Adolescence is a unique life stage in which accelerated growth and development bring increased nutrient requirements and possible heightened risk of deficiencies^1–4. In particular, adolescents are affected by iron deficiency^1,5, which frequently results in anemia and can lead to negative consequences including compromised growth, cognitive impairment, suboptimal immune function, and adverse birth outcomes^2,6–9.

In recent years, attention to the food environment as a critical influencer of health and nutrition outcomes has increased, and the body of research on this topic is steadily growing^10–13. The food environment has been defined as “the consumer interface with the food system that encompasses the availability, affordability, convenience, and desirability of foods,”^10,11 and researchers have used the socio-ecological model to situate various types of food environments—wild, built, and cultivated—under the umbrella of influence from various sectors (e.g., agriculture, labor, media) and broader socio-cultural and political environments¹¹.

Various theories and models have been proposed to help explain the multiple factors (e.g., context, ideals, personal beliefs, and resources) affecting the food choice process throughout the life course^14–16. A 2002 paper offered a theoretical framework, based on social cognitive theory and an ecological perspective, for understanding the various factors—intrapersonal, interpersonal, community, and societal—that influence adolescent eating behavior¹⁷. Adolescence is a critical period for social development and is characterized by a marked increase in autonomy and agency^4,18. As they mature, adolescents become more empowered to engage in decision-making and begin to form dietary habits that may persist in adulthood^3,4. Although family influence remains strong in this life period, adolescents are increasingly influenced by social and cultural norms, which have an impact on their dietary patterns and nutritional status^1,17. Social pressures from peers, body image, and mass media are key drivers of adolescent dietary choices^4,19. Young people have strong opinions regarding their food consumption, and they are enabled to act on these opinions through access to funds in the form of pocket money and/or from their engagement in income generating activities⁴. Adolescents living in poverty and food insecurity experience additional food-related stressors that put their health at risk, and across contexts, adolescents frequently encounter food environments that are unconducive to healthy choices^4,14.

Schools are an important entry-point for reaching children with health and nutrition interventions including those focused on promoting healthy and diverse diets²⁰. For example, it has been estimated that 370 million children globally receive school meals every day²¹. School meals are an important contributor to children and adolescents’ daily intake of foods and nutrients, but the portion of a child’s daily nutrient requirements that is satisfied through school meals varies by context^22–25. In some contexts, children may consume as much as half of their daily energy at school²⁶. A 2018 study in Nigeria found that school meals provided over one third of recommended nutrient intakes (RNI) for protein, vitamins A and C, and zinc, but fell short on meeting a third of the RNI for energy, calcium, and iron²⁴. A 2019 study in Sweden found that school lunch provided approximately half of daily vegetable intake and two-thirds of daily fish intake, and that daily energy-adjusted nutrient intakes were lower for pupils of less educated parents²³.

The 2022 edition of the School Meal Programs Around the World report, based on survey data collected from 139 countries representing 81% of the world’s population, revealed that 125 countries had at least 1 large-scale school meal program²⁷. The Ghana School Feeding Program was initiated in 2005 as a social protection measure in support of achieving the Millenium Development Goals and aims to provide low-income school children with 1 free, nutritionally adequate meal per day²⁸.

The role of the school food environment in shaping adolescent nutrition is increasingly studied^{22,26,29–33}; however, much of this research has focused on high-income countries, leaving a critical evidence gap in low- and middle-income countries. A 2020 systematic review found no standardized methods used to measure the school food environment²⁶. The authors grouped various methods found across the literature into 4 dimensions as defined by Swinburn et al.³⁴: physical, economic, socio-cultural, and policy. These dimensions can be used as a framework to guide a systematic approach to studies of the school food environment.

The purpose of this study was to explore the extent to which the school food environment in Ghana, including but not limited to school meals, supports girls and boys to consume healthy and diverse diets. We used nationally representative data to characterize multiple dimensions of the school food environment that influence adolescent food access, food choice, and ultimately their health and nutrition status.

Across the 4 regions, 239 of the selected 242 schools were reached (response rate: 99%), 117 of which were JHS, and 122 of which were SHS (table not shown). Two schools were not reached because one was closed at the time of data collection and the other was a school for the deaf where a translator was not available. In the third school, student questionnaires were conducted but there was not sufficient time to conduct the school-level questionnaire. Among the 117 JHS, 116 (>99%) were day schools and 1 (>1%) was a boarding school. Among the 122 SHS, 29 (24%) were day schools, 17 (14%) were boarding schools, and 76 (62%) had both boarding and day options. All schools were public schools. Most (63%) were located in an urban setting while 29% were in rural and 8% in peri-urban settings. From these schools, 4833 girls and 1498 boys participated in the survey, resulting in an analytical sample of 6331 adolescents aged 10–19 years (response rate: ^~90%) (Table 1).

