**Salt Reduction Strategies in Portuguese School Meals, from Pre-School to Secondary Education—The Eat Mediterranean Program**

**Ana Isabel Rito 1,2,\* , Sofia Mendes 2,3, Mariana Santos 1,3 , Francisco Goiana-da-Silva <sup>4</sup> , Francesco Paolo Cappuccio <sup>5</sup> , Stephen Whiting <sup>6</sup> , Ana Dinis <sup>7</sup> , Carla Rascôa 7 , Isabel Castanheira <sup>1</sup> , Ara Darzi <sup>4</sup> and João Breda <sup>6</sup>**


Received: 23 June 2020; Accepted: 16 July 2020; Published: 24 July 2020

**Abstract:** High sodium (salt) consumption is associated with an increased risk of developing non-communicable diseases. However, in most European countries, Portugal included, sodium intake is still high. This study aimed to assess the sodium content of school meals before and after the Eat Mediterranean (EM) intervention—a community-based program to identify and correct nutritional deviations through the implementation of new school menus and through schools' food handlers training. EM (2015–2017) was developed in 25 schools (pre to secondary education) of two Portuguese Municipalities, reaching students aged 3–21 years old. Samples of the complete meals (soup + main course + bread) from all schools were collected, and nutritional quality and laboratory analysis were performed to determine their nutritional composition, including sodium content. Overall, there was a significant decrease (−23%) in the mean sodium content of the complete school meals, which was mainly achieved by the significant reduction of 34% of sodium content per serving portion of soup. In conclusion, EM had a positive effect on the improvement of the school meals' sodium content, among the participant schools. Furthermore, school setting might be ideal for nutrition literacy interventions among children, for flavors shaping, and for educating towards less salty food acceptance.

**Keywords:** community-based program; childhood obesity; school meals; salt intake; sodium consumption

#### **1. Introduction**

Given the well-established evidence that excessive sodium consumption (1 g of sodium per 100 g represents 2.5 g of salt per 100 g) is linked to an increased risk of developing non-communicable diseases (NCD) [1], a reduction in population's sodium consumption has been a key focus of both the international and national policy agendas. Reducing salt intake in the general population is not only a practical action that can prevent adverse health outcomes—such as increased blood pressure—but it is also a feasible and cost-effective strategy to reduce the growing burden of NCDs and reduce health-care costs for governments and individuals [2].

The World Health Organization (WHO) recommends a population reduction in salt intake as one of the 'best buys' or cost-effective actions that should be prioritized to tackle the global burden of NCDs [3–5]. Targets of a daily salt intake lower than 5 g for adults and 2 g for children have been recommended [6]. In addition, WHO Member States have agreed to work towards the global target of a 30% relative reduction in mean population intake of salt by 2025 relative to 2010 levels. It is crucial that this target is met in order to achieve the overall goal of a 25% reduction in premature mortality from NCDs by 2025 [2].

The overall number of countries implementing a national salt reduction strategy more than doubled from 2010 to 2015. However, despite the remarkable efforts and actions that have since been taken, more needs to be done. Data from 2013 revealed that population salt consumption in most European countries ranged from around 7 g/day (Bulgaria, Cyprus, Germany, and Latvia) to 13 g/day (Czech Republic) [7,8].

Among all dietary habits, excessive salt intake has the most adverse outcomes. The average daily intake of salt per capita among the Portuguese population is 10.7 g [9], which is double the level recommended by WHO (<5 g) [6]. Portugal ranks the highest among European countries regarding salt intake, with excessive intake reported in 63.2% of women and 88.9% of men [10]. The problem also affects younger groups, as research shows that most children and adolescents exceed daily recommendations [11–13].

Excessive salt intake is associated with an increased risk of obesity—partially due to poor diets that are high in both energy and salt, such as regular consumption of breakfast cereals [14] and highly processed foods [10]. Another reason for this association may be that consumption of salty foods stimulates thirst and increases fluid intake, thereby increasing the consumption of sugar-sweetened beverages, which can further fuel obesity [15]. This scenario is of particular interest in Portugal, where the prevalence of childhood overweight has been among the highest in Europe, affecting around one in three children [16]. It is, therefore, urgent to tackle this issue, as even small reductions in salt consumption can bring great health benefits to children by reducing the risk of developing cardiovascular diseases—the leading cause of death and disability in Portugal and worldwide [17,18].

The Portuguese "National Program for the Promotion of Healthy Eating" [19], in line with internationally recommended interventions [20], strongly advocates for the implementation of strategies to reduce dietary salt intake in children by providing information and education on healthy eating as well as the strengthening of consumer protections, particularly by reducing the salt content of school meals.

Several attempts have been made to reach children in schools to encourage healthier eating habits and improve the nutritional quality of the food served to them. These interventions can potentially impact all children of school age, irrespective of their ethnicity or socioeconomic group [21–25]. Primary and secondary schools serve at least one meal every day and can also determine the types of food and beverages that are available or served at schools (i.e., schools' cafeterias and vending machines). Schools can positively impact eating behaviors and promote healthier eating [26–28], for example, by deploying nutrition education classes.

