Impacts of Global School Feeding Programmes on Children’s Health and Wellbeing Outcomes: A Scoping Review

Inclusion Criteria

Studies exploring school feeding programmes were included if they met the following outcomes: (1) Children’s physical health – Illness, physical activity, and/or diet dietary intake. (2) Children’s emotional and psychological health– emotion regulation, resilience, self-esteem, and behaviour. (3) Children’s social health – social engagement/participation, building and maintaining relationships. (4) Studies published within the past fifteen years to ensure research is up-to-date and relevant. (5) Primary and secondary school aged children (ages 5 to 16). (6) Studies investigating the implementation of universal feeding programmes and its potential challenges. Both experimental and quasi-experimental study designs including randomised controlled trials, non-randomised controlled trials, before and after studies and interrupted time-series studies. Descriptive observational study designs including case series, individual case reports and descriptive cross-sectional studies were also included. Qualitative studies were also included that focused on children’s self-assessed health and wellbeing and perceptions of school feeding programmes.

Exclusion Criteria

Studies were excluded if they met the following criteria: (1) Children below or above school age (below age 5 or above 18). (2) Studies with a focus on SFP and academic outcomes – performance, attainment, or attendance without a health element. (3) Studies that only looked at parental funded meals/packed lunches. (4) Studies not in English and, (5) Studies published prior to 2009. Additionally, any studies that did not focus on the impacts of school meal provisions on children’s health, unless considering factors influencing uptake, were excluded.

Quality of Included Papers

Using the modified version of the Joanna Briggs Institute (JBI) critical appraisal checklist for text and opinion [34], three papers were categorised as high, and the other was categorised as moderate. The three papers categorised as high [24, 69, 72], used a variety of methodologies or a strong evidence base to support their findings/opinion. The paper rated moderate [70] included less peer reviewed evidence to support their findings but methodological quality was satisfactory.

Characteristics of Included Studies

Characteristics of the included studies are summarised in table 3. 11 studies were cross-sectional [37–47]. Five studies had a difference-in-difference design [20, 49–51]. Three were non-randomised control trials [6, 52, 53]. Four were quantitative descriptive [54–57]. Four were mixed methods [58–61]. Six were qualitative [62, 64–68]. Six were reports and/or opinion pieces [22, 24, 70–72]. One was a rapid narrative review [73] and two were case studies [63, 74].

Twenty-three papers were based on universal school feeding programmes [20, 24, 38, 40, 43, 45, 48, 49, 50, 51, 52, 53, 56, 57, 59, 60, 61, 63, 65, 66, 67, 70, 73]. Fifteen had targeted SFP systems [6, 37, 41, 42, 44, 46, 47, 54, 62, 64, 66, 68, 69, 72, 74]. Four papers were categorised as ‘other’. Papers were categorised as other if they looked at breakfast [55], compared both universal and targeted systems [71], compared packed lunch to school meals [39] or were not specified [22].

Nine papers included children in all school ages (primary and secondary) [24, 43, 44, 46, 56, 57, 59, 68, 70], Twelve papers included children just in primary schools [39, 40, 41, 45, 49, 50, 51, 52, 55, 61, 69, 71], Ten papers included children just in secondary schools [6, 38, 42, 53, 62, 64, 65, 66, 67, 74]. Two papers had an overlap of different age groups (children aged 4-14 [37] and children aged 9-16 [48]). One study looked at the impacts of SFP across the life course (from birth to 50+) [20]. One paper included school staff [60]. Five papers did not explicitly state age groups [22, 54, 63, 65, 72].

Seventeen papers were from UK (England n=11 [39, 40, 42, 45, 49, 50, 58, 65, 68, 70, 71], Scotland n= 2 [63, 74], England and Scotland n=1 [51], UK wide n=3 [44, 46, 69]. Six papers were from the USA [43, 54, 55, 56, 59, 60], three papers were from Sweden [20, 24, 38], four from Norway [6, 53, 64, 65], one from Portugal [62] one was Europe wide [22]. Two were from New Zealand [67, 73], two from Brazil [37, 41], two from South Korea [47, 48], two from Africa (Kenya [52] and Nigeria [47], one from Japan [71] and one was cross national [72]. Figure 2 demonstrates the countries of the included papers.

• Download figure
• Open in new tab

Fig 2. Map showing countries included in review.

Overall studies that investigated universal feeding programmes appear to have the most significant positive outcomes on children’s health, specifically in relation to behaviour, health, and socio-economic outcomes. Whilst targeted systems adequately address food insecurity and child hunger, children eligible for targeted provision had a higher probability of experiencing poor mental health and stigma. Findings are discussed in more detail below.

