer that the programme could be an incentive to keep children in school, a major problem in the educational system in Peru. Statistics on different cohorts of children from the schools in this study show a range of 5% to 15% in drop-out rates in the first grade of primary school.

TABLE 7. Parameter estimates and [F values] of the effects of the Peruvian school breakfast programme on cognitive tests adjusted for baseline values, in Huaraz, Peru (N= 326)

^a. Parameters correspond to an ANCOVA (GLM in SAS) model adjusted for baseline test scores and the following confounders: sex, height-for-age z score, weightR, SES, language spoken at home, grade at school, repetition of any grade, school, and age on entry to school.
^b. Error term includes the effect of school nested in study group.
^c. Residual of weight after regressing it on height and age.
^d p <.05.

The coding, reading comprehension, vocabulary, and mathematics tests had relatively robust correlations (r< .65) between the pretreatment and posttreatment evaluations. Among them, the vocabulary test was the only one whose variability in the change of scores was great enough to be explained by the regression model. Moreover, the results of the regression showed that the treatment x weight-R interaction accounted for a significant portion of the variance of the vocabulary scores.

Paradoxically, among children who received breakfast, the heavier children benefited the most. Moreover, among the controls, the heavier children performed less well than the rest of the group. This is in agreement with what was observed for all children before treatment (data not presented). Without reaching statistical significance, weight-R was negatively associated with performance on the six tests. The same result was obtained in a similar analysis using the children who participated in a parallel study in the city of Huaraz.

When the weight-for-height Z score was entered in the model instead of weight-R, it yielded similar statistical results. First, we concluded that the weight-R variable is primarily the sum total of the weight of adipose tissue, water, and skeletal structure, in other words, a measure of current nutrition status. The issue is why stunted children with comparatively high weight should tend to score lower in the vocabulary test and benefit more from breakfast than those with lower weight.

Short stature coupled with normal weight-for-height is characteristic of poor Peruvian populations in both rural [21] and urban [22] environments. This phenomenon is established early in life and reflects a protracted deficit of critical nutrients due to poor diet and infection, although not starvation. According to one report [23], these anthropometric characteristics of poor Peruvian children under five years of age are likely to be a reflection of increased total body water, which may indicate a rise in the hydration of fat-free tissue related to malnutrition. Thus, what we are likely to have observed is that those who tended to score low in the vocabulary test, as well as those who benefited from the breakfast, were of poor health status or undernourished. This is an issue that certainly merits further investigation.

The beneficial effects of the breakfast on the vocabulary test must have been mediated by short-term metabolic changes. A considerable amount of experimental data has been gathered to show that skipping breakfast has adverse effects on cognitive function [3, 24, 25]. As previously noted, the second study related to the school feeding programme in Peru observed that short-term memory retrieval was also delayed under a no-breakfast condition in those children classified as nutritionally at risk. In another publication, we suggested that glucose and insulin changes and correlated hormonal changes activated by prolonged fasting in children may have direct effects on the levels of substrates in the brain that are involved in cognitive function [3].

FIG. 2. Interaction treatment x weight-R for vocabulary test score

At issue now is which particular cognitive processes involved in the vocabulary test were affected by the breakfast. Besides paying attention to the verbal information that must be processed, a subject must recall distinct attributes that define a particular verbal stimulus. In our view, the retrieval and use of verbal material was probably facilitated by the availability of nutrients. Alternatively, the school breakfast programme could have enhanced learning and the acquisition and storage of new information. However, this does not seem likely, because of the very short time between the pretreatment and posttreatment evaluations. The effect on retrieval could be a direct and selective effect on memory or could be secondary to a salutary influence on vigilance and attention. However, there is a more convincing argument in favour of a selective memory effect, because none of the other tests-coding, math, and reading-discriminated between the treatment conditions. Of importance is that, in the second study we conducted of the school breakfast programme, strong evidence showed that the breakfast prevented delays in retrieving information from working memory among stunted and wasted children.

At this point, it is pertinent to address the issue of whether the observed effects were mediated by the breakfast. It may be argued that the study groups differed in factors other than the three we used (language spoken at home, age on entering school, and grade), and that the absence of a placebo created a diversity of unintended effects. However, the inherent nature of the interaction between treatment and weight, which rests on body weight, provides unequivocal support for a nutritional explanation. Furthermore, we documented an increment in dietary consumption associated with the programme's breakfast.

It is worth noting that the characteristics of our sample confine the external validity of the findings to fourth- and fifth-graders. This group of children not only developed reading and writing skills, but also represents a selective cohort that survived the high drop-out rates observed during the first years of primary school [26].

We conclude that the school breakfast programme in Peru has tangible beneficial effects in the short term, and that a long-term evaluation of the programme is warranted.