School is a key setting to deliver health education to children, promote healthy lifestyles and social equality, and to ensure access to nutritionally balanced meals, regardless of the family's socioeconomic status [29]. In Portugal, municipalities are responsible for providing school meals (lunch) for pre-school and primary schools as well as for the management [30] of the menu. During secondary education, the supply of school meals is supported by the Directorate General of Education Institutions (DGEstE) [31], except for schools with their own cooking facilities. In Portugal, a set of guidelines for the school food supply has been established, which includes limits on the salt content of the school meal's components—bread, soup, and the main dish [31].

Assuming that lunch represents 30% of the total energy value [32] and considering the WHO recommendation [6], 1.5 g of salt should be the maximum level in this meal. In Portugal, little is known about the nutritional composition of the complete school meal. The amount of different nutrients in food samples can be measured through laboratory analyses, using standardized techniques recommended by international organizations [33]. In the few studies that have been conducted to estimate sodium content of school meals in Portugal, mean salt content has ranged from 2.83 and 3.82 g [34–36], which clearly should be reduced.

Eat Mediterranean—A Program for Eliminating Dietary Inequalities in Schools (EM) [37], was a European Economic Area (EEA) Grant funded project developed as a Portuguese community-based intervention (2015 to 2017) through a multi-sectorial approach involving health, education, and political stakeholders. The program's goal was to reduce nutritional inequalities among school-aged children through the promotion of the Mediterranean diet. The program comprised a comprehensive approach both at the individual level (child and family) and at the group/community level (nutritional education sessions at schools and improvement of school food environments). One of the objectives and key priority areas in the implementation of EM at the community level was to evaluate and improve the nutritional quality of food available in school meals. Within the school food environment, the EM program proposed a qualitative and quantitative (laboratory) analysis and evaluation of the nutritional adequacy of school meals. The aim was to identify nutritional deviations, according to international/national recommendations [29,31,38,39] and correct them by modifying the food composition of school meals through both training of the schools' food handlers and through the development and implementation of new menus.

#### **2. Materials and Methods**

#### *2.1. Program and Participant Schools*

The EM program was implemented over two school years (2015/2016 (Y1) to 2016/2017 (Y2)) in two Portuguese municipalities: Santarém and Alpiarça. In total, 25 individual public schools and 5773 students (3–21 years old), from pre-school to secondary education, participated in EM.

The entities responsible for the supply of school meals in both pre-schools and primary schools were Santarém Municipality (17 schools) and Alpiarça Municipality (three schools). For secondary schools, DGEstE supplied meals to four schools, while one had their own cooking service.

The specific evaluation and intervention on nutritional adequacy of served school meals (lunch) were organized in three phases:


sheets of the new menus were developed, and their implementation was conducted under the supervision of members of the working group.

• Post-intervention (Y2): A new set of 39 school meal samples was collected from the same kitchens from April to June 2017, and a qualitative and quantitative evaluation of the changes was performed.

Ethical approval was granted by Lisbon and Tagus Valley Regional Health Administration Ethical Committee (089.CES/INV/2015).

#### *2.2. Food Samples and Sample Preparation*

Food samples were collected from all 10 kitchens that served meals to the 25 schools. Of the 10 kitchens, nine served meals at their own schools, so samples were collected at the moment of serving. One school was served by transporting meals from a central kitchen outside the city. In this case, food samples were collected at the school immediately prior to serving.

The samples consisted of the food portions that were served to children at lunchtime. Each food sample consisted of three main items: bread, soup, and the main course (including salad or cooked vegetables and one piece of fruit). These were collected during both evaluation and post-intervention phases in a total of 39 samples in each phase. In one of the schools (school B), it was not possible to analyze the bread samples, as they were not sent to the laboratory. The meal items were weighed on a Mettler-Toledo PB3002-S/FACT (Mettler-Toledo, Inc., Columbus, OH) laboratory scale, with an accuracy of 0.01 g. Samples were collected using latex gloves, placed in sterile polythene bags, and alphabetically coded to maintain confidentiality. The samples were transported to the laboratory, refrigerated, homogenized, and milled using a high-speed grinder, a knife mill Grindomix GM 200; Retsch, Haan, Germany equipped with titanium knives to prevent contamination. The prepared samples were stored in vacuum bags at the freezing temperature (−20 ◦C) until processing.

#### *2.3. Laboratory Analysis and Interpretation*

The analysis was performed in accordance with the methodology recommended by the Official Methods of Analysis of AOAC International [33], under quality assurance conditions complying with the requirements described in standard EN ISO/IEC 17025: 2005 [41]. For sodium determination, the samples were analyzed in triplicate using an inductively-coupled plasma optical emission spectrometer, ICP OES, model iCAP 6000, Thermo Fisher Scientific, Madison, WI, USA for the determination of sodium (Na) content. There are several common sources of sodium in food, including from salt added during preparation or during processing, as well as from the sodium in seasoning (e.g., sodium phosphate, sodium bicarbonate, MSG mono-glutamate, etc.). However, this study assumed that all sodium in food was in the form of sodium chloride and equivalents, so all results were expressed in terms of "salt".