Physical Health: BMI and Body Weight Outcomes

Five papers investigated the impacts of SFP on children’s BMI and weight outcomes. Two with targeted provision [6, 37] and three with a universal system [48, 50, 56]. Studies had contradicting results.

Batista et al, [37], investigated the prevention of overweight children in Brazil through targeted SFP. The study which utilised a cross-sectional design and included 7,017 participants from 21 schools found that being overweight was prevalent in 30.6% of children and being underweight was present in 1.9%. This study highlighted that a large proportion of the foods served were ultra-processed, recommending a re-evaluation of food standards within schools. Researchers noted an inability to observe extracurricular activities relating to food consumption and physical activity which may also influence children being overweight.

Bethman and Cho [48], investigated the impacts of reintroducing universal free school meals (UFSM) in South Korea after the policy had been abolished a year prior. Findings indicated the removal of the policy had adverse effects on children’s BMI outcomes. The reintroduction of the policy on the other hand reversed these impacts and saw more children with healthier weights. However, the sample size of the study is not explicitly stated and may not represent the intended population.

Holford and Rabe [50], suggested that longer exposure to a universal infant SFP positively influenced children’s weight outcomes. The study included 153,522 primary school children aged between four and seven and used a difference-indifference analysis to measure children’s weight outcomes over time. Findings concluded that children are more likely to be a healthy weight after exposure to the provision over the course of the first year of school. Though participants were very young, it is recommended that participants are tracked over a longer time period to see if this is a lasting effect.

A study by Vik et al, [6] conducted in a secondary school in Norway, investigated changes in children’s dietary habits, BMI and waist circumference modulated by socio economic status. The study conducted a non-randomised control trial consisting of two groups, non-exposure to SFP and exposure to SFP under a targeted system. Findings suggested free school meals (FSM) increased the intake of healthy foods, particularly among children with lower socio-economic status, leading to an increase in the Healthy Food score of the intervention group between baseline and follow-up. However, children in the intervention group had a significant increase in BMI, though it is not explicitly stated if the children were underweighting prior to the trial. The study concluded that interventions to promote healthy eating are more effective in lower socio-economic groups if it is free or at reduced price.

Davis et al. [56] saw similar results when they examined the effects of a universal SFP in a deprivation area in the USA. Reporting that BMI increased in provision schools versus non provision schools. They further stated that children in provision schools had a higher BMI on average.

In conclusion, there is no evidence that SFP lower BMI, but children are more likely to be a healthy weight. Factors such as the prevalence of ultra-processed foods, socio-economic status, and policy reintroduction played significant roles in these outcomes. Longer exposure to SFP, particularly universal programs, was associated with healthier weight outcomes in younger children, though lasting effects require further study.

Nutritional Intake

Ten papers [24, 53, 51, 39, 45, 61, 52, 47, 38] explored the impacts of SFP on nutritional intake. Outcomes explored were diverse. An Article piece [24] claimed that Sweden was one of the countries with the highest vegetable intake, with a substantial intake at lunch. Sweden has the longest running universal SFP in the world and supplies food to all school aged children. The article notes that school meal provision in Sweden is seen as a universal welfare service and a part of public health work.

Vik et al. [53] used a non-randomised control trial including 164 secondary school children to investigate whether serving a free healthy school meal for one year resulted in a higher intake of fruit and vegetables and a lower intake of unhealthy snacks. Vik et al. [53] reported that a free healthy school meal for one year was associated with higher weekly intake of vegetables on sandwiches in the intervention group compared to the control group. However, it was not associated with a lower weekly intake of unhealthy snacks in the intervention group compared to the control group. It is important to note that the study had a small sample size and relied on self-reported data.

Parnham et al. [51] examined the impact of a universal infant SFP policy on dietary intake in primary school children in England and Scotland. The implementation of the policy led to more children participating in school meals. No impact was seen on fruit and vegetable intake but there was evidence that the policy lowered consumption of foods associated with packed lunches, such as crisps, and some nutrients, such as total fat and sodium.

A further study by Evans [39] also reported that school meals were a healthier option than packed lunches. The study utilised a cross-sectional design and included 2709 primary school children in England. Findings suggested children having a packed lunch consumed on average 11g more total sugars than those who had a school meal whereas, children having a school meal were more likely to consume different types of vegetables and drink water. Overall, children with a packed lunch consumed a lower-quality diet over the whole school day. The study did not look directly at school meal provision and so it is difficult to note whether those on provision would yield differing results.

