Effects of eating breakfast on children and adolescents: A systematic review of potentially relevant outcomes in economic evaluations

Background Breakfast is often described as the most important meal of the day. Several studies have focused on examining if breakfast habits have any short-term effects on school attendance, academic achievement, and general health in children and adolescents. Informed decisions of whether to promote eating breakfast or not require a more long-term perspective. Objective The aim of this study was to conduct a systematic review of scientific publications studying the effects identified as potentially relevant for the economic evaluation of eating breakfast in children and adolescents. Design A systematic literature review was conducted. Studies were identified by searching the electronic databases PubMed, CINAHL, Web of Science, and PsycINFO between January 2000 and October 2017. The inclusion criteria applied were published articles from peer-reviewed journals with full text in English, quantitative studies collecting primary data with school-aged children, and adolescents aged from 6 to 18 years as participants, performed entirely or partly in countries with advanced economies, except Japan and Taiwan. Results Twenty-six studies fulfilled the inclusion criteria, and studies that were judged to be of at least moderate quality were included in the analysis. The results of the review of eating breakfast studies showed positive and conclusive effects on cognitive performance, academic achievement, quality of life, well-being and on morbidity risk factors. Conclusions The overall assessment of the studies indicated positive effects of eating breakfast. How the identified effects influence societal costs and an individual’s quality-adjusted life years require further research.

B reakfast is often described as the most important meal of the day. Several studies have focused on examining if breakfast habits have any bearing on school attendance, academic achievement, and general health in children and adolescents. Research findings indicate that the regularity of wholesome, daily meal patterns comprised of healthy foods is important for the physical and mental development of children and adolescents (1). The association between adiposity and breakfast habits among children and adolescents is a well-studied Popular scientific summary • The article reviews the potential study outcomes for the economic evaluation of long-term effects of eating breakfast in children and adolescents. • The studies showed positive and conclusive effects on cognitive performance, academic achievement, quality of life, well-being, and on morbidity risk factors. • How the identified effects influence societal costs and an individual's quality-adjusted life years require further research. • Development of simulation models to estimate the long-term costs and effects of eating breakfast should be considered.
topic (2). The quality and regularity of breakfast consumption have also shown to be associated with mental performance, academic achievement, physical activity, and quality of life (3)(4)(5)(6)(7)(8)(9). Despite this, young people often skip breakfast (10), and the prevalence of children and adolescents skipping breakfast is increasing (11). In addition, it has been shown that breakfast skipping is particularly common amongst children of lower socioeconomic status (12). As societal resources are scarce, and needs are endless, choices between health interventions must be made. Several published reviews have examined varying shortterm effects of eating breakfast on children and adolescents (6,(13)(14)(15). However, informed decisions of whether to promote eating breakfast or not require a more longterm perspective. Economic evaluation seeks to identify, quantify, and compare the long-term costs and effects of different interventions for well-founded and informed decision-making. The costs are weighed against the health effects measured in such a way that it corresponds to a value, usually quality-adjusted life years (QALYs). This measure should ideally encapsulate the impact of an intervention on a person's length of life and the impact on their health-related quality of life. To our knowledge, it remains unstudied which short-term effects of breakfast eating that may serve as parameters for modeling longterm effects of eating breakfast in economic evaluations. As a starting point for future economic assessments, our intention was to review study outcome measures and identify those that could transform into long-term QALYs.
The aim of this study was to conduct a systematic review of scientific publications studying effects identified as potentially relevant for the economic evaluation of eating breakfast in children and adolescents.

Search methods for identification of studies
Studies were identified by searching the electronic databases PubMed, CINAHL, Web of Science, and PsycINFO between January 2000 and October 2017.
To capture all relevant articles published in the field, two different search strategies were constructed (see Table 1).

Selection of studies
One of the authors (ML) conducted the search in October 2017. Initially, two of the authors (ML and NEV) read all titles and abstracts of the identified studies to determine the relevance of each article. If title and abstract met with the inclusion criteria, the study proceeded to the next stage of the review process. Studies with insufficient information in title and abstract were also proceeded to the next stage of the review process. After this selection, all authors (ML, L-ÅL, and NEV) read the remaining studies in full text, in order to confirm the inclusion eligibility and conduct the quality assessment.
The inclusion criteria applied in the review were determined before the initial search. The criteria are presented as follows: • Published articles from peer-reviewed journals with full text in English. • Studies collecting primary data. • Quantitative studies. • Study participants in the age of 6 to 18 years. • Studies performed entirely or partly in countries listed by Central Intelligence Agency (CIA) as advanced economies, except Japan and Taiwan (16). • Studies using clear definitions of eating and not eating breakfast.