The salt content in g/100 g of food was calculated by the formula: salt (g) = sodium (g) × 2.5 [41]. Considering a school meal (lunch) makes up 30% of the daily total energy intake [32], 1.5 g of salt was the reference value used in the present study (according to the WHO recommendation of salt intake [6]: 0.30 × 5 = 1.5 g).

#### *2.4. Statistical Analysis*

Data sets were produced using Microsoft Excel® spreadsheets, and statistical analyses were performed using IBM SPSS® statistics for Windows, version 22.0, Armonk, NY, USA [42]. Results were reported as mean (+ standard deviation). Non-parametric tests for comparing means were carried out for paired samples. A significance level of α = 0.05 was considered statistically significant.

#### **3. Results**

The quantitative analysis of the school menus found that the standardized serving portions collected during the evaluation phase and the post-intervention phase were similar. Regarding the reduction in sodium and salt equivalent of the individual meal components, there was a 34% reduction per serving portion of soup. There were no significant changes in sodium and salt equivalent per serving portion of bread or per serving portion of the main course. In the complete meal, including the three components, there was a 23% reduction in sodium and salt equivalent per serving portion (Table 1).

**Table 1.** Sodium and salt content of school meals components (soup, main course, and bread) and of the complete meal (all components) analyzed at the evaluation and the post-intervention phases of the eat Mediterranean program.


(a) There were no statistically significant differences between serving portions (g) (*p* > 0.05); (b) The salt content was calculated by the formula: salt (g) = sodium (g) × 2.5 [35].

Changes in the mean salt content of the complete school meal in grams (g) at evaluation and at the post-intervention phase are shown, for individual schools and for all schools combined, in Figure 1. For all schools except for two (B and J), there was a decrease in salt content between the two time-points.

**Figure 1.** Mean salt content (g) of the complete school meals analyzed at the evaluation phase and the post-intervention phase of the eat Mediterranean program and its adequacy regarding the reference value (maximum 1.5 g of salt/meal), by the school.

#### **4. Discussion**

While community-based programs designed to improve the nutritional quality of school meals have been shown to be effective previously [43,44], EM was one of the first programs in Portugal to address qualitative and nutritional laboratory analysis together. Through a multidisciplinary approach targeting the school food environment, a key objective of the program was to improve the nutritional composition of school meals served to young people during lunchtime. In Portugal, addressing the quality of school meals is an important way to promote healthy diets as at least one meal is offered every day, and at pre- and primary education levels, almost every child has lunch at school [45].

The qualitative assessment of the 386 school menus has been presented elsewhere [46]. As part of evaluating EM, this study focused on identifying the nutritional deviations of sodium and salt equivalent content of school meals from international and national recommendations [29,31,38,39] and aimed to correct them by modifying the nutritional composition of school meals. This was done through training of the schools' food handlers and the development and implementation of new menus. The results showed that EM had a positive effect on the improvement of the school meals' salt content among the participating schools, achieving an overall reduction of 23% of the salt content of school meals served at lunchtime.

At the beginning of the EM program, the mean salt content of school meals was 3.75 g of salt per meal. These findings were similar to those reported in previous Portuguese studies [34–36], as well as in studies from other countries that assessed the salt content in school meals served in canteens [47]. Interventions as part of the EM program led to a significant reduction (*p* < 0.05) of salt content (from 3.75 g to 2.90 g of salt per meal, i.e., ~23%); however, it was still far from the reference value of lunch salt content (1.5 g of salt), and it was estimated that it would need to be met to achieve recommended salt consumption levels.

Looking separately at each component of the meal, the main dish was the component with the highest contribution to the salt content of the whole meal. This was also found in the study conducted by Barbosa et al. 2018 [48] in Portuguese University Canteens, in which it was suggested that one possible explanation for this result was the presence of intrinsic sodium in foods, such as meat and fish, which is higher than the sodium intrinsically present in vegetables used for soups [49]. However, we found a significant reduction in the salt content of soup from 1.48 g per serving portion before the intervention to 0.98 g per serving portion after the intervention (~34% reduction).

The values for lunch salt content, after EM intervention, are yet slightly higher than those reported by other Portuguese studies [50,51]. According to the Portuguese 2018 guidelines for menus and school canteens [31], the maximum value that can be added to soups and main dishes during the cooking process is 0.2 g of iodized salt. In addition, it is recommended that salt be replaced by glasswort or aromatic herbs. Regarding the serving of bread included in the school meals, according to the recommendations for Portuguese school meals [31], it should be one small piece of bread of 25 g for pre-school and primary school and 45 g for elementary and secondary school, with a maximum salt composition of 1%, meaning 0.25 g and 0.45 g of salt per serving of bread, respectively. As there was no intervention targeted at reducing bread provided during school meals, this study observed that, despite the level of education, the mean serving of bread was around 45 g, and the salt content of bread per serving, both before and after the intervention, was above the guidelines (0.46 g–0.49 g). In Portugal, there is a culture of always serving bread at mealtime, which is reflected in the official guidelines [31]. It could be suggested that if there was a non-mandatory offer of bread at school meals, at least for young children, and if carbohydrate intake recommendations were met through foods with less salt, a further reduction on overall salt intake could have been observed through this action alone.