Spence et al. [45] explored the effect of pre– and post-Universal Infant Free School Meals (UIFSM) and school on pupil’s dietary intakes using a cross-sectional research design. At lunchtime, there was a statistically significant decrease in students’ non-milk extrinsic sugars intake pre implementation and increases of intakes of cake’s pudding post implementation. However, only two schools were included, thus limiting the generalisability of results. It is recommended that schools consider healthier policies when implementing SFP.

Yamaguchi et al. [61] indicated that universal SFP were beneficial to individuals with low socio-economic backgrounds. Their study which used a mixed methods approach and included 719 primary school children in Japan, saw that children whose mothers were less educated had greater reliance on school lunch for their vegetable intake. Children with lower household income had more contribution from school lunch to their fruit intake. Household income was not explicitly considered in the study and mothers’ education was used to determine socio-economic status, so it is unclear whether families were from lower socio-economic backgrounds. Further studies are needed to conclude the correlation between vegetable intake and household income in Japanese children.

Four papers, Neervoort et al. [52], Zailani et al. [47], Colombo et al. [38] and Goel [55] considered the impacts of SFP on vitamin intake as well as nutritional deficiencies.

Neervoort et al. [52] conducted a non-randomised control trial including sixty-seven primary school children to investigate the impacts of a universal SFP on anaemia, stunting, wasting and malnutrition. Findings indicated the programme reduced anaemia and malnutrition, and improved child growth in the study group. Whilst improvements were seen, the study had several limitations. Such as a very small sample size and participants were from only one school. The study also lacked information on the content and nutritional value of provided meals, which could influence program effectiveness.

Zailania et al. [47] investigated portion sizes of targeted school meals as well as nutrient intake and found that the meals served through the SFP contributed at least 33% for energy, protein, iron, calcium, sodium, vitamin A, and zinc intake. However, there was no consumption of fruits, meat, poultry, and fish within the programme. The study did not consider the contribution of other food sources, including snacks and homemade meals which may account for nutritional values.

Colombo et al. [38] used a cross-sectional design to explore School Lunch Dietary intake under a universal SFP in Sweden. Results found that a quarter of the overall energy intake; between 22% and 30 % of selected nutrient intakes; almost half of vegetable intakes; roughly two-thirds of fish intakes; and around a third of red/processed meat intakes. These findings imply that school meals make an important contribution to children’s diets on weekdays. Researchers also reported school meals as more nutritious than meals consumed outside of school, though meals consumed outside of school meals were not measured.

Goel et al. [55] examined total sugar in free breakfasts served in elementary (primary) schools in Virginia, USA through an observational study. Findings suggest that meals offered might contribute to excessive overall sugar availability for children. Nutritional information for added sugars was not provided within the study, however, the findings may explain contradictions in BMI above. Goel et al. [55] states districts and policymakers should collaborate to implement more effective guidelines concerning sugar availability in breakfast items to optimise children’s dietary intake.

In summary, the evidence for improved nutritional intake is limited. However, school meals are an important contribution to diets during weekdays. Consuming school meals alleviated dietary discrepancies related to social inequalities.

Psychological and Emotional Health Outcomes

The impacts of SFP on children’s psychological and emotional outcomes were investigated in four papers [57, 60, 48, 46].

A randomised control trial by Altindag et al. [57] evaluated the effects of targeted school meals on students’ behaviour, bullying, and violence in schools. Behavioural incidents reduced by 35% since the implementation of a universal system in 2010 (particularly physical fights among students). It is believed that this reduction is the result of reduced stigma among students as socio-economic status cannot be identified via the system. Provision did not impact students’ nutritional intake but reduced the stigma of receiving free meals through a targeted system by mitigating the possibility of identifying peers’ socioeconomic statuses.

Taylor et al. [60] included 116 school staff in the USA to explore the impact of a universal SFP on school climate, behaviour, and attainment. Taking a mixed-methods approach, over 725 of school staff surveyed reported that serving universal school meals has improved social climate as well as reductions in student stress, family financial stress and school administrator stress. This is thought to be a result of income differences being less visible, and the school community feeling more inclusive. It is recommended that further research include interviews with students and families to gain a better understanding.

As previously mentioned Bethman and Cho [48] saw positive results when looking at children’s BMI. Similarly, the reintroduction of a universal policy also saw mental health improvements among students. Proclaiming that free school lunches help to improve health and benefit student welfare. However, mental health was assessed using a ‘crying without any reason’ measure. The measure used is not reported so validity and reliability cannot be determined. No other Mental Health outcomes were assessed.