The mean age among girls was 16.4 (95% CI: 16.3, 16.4), and 92% had reached menarche (Table 1). Among boys, the mean age was 16.5 (95% CI: 16.5, 16.6), and 69% had experienced voice change. Anemia prevalence was significantly higher among girls compared to boys (22% vs 11%, p-value <.0001) (Supplemental Table 1). Compared to boys, a significantly larger proportion of girls had heard of anemia (p-value <.01) and iron-folic acid tablets (p-value <.0001) and taken supplements containing iron and/or folic acid in the past 7 days (p-value <.0001).

Among SHS, 66% routinely provided free meals to all students, and 27% provided free meals to some students, while in JHS, 90% did not provide meals to any students. The presence of vendors (defined as anyone selling food on campus) was high at both JHS (88%) and SHS (99%) (table not shown). Sixty-four percent of students reported eating meals provided by the school while 54% and 85% reported bringing food to school and buying food at school, respectively (Supplemental Table 2). Most students reported eating meals sold or prepared at school 3–5 days each week (34%) or ≥6 days each week (57%, driven by boarding schools). Seventy-nine percent and 87% reported eating breakfast and lunch, respectively.

Among the 21 food groups observed by data collectors at SHS, the most prevalent were beans, peas, and lentils (92%), porridges, bread, rice, pasta, and other local starchy foods (91%), and other vegetables including tomatoes, okra, and peppers (90%) (Supplemental Figure 1). Conversely, at JHS, the most prevalent food groups observed on school grounds were eggs (76%), other vegetables including tomatoes, okra, and peppers (76%), and locally produced juices (72%). The proportion of SHS where the MDD-W food groups were observed was significantly higher for 7 out of the 10 food groups compared to JHS (Table 2). On average, 6.2 food groups were available at JHS and 8.1 at SHS, a difference that was significant (p-value <.0001). The mean FES was 10.9 (95% CI: 10.1, 11.7) and 14.3 (95% CI: 13.6, 14.9) for JHS and SHS, respectively (Table 3).

Nearly all adolescents reported that the food group made up of grains, cereals, white roots, and tubers was typically included in school meals (99%), brought to school (92%), and purchased at school (95%) (Figure 1). Other items typically included in school meals were pulses (91%), fish and seafood (57%), and other vegetables (52%) while items commonly brought from home were milk and milk products (62%), sweets and desserts (48%), and sweetened drinks (43%). When students were asked what food items they typically buy on campus, the most reported items after starchy foods (95%) were pulses (66%), meat and poultry (57%), eggs (53%), fish and seafood (48%), and other vegetables (47%).

Per 24-hour dietary recall of 21 food groups, the food group with the highest prevalence of consumption was porridge, bread, rice, pasta, and other local starchy foods (93% girls; 91% boys), followed by other vegetables including tomatoes, okra, and peppers (68% both sexes), plantains, white potatoes, yams, cassava, and other white tubers (53% girls; 56% boys), and fish and seafood (46% both sexes) (Supplemental Figure 2). The food groups with the lowest prevalence of consumption among girls and boys were mango and papaya (4% both sexes), organ meats (6% girls; 8% boys), and cheese and yogurt (9% girls; 7% boys).

We found significant differences in the consumption of 3 of the 10 MDD-W food groups by sex (Figure 2). A significantly greater proportion of boys consumed pulses compared to girls (45% vs 35%, p-value <.01), while girls were more likely to consume other vitamin A-rich fruits and vegetables (14% vs 11%, p-value = 0.03) and other fruits (23% vs 17%, p-value <.01) compared to boys. ADDS differed by school level (JHS: 4.8; SHS: 4.2) and wealth status (high: 4.4; medium: 4.2; low: 4.1) but not by sex (girls: 4.3; boys: 4.2) (Figure 3). The proportions of adolescents with minimally diverse diets based on the MDD-W (≥5 food groups) and MDD-A (≥6 food groups) also differed significantly by school level and wealth status. Among JHS students, 53% achieved MDD-W and 32% achieved MDD-A while among SHS students, 43% and 24% did so, respectively. The proportion of students achieving both MDD-W and MDD-A was positively associated with wealth status.

Being from a school with a higher FES was a significant predictor of higher ADDS and consuming ≥5 food groups among SHS students but not JHS students (Table 4). Among both girls and boys, FES was inversely associated with anemia. For girls, adjusted anemia prevalence was 27% in low FES schools and 22% in high FES schools (p-value <.01), with 24% lower odds of anemia for those from high FES schools (odds ratio [OR]: 0.76, CI: 0.63, 0.93) compared to low FES schools (Table 5). Among boys, adjusted anemia prevalence was 19% in low FES schools and 13% in high FES schools (p-value <.01), with 46% lower odds of anemia for those from high FES schools (OR: 0.54, CI: 0.34, 0.84) compared to low FES schools.