There were several challenges in the implementation of the EM program. One of these was to reduce the amount of salt added by cooks while preparing the meals, as while the technical guidance clearly requires that the amount of salt added during meal preparation be accurately measured, several cooks still opted to measure by "hand". This was also described by Gonçalves et al. [52], who found that the amount of added salt was influenced by the taste of the cook, even though many food handlers acknowledged that they did not taste the food before adding salt. That study also pointed out that food handlers were aware of the health problems associated with excessive salt intake as well as the recommended salt intake values, but they mentioned that the greatest difficulty in salt reduction was the opinion and acceptance of the consumers toward less salt in foods [52]. Such limitations reinforce the importance of educating both consumers and food handlers so that programs aiming at reducing the salt content in school meals and other settings can be more effective.

Action to reduce salt consumption is urgent, including among younger populations, in order to reduce the risk of developing cardiovascular diseases. The offer of high sodium meals in a school environment can contribute to individuals acquiring long-term poor eating habits, including increased consumption of processed food, which is already a pattern in Portuguese children [10]. Additionally, emerging evidence suggests dietary sodium intake may be associated with obesity, both through pathophysiological mechanisms and through the induction of thirst and increased consumption of high energy drinks [53–57], enhancing the need to address and tackle this public health issue.

The improvement of the menus introduced by the EM program has shown that it is possible to successfully reduce the salt content of school lunches through existing mechanisms. To ensure school meals are nutritionally adequate, it is essential that trained cooking staff and all responsible parties strictly comply with the technical sheets provided. It is also crucial to continue the education of students, parents, educators, teachers, as well as the monitoring of all stages of preparing and serving meals by the cooking staff. The integrated and concerted work among health departments, research institutions, municipalities, and educational communities was a strong part of the success of the EM program.

Among the common challenges educators face when trying to reduce the amount of added salt to meals is rejection by consumers due to "lack of flavor" [52]. However, the adaptive capacity of the flavors-linked neurological system to small reductions in salt in meals has been well described [57]. Thus, school settings may be ideal not only for nutrition literacy interventions among children but also for flavors, shaping and educating towards less salty food acceptance.

One of the pitfalls of this study was the limited time frame of the intervention to carry out all the activities projected in this comprehensive program without including a post-intervention longer monitoring period. This would have been important to continuously assess the adaptation to the changes implemented in the context of the new school food environment, in particular, the acceptance/preference of less salty meals by the children.

Nonetheless, recognizing the relevance of consumer acceptance in order to obtain long-lasting changes, EM paved the way for further work towards providing healthier meals in the participating schools, including a monitoring system to assess students' acceptance of school meals changes and also regarding food waste. The training, capacity building, and nutritional education offered during the EM intervention to all school community (teachers, parents, children, food handlers, and others) would hopefully contribute to the sustainability of the progress achieved in improvements and support continuous improvements.

#### **5. Conclusions**

This study demonstrated the success of the EM program in reducing the salt content of lunch meals served in schools. School meals must be nutritious, and reinforcement of this through regular monitoring and evaluation is a key factor to ensure school food quality. In order for school meals to be nutritionally adequate, trained cooking staff and all responsible parties would need to strictly comply with the provided technical sheets. It is crucial that the health literacy of students, parents, educators, and teachers is developed through continuous education, and the monitoring process at all stages of preparing and serving meals by school food handlers is strengthened. This comprehensive program was built through a collaboration between different stakeholders (health departments and units, research institutions, municipalities, and educational communities), which was both key to its success and ensured a holistic approach towards promoting healthier behaviors.

**Author Contributions:** The authors' contributions are as follows: A.I.R., was part of the coordinating team of eat Mediterranean and worked in all stages of the research, such as project design, methodology, analysis and interpretation of the data, and conceptualization, preparation and writing of the original manuscript; S.M. worked on the analysis and interpretation of the data; M.S. contributed to the laboratory analysis; A.D. (Ana Dinis) and C.R. coordinated and implemented the eat Mediterranean program, worked on the project design, data collection, and interpretation of the data. A.I.R.; S.M.; M.S.; F.G.-d.-S.; F.P.C.; S.W.; A.D. (Ana Dinis); C.R.; I.C.; A.D. (Ara Darzi); J.B. contributed to the writing, review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** The eat Mediterranean program (2015–2017) was coordinated by ARSLVT, Portugal, co-funded by the Public Health Initiatives Program (PT06) of the EEA Grants (grant application 171 NU2).

**Acknowledgments:** The authors wish to acknowledge all the nutritionists, psychologists, and other health professionals, children, parents, educators, teachers, school cooking staff, and municipality technicians for their contribution on the fieldwork, as well as the following Institutions for their partnership and support: Agrupamentos de Escolas Ginestal Machado, Sá da Bandeira e de José Relvas; Hospital Distrital de Santarém; CEIDSS—Centre for Studies and Research in Social Dynamics and Health; ISCTEIUL—Instituto Universitário de Lisboa; Municipalities of Alpiarça and Santarém and Instituto Nacional de Saúde Doutor Ricardo Jorge.