Acknowledgements

The study was partially funded by the Kellogg Co., Battle Creek, Michigan, USA, and by the Fondo Nacional de Compensación pare el Desarrollo Social (FONCODES) of the Government of the Republic of Peru. The critical comments and suggestions of Roberto Frisancho, Ph.D., Rudolph Leibel, M.D., and Richard Trowbridge, M.D., are gratefully acknowledged.

References

1. Paige DM, Cordano A, Huang S. Nutritional supplementation of disadvantaged elementary-school children. Pediatrics 1976;58:697-703.

2. O'Sullivan JB. Age gradient in blood glucose levels: magnitude and clinical implications. Diabetes 1974;23: 713-5.

3. Pollitt E, Lewis NL, Garza C, Schulman J. Fasting and cognitive function. J Psychol Res 1982-83;17:169-74

4. Meyers AF, Sampson AK, Weitzman M, Rogers BL, Kayne H. School breakfast program and school performance. Am J Dis Child 1989;143:1234-9.

5. Pollitt E, Gersovitz MS, Gragiulo BS. Educational benefits of the United States school feeding program: a critical review of the literature. Am J Publ Health 1978; 68:477-81.

6. Radzikowski J, Gale S. The national evaluation of school nutrition programs: conclusions. Am J Clin Nutr 1984;40:454-61.

7. Levinger B. School feeding programs: myth and potential. Prospects 1984;XIV(3):369-76.

8. Beaton GH. Evaluation of nutrition interventions: methodologic considerations. Am J Clin Nutr 1982;35: 1280-9.

9. Beaton GH, Ghassemi H Supplementary feeding programs for young children in developing countries. Am J Clin Nutr 1982;35:863-916.

10. FONCODES, UNICEF, Instituto CUANTO. El Mapa de la Inversión Social: Pobreza y Actuación de FONCODES a nivel departamental y provincial. Lima: FONCODES, 1994.

11. WHO. Energy and protein requirements. Report of a joint FAO/WHO/UNU expert consultation. Technical report series no. 724. Geneva: WHO, 1985.

12. National Research Council. Recommended dietary allowances. 10th ed. Report of the subcommittee on the tenth edition of the recommended dietary allowances. Food and Nutrition Board, Commission on Life Sciences. Washington, DC: National Academy Press, 1989

13. FAO/WHO. Expert consultation. Requirements of vitamin A, iron, folate and vitamin B₁₂. Rome: FAO, 1988.

14. Instituto de Investigación Nutricional. Evaluación del impacto del Programa de Desayunos Escolares implementado por FONCODES Lima: FONCODES, 1994.

15. Donner A, Birkett N, Buck C. Randomization by cluster. Am J Epidemiol 1981;114:906-14

16. Gibson RS. Principles of nutritional assessment. Oxford: Oxford University Press, 1990:37-42

17. Wechsler D. Manual for the Wechsler intelligence scale for children-revised. San Antonio, Tex, USA: The Psychological Corporation, 1974.

18. Manuel HT. Technical reports, tests of general ability and tests of reading. Interamerican Series. San Antonio, Tex, USA: Guidance Testing Associates, 1967.

19. Pollitt E, Gorman KS, Engle PL, Martorell R, Rivera J. Early supplementary feeding and cognition. Monographs of the Society for Research in Child Development. Serial no. 235, 1993;58(7).

20. Pollitt E. Iron deficiency and cognitive function. Annu Rev Nutr 1993;13:521-37.

21. Graham GG, MacLean WC, Kallman CH, Rabold J, Mellis ED. Urban-rural differences in the growth of Peruvian children. Am J Clin Nutr 1980;33:338-44.

22. Lopez de Romana G, Brown KH, Black RE, Kanashiro HC. Longitudinal studies of infectious diseases and physical growth of infants in Huascar, an underprivileged pert-urban community of Lima, Peru. Am J Epidemiol 1989;129:769-89.

23. Trowbridge FL, Marks JS, Lopez de Romaña G, Madrid S, Bouton TW, Klein PD. Body composition of Peruvian children with short stature and high weight-for-height. II. Implications for the interpretation of weight-for-height as an indicator of nutritional status. Am J Clin Nutr 1987;46:411-8.

24. Conners CK, Blouin AG. Nutritional effects on behavior of children. J Psychiatr Res 1982-83;17:193-201.

25. Simeon DT, Grantham-McGregor S. Effects of missing breakfast on the cognitive functions of school children of differing nutritional status. Am J Clin Nutr 1989;49: 646-53.

26. Unesco Regional Office for Education in Latin America and the Caribbean (Unesco/OREALC). The state of education in Latin America and the Caribbean. Major projects in the field of education 1980-1985. Santiago, Chile: Unesco, 1989.

Contents - Previous - Next