Review of quality
In pairs, the authors read the articles in full to assess both eligibility and scientific quality. If the two authors who made the initial assessment did not agree, the third author will also read the article. The scientific quality assessment was based on the following criteria: adequate control group(s), randomization, sufficient statistical power, control of confounders, sufficient descriptions of experimental design, relevance of outcome measures, and consistency between reported results and conclusions (17,18). During the selection process, the authors assessed the relevance of the outcome measures (i.e. study effects). Study effects deemed to have the potential to be transformed into long-term effects were selected and categorized into four topics: academic achievement, quality of life and well-being, morbidity risk factors, and cognitive Effects of eating breakfast on children and adolescents performance. Studies focusing on the association between adiposity, obesity, overweight, and eating breakfast were excluded, because the topic has already been studied thoroughly. Studies using other outcome measures were also excluded.
In order not to let the results of the study affect the quality assessment, this was done separately without, as far as possible, reading the study results. The criteria for quality assessment varied with different study types. The quality of the studies was rated as high, moderate, or low. A study rated "high quality" had to fulfill all the quality criteria. Studies of moderate quality fulfilled several of the criteria. Finally, low-quality studies either failed to meet several of the criteria or had major shortcomings in certain criteria. Studies rated as being of high or moderate quality were included in the further analysis. The low-quality studies were excluded. They are presented with a comment in Appendix 1. Table 2 contains the following information from the included studies: reference, participant characteristics (number, age, gender), study design, study purpose, outcomes, authors conclusions, and the scientific quality assessment.

Result compilation
A compilation of the study results based on statistical inference is presented in Table 3. A study was deemed positive if it had at least one statistically significant positive outcome measure, a study was deemed negative if it had at least one statistically significant negative outcome, and a study was deemed "no effect" if it showed no statistically significant results. A two-tailed p-value of 0.05 was considered statistically significant.

Result of the search
The flow chart presented in Fig. 1 illustrates the work process. Database searches identified 5,200 articles. After the removal of duplicates, 2,958 unique articles remained. Exclusion based on information given in title and abstract resulted in the removal of 2,908 articles. The full-text reading of the articles resulted in the additional exclusion of 16 articles that did not meet the inclusion criteria, and eight articles were excluded because of low quality (19-26) (see Appendix 1). Finally, 26 articles met the inclusion criteria and remained for further analysis.

Results
The 26 studies included in the analysis are summarized in Table 2.
All studies were published between 2003 and 2017. The countries represented were the UK (seven studies); the United States (three studies); Canada, Norway, Australia, and Sweden (two studies each); and Denmark, Germany, Italy, and the Netherlands (one study each). Four studies were multinational. Nineteen of the included studies were observational. Four of these were also longitudinal with follow-up periods of 27 years in two studies, 21 years and 6 months in one study each. Five studies employed randomization. One study was a cluster-randomized controlled trial, three studies employed randomized crossover designs, and one study used a cluster-randomized crossover design. One study was a non-randomized controlled trial, and one study employed a separate-group design.
The number of participants included in the studies varied from 29 to 28,608. Based on the predetermined inclusion criteria, the age range of all children and adolescents included in the studies were from 6 to 18 years. Four of the studies had participants with a mean age below 10 years (27)(28)(29)(30), the remaining 22 focused on studying participants with a mean age of 10 or more years (5,. The studies included a variety of outcome measures and instruments. The outcome measures that met inclusion criteria were grouped into four categories. Nine studies comprise the category of morbidity risk factors associated with eating or not eating breakfast (29)(30)(31)(32)(33)(34)(35)(36)(37). Eight studies form the academic achievement category, with outcome measures that study links between breakfast and academic achievement (5,28,(43)(44)(45)(46)(47)(48). Six studies analyzed the association between breakfast eating and cognitive disorders, comprising the category of cognitive performance (27, 38-40, 49, 50). Finally, the quality of life and/or well-being category is made up of three studies, analyzing links between breakfast eating and QoL and/or well-being (41,42,51). No study concerning the costeffectiveness of eating breakfast was found. Table 3 shows the compilation of the study results.

Morbidity risk factors
All of the nine studies with morbidity risk factors as outcomes were observational studies. Five of the studies examined the association between breakfast consumption and the metabolic syndrome. Marlatt et al. found that higher levels of breakfast consumption were significantly associated with lower risk for insulin resistance as well as other risk factors for metabolic syndrome (32). This is in line with what both Sese et al. and Wennberg et al. concluded in their studies (34,36,37). The two studies conducted by Wennberg et al. were longitudinal with 27 years of follow-up. In both studies, they found an association between poor breakfast habits and future risk of developing metabolic syndrome. Hallstrom et al. studied the association between breakfast consumption and cardiovascular disease risk factors. Their results indicated that adolescents who regularly consumed breakfast had higher cardiorespiratory fitness and lower total adiposity (31). Significant effects of breakfast consumption on blood lipid levels, blood pressure, or insulin resistance were found in overweight/obese male participants, while no significant effects were found among female participants, Effects of eating breakfast on children and adolescents Effects of eating breakfast on children and adolescents Effects of eating breakfast on children and adolescents  (27). Widenhorn-Müller showed beneficial effects of breakfast on short-term memory and mood but not on sustained attention (40).