**Conflicts of Interest:** The authors declare no conflict of interest. The writing group takes sole responsibility for the content of this article, and the content of this article reflects the views of the authors only. S.W. and J.B. are staff members of the WHO. The WHO is not liable for any use that may be made of the information contained therein.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Is the Perceived Fruit Accessibility Related to Fruit Intakes and Prevalence of Overweight in Disadvantaged Youth: A Cross-Sectional Study**

### **Narae Yang and Kirang Kim \***

Department of Food Science and Nutrition, College of Natural Sciences, Dankook University, Dandae-ro,

Dongnam-gu, Cheonan-si, Chungnam 31116, Korea; skfo2581@daum.net

**\*** Correspondence: kirangkim@dankook.ac.kr; Tel.: +82-41-529-6373

Received: 23 September 2020; Accepted: 19 October 2020; Published: 29 October 2020

**Abstract:** Background: Few investigations have studied the relationship between home and school food environments, fruit intakes, and prevalence of overweight in children and adolescents from disadvantaged backgrounds. This study aimed to determine whether food environments for fruit intake at household and school levels affect fruit intakes and risk of overweight among children and adolescents with low household income. Methods: Students (*n* = 3148) in Seoul, Korea completed questionnaires pertaining to select aspects of their food environments, frequency of fruit intakes, and weight status. Chi-square tests and logistic regressions evaluated associations between the aforementioned variables. Results: Participants consumed fruit an average of 0.77 times per day, though its frequency increased when fruit accessibility was perceived positively. The percentage of overweight participants was 23.5% for boys and 22.8% for girls. Generally, fruit intake frequency was linked to a lower prevalence of overweight. Regular provision of fruit in school lunches was associated with a reduced risk of overweight among elementary school girls (odds ratio (OR): 0.52, 95% confidence interval (CI): 0.30–0.92), and having someone at home to prepare fruit was associated with a reduced risk of overweight in elementary school boys (OR: 0.64, 95% CI: 0.43–0.94) and girls (OR: 0.63, 95% CI: 0.43–0.93). Conclusions: The frequency of fruit intake was low among disadvantaged youth. Increasing access to fruit in their food environments appears to enhance consumption and lower the risk of overweight, especially for elementary school girls.

**Keywords:** food environment; fruit; children; adolescents; obesity; overweight; home; school

#### **1. Introduction**

The epidemic of pediatric overweight and obesity has expanded steadily over the past few decades, currently reaching more than 370 million children worldwide [1]. Consequences of early excess adiposity include a heightened risk of psychological problems such as depression, low self-esteem, and disordered eating in the immediate term, as well as cardiovascular disease and various cancers upon adulthood [2–5]. Adult obesity, implemented in child obesity, makes it difficult to lose weight [6]. Therefore, it is important to prevent obesity in childhood [6]. The prevalence of obesity among children and adolescents has increased in Asia [7]. Although the rising trends of body mass index (BMI) have flattened in high-income countries, the prevalence of obesity is still high in worldwide [7].

Obesity has been explained with reduced intakes of healthy foods and unhealthy eating behaviors [8–11]. Especially fruit, which are rich in water and fiber, to enhance satiety, and low in energy density [12], have been known to prevent obesity, coronary heart disease, stroke, cardiovascular disease, total cancer, and all-cause mortality [13–16]. Obesity prevention interventions that take into account healthy food intake, such as fruit, are needed. For effective obesity intervention, it is important

to identify the relationship between eating healthy foods and obesity and to find the main determinants that affect obesity. Previous intervention methods for eating healthy foods to prevent obesity have focused on individual behavior changes. However, individual level interventions fundamentally could not change the obesogenic environment, which have been reported to have little effect on long-term improvement [17,18]. In recent years, the ecological model has been applied to promote health in the public health sector. The multilevel interventions that take an account for both individual and environmental factors have been paid attention for effective behavior change. Among environmental factors, food environments include political, economic, social, physical, and natural environment factors which affect accessibility, availability, and affordability of foods [19,20]. The food environments have been known as an important factor related to food choice and dietary intakes [19,20]. For children and adolescents, the household is the first physical and social environment to learn about food intake [21], and the school is responsible for their lunches and another setting to make them eat healthy food [21]. Thus, for effective intervention, factors at household and school levels that affect the availability and accessibility of foods as well as at the individual level should be included [22].

It has been reported that the food environment has a greater effect on health and nutritional status in the vulnerable group than in the general group [23], but most of the studies have been conducted on the general group [24], and there is a dearth of research exploring how food environments at home and school affect fruit intakes and risk of overweight among disadvantaged children and adolescents. Therefore, research is needed to determine how the food environments of vulnerable children is related to healthy food intake and obesity. For this study, we surveyed low socioeconomic status youth to determine how select aspects of their food environments associate with consumption of fruit and weight status. Therefore, the objective of this study was to investigate whether food environments at household and school levels affect fruit intakes and risk of overweight among disadvantaged children and adolescents.