A further paper by Yang et al. [46] assessed targeted school meals in the UK on children’s mental health outcomes. The study used a cross-sectional design and included 2166 primary and secondary children. Yang et al [46] reported that poor mental health was observed in food insecure children receiving provision but also food insecure children not receiving provision. Interestingly, food insecure (i.e., not having access to sufficient food) children receiving provision had a higher probability of poor mental health than those who were food insecure and not receiving provision. It is thought that this is because targeted approaches are accosted with stigma.

In conclusion, universal approaches were associated with improved mental health outcomes, such as reduced behavioural incidents in school and an improved social climate. This was not observed with targeted approaches.

Social Health Outcomes

Five papers [58, 65, 64, 66, 68] investigated the social impacts of SFP. The impacts of SFP are multifaceted with outcomes seen with social interactions, reduced stigma and narrowing inequalities. Findings are discussed further below.

A mixed method, quasi-experimental evaluation exploring the impact of universal provisions in two London secondary schools by Carlisle et al [58] saw improved social skills as children were eating with peers. Improved behaviours, less stigma as well as students eating more varied foods. These findings were as a result of paired student interviews, this may have caused socially desirable responses.

A further study by Jessiman et al. [65] which also included a sample of secondary school students in London also reported social health benefits, Students perceived feeling equal under a universal system whereas they reported feeling ‘weak’ under the targeted system.

Illøkken et al. [64] conducted a qualitative study in a Norwegian secondary school, involving 13 students and five teachers to investigate the effects of a targeted school feeding program on social inequalities. Participants viewed school meals as a social event where students made new friends and learned new skills. It was also reported that social equality among students increased. However, it is unclear whether this was the result of the feeding programme or school meals in general.

Another qualitative study conducted in a Norwegian secondary school by Mauer et al. [66], also reported positive social outcomes. Students expressed that social time while eating school meals was important to them. This study also noted that popularity of the food was also important for attracting students to school meals.

Sahota et al. [68] investigated factors influencing take-up of targeted school meals through focus groups with children in both primary and secondary schools in England. Utilising focus groups to gain children’s perspectives. Food choice, queuing, and the social aspects of lunch time, such as eating with friends was a major influence in uptake. Recognition of the importance of the social aspects of dining for pupils and facilitation of social interactions through the spatial (including flexible locations, e.g. outside) and temporal organisation of lunchtimes. It is reported that the schools involved in the study had a high level of targeted school meal entitlement which may have resulted in the normalisation of school meal uptake. It is recommended that social aspects of school meal provision be further investigated in schools with lower levels of eligibility to see if results differ.

Reducing Food Insecurity and Child Hunger

Three papers [73, 59, 41] looked at the impacts of SFP on food insecurity and child hunger. Most papers in this review touched on food insecurity, however, these three papers specifically looked at impacts of food insecurity outcome. A rapid narrative review by Garton et al. [73] concluded that SFP in New Zealand significantly reduced hunger and food insecurity in primary schools. Hecht [59] suggested that access to free meals through a universal provision in a high deprivation area in the USA, reduced food insecurity among children as well as decreased child hunger. Lastly, a cross-sectional analysis by Horta [40], investigated impacts of SFP on Vulnerability risk. Findings saw positive impacts from consuming school meals on children’s diets, particularly among children living in high/very high social vulnerability risk areas.

Factors Influencing Uptake

Ten papers explicitly investigated SFP participation and factors associated with non-take-up among students. Two studies gained perspectives through qualitative methods [67, 62]. A paper by McKelvie-Sebileau et al. [67], involving universal provision, found that lack of knowledge of the programme and loss of agency over meal choices were major drivers of non-take-up but, positive influences such as participants’ food security, better nutritional knowledge and improved wellbeing were also perceived. Cardosa et al. [62] who evaluated a targeted approach in Portugal noted that quality of food was a concern when participating in school meals. Children who were not entitled to free or discounted meals reported eating at school less often than children with free or discounted lunches.

Three reports, Kitchen et al. [70], Rahim et al. [71] and Harper and Wood [69], all reported increased take-up under a universal policy, for both previously non eligible children and eligible children. It is suggested that improvements were seen as a result of reduced stigma and familiarising parents with school meals. Participants believed the pilot increased the range of food that pupils would eat, built their social skills at mealtimes and, for some pupils, resulted in health benefits associated with having a balanced meal, such as more energy, concentration and alertness and improved complexion. take-up was higher among primary school children than secondary school children. However, Rahim et al. [71] in particular, found schools experienced difficulty predicting the appropriate quantities of food required for each menu option, which meant that some meals either ran out early or were wasted.