Quality of life and well-being
The three included studies used different measurements to study the association between breakfast consumption and QoL and/or well-being. However, the results are consistent. Page et al. found that eating breakfast was one of Number of studies indicating negative effects 1 (14%) 0 (0%) 0 (0%) 0 (0%) Number of studies indicating no effects 0 (0%) 0 (0%) 0 (0%) 0 (0%) + = positive effect, − = negative effect, 0 = no effect, N/A= not applicable. the 12 significant predictors of self-rated health measures (41). Richards and Smith provided evidence showing that stress, anxiety, and depression were associated with skipping breakfast (42). Nevertheless, they highlighted the possibility that mental health might influence whether a person consumes breakfast or not. Smith et al. also found that breakfast consumption is linked to increased well-being (51).

Academic achievement
All the eight studies that investigated breakfast consumption and academic achievement found positive associations between breakfast consumption and academic achievement. In a study conducted by Lien in Norway, the implications of skipping breakfast on mental distress and academic achievement were stronger for boys than girls and stronger for native Norwegians than immigrants (45). Burrows et al. concluded that dietary behavior overall was associated with higher academic achievement, but that breakfast was only significantly associated with the domain writing (44).

Discussion
This review set out to find relevant outcome measures of eating breakfast on children and adolescents, for use in economic evaluations. The studies included focused mainly on studying the effects of eating versus not eating breakfast, and they were divided into four categories of study outcomes, namely morbidity risk factors, cognitive performance, quality of life and well-being, and academic achievement. All studies included in the analysis met the inclusion criteria of this review and were judged to be of at least moderate quality. Consequently, the excluded studies either failed to meet the inclusion criteria or had a low study quality rating.
There was a notable variation in the study design. In general, studies were either of observational or experimental design. Even though, randomized control trial is the only study design for drawing causal relationship, the observational study still is important to identify associations. Whether children eat breakfast or not and whether it has any substantial effect on study outcomes might very well be derived from a large number of unobserved influential factors, such as the children's social environment and parental support. Thus, assigning the effects of eating breakfast to its sole nutritional value may be problematic.
The review showed associations between eating breakfast and cognitive performance, academic achievement, quality of life, well-being, and morbidity risk factors. When results are consistently positive, different forms of bias need to be discussed. There is always a risk of publication bias that could contaminate this type of review, which could have overestimated the strength of the evidence (52). In addition, reformulations of initial study hypothesis to better suit data might affect results. This is known as HARKing-hypothesizing after the results are known (53). Also, the lack of control for confounders may have affected the results. As it is not possible to adjust for all confounders, it is difficult to establish a clear causal link between intervention and outcome (54). Studies that examine the association between breakfast and weight/obesity have been thoroughly studied and reviewed before. Therefore, such studies were excluded from this review. However, despite the exclusion of the weight/ obesity effects of eating breakfast in this review, it should be noted that it is relevant to include weight/obesity in the economic evaluations of breakfast.
To only include studies performed entirely or partly in countries listed by the CIA as advanced economies, except Japan and Taiwan, restricts the relevance of results to these countries. The potential exclusion of relevant studies from other nations is thus one limitation. In the assessment of study quality, the risk of incorrect classification is contingent. Although, by proactively disregarding the study results in the articles included when performing quality assessment, the authors took measures to limit the risk of study quality misclassification.
Our interest in finding and valuing effects that can be attributed to breakfast eating emanates from the question of the cost-effectiveness of breakfast interventions.
Our study cannot answer that question, but shows that breakfast eating is associated with effects that could be used in economic evaluations, especially when using decision analytic modeling, calculating the long-term cost and effects of intervention. For instance, improved cognitive performance and academic achievements may influence long-term effects on an individual's productivity through improved school results, grades, and higher education. Further, this gives higher human capital, which can be reflected in higher income, both for the individual and for society. Improvements in quality of life and well-being will have an immediate effect on the QALY weight of a child or adolescent but also, if the effects are sustained, on long-term QALY gain. Finally, the health effects mediated by a reduction of long-term morbidity risk factors, such as obesity, will contribute to lower morbidity/mortality and a reduction in healthcare costs. In order to answer the questions regarding the cost-effectiveness of breakfast and interventions promoting breakfast eating, we need to know more about how the identified effects influence the societal cost and the individual's QALY gain. There is also a need to develop models that can simulate breakfast and breakfast interventions' long-term costs and effects. In this study, we have started that work by identifying important model parameters.

Conclusions
The overall assessment of the studies indicated positive associations between eating breakfast and study outcomes that measure cognitive performance, academic achievement, quality of life and well-being, and morbidity risk factors in high-income nations. How these outcomes influence societal costs and individuals' QALYs require further research.

Conflict of interest and funding
The present study was sponsored by a grant from Arla Food AB. The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit for publication. The authors had full access to all the data in the study and had final responsibility for the decision to submit for publication. The authors declare no conflicts of interest.