#### **2. Methods**

#### *2.1. Study Population*

Participants were recruited in 2015 from the Community Childcare Center, located in Seoul, Korea. They provide welfare services (e.g., protection, education, and meals) after school to youth from disadvantaged backgrounds (e.g., beneficiaries of national basic livelihood or single-parent families, etc.) [25]. During this time, the Community Childcare Center participated in the Healthy Fruit Basket Program, which aims to prevent the development of chronic diseases by providing access to fresh fruits [25,26]. In total, 4154 students (mean age; elementary school: 10.4 y, middle and high school: 14.3 y) were recruited; however, only 3148 were included in the final analyses because they supplied information for all the variables of interest and were of an eligible weight status (i.e., normal and overweight) [27,28]. The Institutional Review Board approved this study, and all participants gave written informed consent and assent (DKU 2015-10-016).

#### *2.2. Weight Status*

Height and weight were measured by a trained measurer at the public health centers or the Community Childcare Centers using a nationally certified weight and height scale. BMI was calculated as weight (kg) divided by height squared (m<sup>2</sup> ). According to the 2007 Korean National Growth Charts for children and adolescents [29], classification of normal weight was defined as having a body weight between the fifth and eighty-fourth percentiles, based on one's gender and age; classification of overweight was defined as having a body weight in the eighty-fifth percentile or above, or BMI 25 kg/m<sup>2</sup> or greater, based on one's gender and age.

#### *2.3. Fruit Intakes*

Fruit frequency questionnaires examined fruit intakes from the month preceding study enrollment. Scoring ranged from 0 (never) to 9 (3+ times per day). All responses were converted into times per day and collapsed into two categories (<0.5 times per day vs. ≥0.5 times per day) to determine the association with overweight.

#### *2.4. Food Environments*

Participants were asked five questions about their food environments, adapted from those employed in a previous investigation [30]. Here, three dimensions of the physical environment were evaluated: (1) Availability of fruit at home, (2) accessibility of fruit at home (i.e., is there someone who prepares fruit for children to eat?), and (3) accessibility of fruit at school (i.e., does the school provide fruit twice a week?). For the social environment, questions inquired about the habit of frequently eating fruit among parents and friends. Responses for the questions ranged from 1 (strongly disagree) to 5 (strongly agree), and all were collapsed into categories of disagree (strongly disagree, disagree), neutral, and agree (agree, strongly agree).

#### *2.5. Statistical Analysis*

Chi-square tests assessed the distribution of participants' general characteristics, the food environment, and weight status according to the frequency of fruit intakes and aspects of the food environment. T-tests evaluated differences in the frequency of fruit intakes by sex and grade level. An analysis of variance and Scheffe's post-hoc test determined differences between the food environments. Logistic regressions analyzed the effects of fruit intakes and food environments on weight status, showing odds ratios (ORs), 95% confidence intervals (95% CI). All analyses were performed by SPSS Statistics (v. 23.0); *p* ≤ 0.05 was considered statistically significant.

#### **3. Results**

General characteristics of the participants are shown in Table 1. Overall, 49.0% were boys and 71.5% were in elementary school. The percentage of overweight for boys and girls was 23.5% and 22.8%, respectively. The percentage of overweight for elementary school, and middle and high school students was 24.3% and 20.4%, respectively. There was a significant difference of weight status between grades, and the percentage of overweight was high in elementary school students.


**Table 1.** General characteristics of subjects <sup>1</sup> .

<sup>1</sup> Qualitative variables are presented as n (%). <sup>2</sup> *p*-values for differences between the weight statuses were obtained by a chi-square test.

Table 2 details aspects of the participants' food environments. Overall, 59.5% to 63.2% of participants responded positively (i.e., affirmed the availability of fruit at home, a regular provision of fruit in school lunches, having someone at home to prepare fruit for them, and a family habit of frequently eating fruit). Between the sexes, girls perceived their food environments more positively than boys (*p* < 0.001). Between the grade levels, elementary school students perceived their food environments more positively than middle and high school students (*p* < 0.001).


**Table 2.** Select aspects of participants' food environments <sup>1</sup> .

<sup>1</sup> Values are presented as absolute numbers and percentages. <sup>2</sup> *p*-values for differences between sexes or grade levels, obtained by a chi-square test.

Table 3 shows the relationship between aspects of participants' food environments and frequency of fruit intakes. Regardless of sex or grade level, fruit intakes differed according to how one's food environment was perceived. Here, the group who perceived their food environments positively was found to consume fruit 0.87 to 0.95 times per day, whereas the group who perceived their food environments negatively was found to consume fruit 0.42 to 0.67 times per day (*p* < 0.001). Interestingly, there were no differences in fruit intakes between sexes when they shared the same perception about their food environments. However, when both grade levels responded positively about their food environments, consumption of fruit was found to be higher among elementary school students. For groups displaying a negative or neutral perception, frequency of fruit intakes was comparable.


**Table 3.** Relationship between aspects of participants' food environments and frequency of fruit intakes <sup>1</sup> .