A study by Holford and Rabe [49] also found that take-up of school meals in England by non-eligible children rose from a consistent 30-35% in the eight years preceding the policy to approximately 85% in the universal period.

Additionally, two case studies [74, 63], one exploring a targeted system and the other a universal system saw similar results. Under the targeted provision [74] stigma, food quality and social aspects played a vital role in take-up. Results specifically noted that, being separated from friends decreased take-up. Implementation of a universal system [63] however, yielded more positive results such as, families who were previously above the eligibility threshold now being able to access provision. It was also documented that school-provided meals were of a higher nutritional quality than those from home, exposing children to a greater variety of foods and establishing healthier eating habits.

Finally, a cross-national comparison by Oostindjer [72] indicated that removing popular foods such as meat from school menus significantly reduced take-up of school meals. This was particularly evident in Finland where meat and fish were removed from the school menu on some days reducing participation rate, consequently producing up to 60% plate waste. Please note that this paper did not explicitly mention SFP but instead, school menu choices and take-up therefore, more research is needed on food choices under SFP to evaluate if there is an effect on take-up.

Wider Impacts

The remaining seven papers did not fit into a thematic category and will therefore be summarised individually below.

A UK wide study by Parnham et al. [44] exploring access to targeted SFP, reported that receiving an FSM was associated with increased odds of recently using a food bank [75] but not reporting feeling hungry. In the month after the COVID-19 lockdown, 49% of eligible children did not receive any form of FSM. In this study the sample size was small, and this may lead to chance findings or in ability to detect differences.

A policy recommendation report by Guio [22], explored evidence of the short and longer-term benefits of school meal provisions in EU countries. The report found that health outcomes are dependent on the programme take-up and quality of food. Additionally, while eating formally, pupils learn to be sociable and develop interaction skills. There are some findings on health outcomes associated with school meal provision, but these are more difficult to establish. School-based food and nutrition interventions were able to improve dietary behaviour, healthy eating, and anthropometry, but the design of the intervention affects the magnitude of the effect.

James [42], used a cross-sectional design to investigate stigma associated with targeted provision and how this influences the peer effect. The study included 21,000 secondary school students in England. According to James [42], the presence of stigma dampens the peer effect and so students may not be benefiting from provision as much as possible. However, information about the provision has the opposite impact on the peer effect. These findings suggest that information is a more important part of the peer effect for those living in areas of greater deprivation and stigma is more important for those in the least deprived regions.

Long et al. [43] assessed the impacts of a universal system on meal cost and food quality in 508 US schools. This study finds that participation in the provision was associated with lower per-meal full cost with no differences in dietary quality. This indicates that in the US, universal SFP can provide nutritious meals to more students without a financial disadvantage. Meier et al. [54] conducted a quantitative exploratory study including 576 parents. Completed in the USA, the study aimed to gain parents overall perceptions of the targeted school meals program. This study found that parents of enrolled children were more likely to report positive perceptions of the school meals’ program. Whereas parents of children not receiving the program were less likely to perceive the school meals program in a positive light. Two limitations were noted that may have influenced results. Firstly, Parents of children not receiving provision made up much of the sample and second, parents volunteered to participate in this survey, therefore participation bias could have resulted from those who are more involved or have strong opinions about school lunch.

A longitudinal study, utilising a difference-in-difference design by Lundborg et al. [20] evaluated the long-term impacts of the universal program on children’s economic, educational, and health outcomes throughout the life course. The study involved 1,529,760 participants from birth to age 50+. The study noted several findings, for instance, children exposed to the program during their entire primary school period have 3% higher lifetime income. This effect was greater for pupils that were exposed at earlier ages and for pupils from poor households, suggesting that the program reduced socioeconomic inequalities in adulthood. Additionally, a year of exposure to UFSM increased height growth and nine years of school lunch exposure increase the likelihood of being of near perfect health. Exposure to school lunches also decreases the probability of being diagnosed with any health condition.

The final study by Goodchild et al. [40] investigated factors associated with a universal infant free school meal take-up and refusal in a multicultural urban community involving 676 parents of nursery aged children (4-7 years). The study explored two groups; non-provision, n = 159 or took provision, n = 517. The non-provision group were more likely to be White-British, have higher socio-economic status, have English as a first language, and involve their child in the decision over whether to take the school meal, compared to the provision group. It should be noted that the area the study was conducted is multicultural and school meals must cater for children from a variety of cultural backgrounds. Parents who did not complete questionnaires correctly were more likely to have lower socio-economic status, be non-white British and have English as an additional language, meaning some groups were under-represented.