<sup>1</sup> Values are fruit intake per day, presenting as means ± standard errors. <sup>2</sup> *p*-values for differences between sexes or grade levels, obtained by ANOVA; letters (a,b,c,d,e) indicate significant differences between groups (Scheffe's post-hoc test, *p* < 0.05).

Table 4 shows the relationship between aspects of participants' food environments, frequency of fruit intakes, and weight status. The fruit intake frequency was not related to weight status in total subjects but as classified by grade; the lower intake frequency was shown in overweight groups of elementary school students. In terms of fruit environments and overweight, regular provision of fruit in school lunches and having a person at home to prepare fruit were negatively correlated to an overweight status for all participants. A low frequency of fruit intakes was associated with an overweight status among elementary school students. The proportion of overweight students was lower among those who perceived their food environment at school positively compared to those who did not (*p* = 0.044), especially for girls (*p* = 0.047) and elementary school students (*p* = 0.005). In addition, the proportion of overweight students was lower among those who had someone at home to prepare fruit compared to those who did not (*p* < 0.001), notably again for girls (*p* = 0.007) and elementary school students (*p* = 0.001).


**Table 4.**Relationship between aspects of participants' food environments, frequency of fruit intakes, and weight status1.

1 Values are presented as absolute numbers and percentages. 2 *p*-values for differences between sexes and grade levels, obtained by a chi-square test.


**Table 5.**Relationship between aspects of participants' food environments and odds of being overweight.

1 Odds ratios (ORs) were derived from logistic regression models. 2 Adjusted ORs adjusted for fruit intake frequency and were derived from logistic regression models. 3 *p*-values for trends.

Table 5 shows the relationship between aspects of participants' food environments and odds of being overweight, unadjusted and adjusted for fruit intake frequency. For boys, those who had someone at home to prepare fruit were unlikely to be overweight (OR = 0.64, 95% CI = 0.43–0.94, *p-trend* = 0.023), and these results remained significant after adjusting for frequency of fruit intakes (OR = 0.63, 95% CI = 0.420.93, *p-trend* = 0.019). For elementary school girls, regular provision of fruit in school lunches (OR = 0.52, 95% CI = 0.30–0.92, *p-trend* = 0.005) and having someone at home to prepare fruit for them (OR = 0.63, 95% CI = 0.43–0.93, *p-trend* = 0.004) were negatively related to an overweight status, even after adjusting for frequency of fruit intakes (OR = 0.54, 95% CI = 0.31–0.95, *p-trend* = 0.01 for regular provision of fruit in school lunches; OR = 0.65, 95% CI = 0.44–0.96, *p-trend* = 0.009 for having someone at home to prepare fruit). Among middle and high school girls, having someone at home to prepare fruit was significantly associated with a lower risk of being overweight after adjusting for frequency of fruit intakes (OR = 0.51, 95% CI = 0.27–0.96, *p-trend* = 0.06).

#### **4. Discussion**

Because of the prominent influence that food environments exert on an individual's tendency toward obesity, we explored the relationship between aspects of the home and school food environments, fruit intakes, and overweight status among disadvantaged children and adolescents. We found that the frequency of fruit intakes increased when participants perceived their food environments positively, and this was associated with a reduced prevalence of overweight. In particular, regular provision of fruit in school lunches and having someone at home to prepare fruit was associated with a healthier body weight among elementary school students and girls.

Fruit was consumed an average of 0.77 times per day by our participants, which was below the recommended level of the Dietary Reference Intake for Koreans of at least twice per day [31]. According to the Korea National Health and Nutrition Examination Survey [32], low-income families consume less fruit than medium-income families (100.1 vs. 135.2 g/day). In this study, frequency of fruit intake varied according to a participant's perception of their food environment, with higher rates of consumption among those who responded positively. These results provide evidence that food environments to increase fruit intakes would play an important role in fruit consumption.

Previously, review papers and meta-analyses have shown that increasing the availability and accessibility of fruits and vegetables for children at school is effective at preventing obesity [33,34]. In this study, regular provision of fruit in school lunches was positively related to fruit intakes among elementary school students, but not middle and high school students. These findings do not align with those from a related investigation in the United States, wherein high school students who received fruits and vegetables from The Fresh Fruit and Vegetable Program were likely to eat fruit more often than those who did not (59.1% vs. 40.9%) [35]. Given that middle and high school students appear to have more established eating habits, it may be necessary to provide fruit at a higher frequency in order to enhance intakes [35,36]. The school lunch service in Korea could provide fruits as a dessert within the school budget, but the current frequency of provision of fruit at school lunch would not be enough to meet the consumption of fruit for the disadvantaged middle and high school students in Korea. Therefore, additional funding for more frequent provision of fruit at school lunch should be needed to increase their fruit consumption.

Regarding the food environment at home, having someone to prepare fruit for students was found to relate favorably to fruit intakes and weight status. This result was consistent with those from other studies reporting that social support for healthy food intakes aides in the prevention of obesity among vulnerable children [37,38]. A few reasons might explain this. First, caregivers give children fruit in a form that allows them to eat it easily. Several studies have shown that providing fruit in a ready-to-eat form or making it visible to children promotes fruit intakes [39,40]. Because the process of washing, cutting, and peeling fruit has been described as an impediment to consumption, providing or storing it in an accessible form may help children eat fruit more frequently [40,41]. Indeed, analyses from the Healthy Habits randomized trial [42] revealed that the frequency of fruit provision from a parent

positively impacted their children's fruit intakes after 12 months. Second, caregivers could promote fruit intakes simply by encouraging their children [43]. Considering that participants in our study hailed from disadvantaged backgrounds, it is unlikely that their caregivers purchase much fruit for them or monitor their intakes at home. Thus, incorporating the Healthy Fruit Basket Program into all Community Childcare Centers and related institutions nationwide may help these individuals increase fruit intakes and attenuate risk of overweight or obesity [25,26].

It is well established that availability of food in the home is a key determinant of consumption among children and adolescents [44–46], and our analyses confirmed this phenomenon for fruit, in particular. Additionally, we observed that a family's habit of eating fruit frequently was connected to high fruit intakes across all participants, similar to findings from other publications [39,42,43].

Interestingly, this investigation detected no difference in fruit intakes between sexes sharing the same perception about their food environments. However, with respect to the grade levels, elementary school students were found to consume fruit more frequently than middle or high school students when their food environment was perceived positively. Childhood is a period of social modeling, as individuals learn how to interact with their environments and behave appropriately [11,47,48]. Hence, the benefits of food environments on fruit intakes may apply more strongly to children than adolescents. For another explanation, a recent study of adults with low socioeconomic position found that self-efficacy on fruit and vegetable consumption was more strongly associated with fruit and vegetable consumption than perceived food environments, which implies the importance of capacity building to partially overcome the poor food environment [49]. As several studies have shown the positive effect of nutritional education on fruit and vegetable intakes [50–52], nutritional education intervention should be included for adolescents with negative perception of food environment who especially have low self-efficacy to increase their fruit and vegetable intakes.

When we examined the relationship between fruit intakes and weight status, a high frequency (>0.5 times per day) was tied to a lower prevalence of overweight in girls. According to systematic reviews, consumption of fruit as a means to prevent pediatric obesity did not always produce consistent results in terms of gender or amount (38,39). For example, one investigation noted that eating fruit more than twice a day was only protective against obesity in boys [53], and another showed that excessive fruit intakes actually engendered obesity [54]. Heeding our results and those from earlier studies, consumption of fruit alone may not be adequate to prevent obesity. Specifically, our investigation found that the effects of some aspects of the food environment on the prevalence of obesity remained significant after adjusting for fruit intake frequency, and a related study observed a positive link between parental concern for their children's diet and fitness practices (*r* = 0.552, *p* < 0.001) [55]. It could suggest that the home environment supporting children to eat fruits could also support other behaviors that can prevent obesity in children, such as encouraging exercise. Further research exploring caretakers' interest in their children's health will provide a more holistic understanding of how environments moderate their risk of obesity.

The present study had several limitations and strengths. Concerning limitations, this was a cross-sectional study, so any causal associations between food environments, fruit intakes, and overweight status have yet to be determined. In addition, the questionnaire evaluating fruit availability and accessibility in the home and school food environments was not validated, although it had been employed in a previous study [30]. Moreover, we did not collect information pertaining to other confounders of overweight, which could offer deeper insight into factors underlying the present findings [25,26]. Granted, this was a highly homogenous cohort since all participants were recruited from the Community Childcare Center, so it is unlikely they possessed any remarkable characteristics that would alter our results. Nonetheless, future studies examining a more diverse population are needed to confirm the relationship between the food environment and risk of obesity by adjusting the confounding factor. Despite limitations, these results provide evidence that food environments play an important role in overweight prevention as well as fruit consumption among underprivileged Asian children and adolescents, using a relatively large sample size.

#### **5. Conclusions**

This study found that the frequency of fruit intakes was generally low among children and adolescents from disadvantaged backgrounds; however, those who perceived their food environments positively consumed more fruit and were less likely to be overweight than those who perceived them negatively. Noteworthily, regular provision of fruit in school lunches and having someone at home to prepare fruit were predictive of a high frequency of fruit intakes and lower prevalence of overweight, and these discoveries were most apparent among elementary school students and girls. On the whole, our findings demonstrate that augmenting access to fruit within any realm of the food environment is associated with increased consumption and healthier body weights for low-income youth. Going forward, the school lunch service in Korea could consider increasing fruit servings since current provisions are not sufficient for enabling disadvantaged middle and high school students to meet the dietary guidelines. In addition, a nation-wide program to increase fruit consumption at the Community Childcare Center where they usually spend time would be essential to improve their health.

**Author Contributions:** K.K. conceived and designed the study; N.Y. analyzed the data and wrote the first draft of the paper; K.K. supervised the data analysis and revised the draft paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Research Foundation (NRF) of Korea (2019R1A2C1084372).

**Conflicts of Interest:** All authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