Maturation status as assessed by puberty was a significant predictor of higher ADDS, consuming ≥5 food groups, and consuming ≥6 food groups among SHS students but not JHS students (Table 4). Among girls and boys, puberty was a significant predictor of higher Hb, with pre-pubertal students having lower Hb by 0.1 g/dL and 0.3 g/dL for girls and boys, respectively (table not shown). Having reached puberty was associated with 43% lower odds of anemia (OR: 0.57, CI: 0.41, 0.78) in boys and 19% lower odds in girls (OR:0.81, CI:0.61, 1.08), but was significant only among boys (Table 5).

Wealth status was a significant predictor of higher ADDS, consuming ≥5 food groups, and consuming ≥6 food groups among both JHS and SHS students (Table 4). Among JHS students, the positive association between wealth and consumption of ≥5 food groups was particularly notable, with 62% of higher wealth students achieving this cutoff while only 46% of lower wealth students did so.

Previous work has outlined limitations of the evidence base for adolescent nutrition due to a lack of data on adolescent diets, especially from low- and middle-income countries (LMICs), and called for future data collection and analysis efforts to include disaggregation by age, sex, socioeconomic status, and area of residence ^1,4,51. A 2022 paper presented dietary data from adolescents in China, Mexico, and India ⁴ while others have focused on adolescent diets in single-country settings including the Philippines ⁵², Tanzania ⁵³, and the United States ^14,54,55. Despite this important progress, a 2024 systematic scoping review found that nationally representative dietary data for adolescents are still limited, and that among LMICs, only 8% of data sets contained dietary details for ≥7 food items or groups ⁵⁶. Our work contributes to the evidence base by offering a detailed account of adolescent food-related behavior, dietary intake, and nutrition outcomes disaggregated by school level, sex, maturation status, and wealth status in Ghana.

A limited number of publications have focused on the school food environment in Ghana and its impact on child dietary behavior and growth ^30,57,58. A 2017 study used data from household and school questionnaires to examine the role of the school food environment on child and adolescent dietary behaviors, with a focus on school meals, distance child lives from school, predictors of children bringing money to school, and broad categories of foods available on campus (i.e., meals, confectionary, fruit, and sugar-sweetened beverages) ³⁰. Similarly, a 2020 study in a single municipality in Ghana described students’ typical purchases during morning break, focusing on energy-dense foods and general categories such as meals, packaged snacks, fried foods, and sweetened beverages ⁵⁸. Our study builds on this previous work by using nationally representative data from JHS and SHS to characterize multiple dimensions of the school food environment and provide a detailed assessment of multiple foods and food groups that are commonly accessible and consumed by in-school adolescents in Ghana.

Our study revealed notable differences in the foods served at lunchtime or available for purchase at JHS and SHS. This difference may be partially driven by the fact that the majority of SHS in our study offered meals to all or some students while most JHS did not offer meals to students. These trends are reflective of national norms. Many SHS are boarding schools and therefore provide meals to students. Conversely, most JHS are day schools and do not routinely serve meals, and thus students in these schools rely more heavily on food brought from home or available for purchase on campus or in the vicinity. Seven of the 10 MDD-W food groups were observed at a larger proportion of SHS compared to JHS, which could reflect the effort invested in menu planning where school meals are routinely served.

The mean FES, which takes into consideration both proximal and distal contributors to health and nutrition, was 3.4 points lower in JHS compared to SHS. Qualitatively, this difference seems largely to be driven by the physical elements of the school food environment (e.g., routine provision of meals, the number of food groups served or available for purchase, having a designated place where students eat, access to drinking water, etc.), for which proportions were generally higher at SHS compared to JHS. Simultaneously, the number of food groups observed at SHS was significantly higher than at JHS. Yet ADDS and the proportion of adolescents consuming ≥5 and ≥6 food groups were higher among JHS than SHS students. This disparity could be driven by a relatively less insular food environment in JHS, including the lack of school meals provision, which leads students in these schools to obtain most of their meals at home or by purchasing from vendors, either on campus or off school premises. Meals eaten at home, brought to school, or purchased by JHS students from vendors may offer greater variety and choice than SHS students have on campuses where meals are served to them. Further, the presence of more food groups on SHS campuses does not guarantee their desirability or affordability and thus does not necessarily translate to better dietary diversity for those students. A study in Tanzania found that diets in boarding schools were monotonous and composed mainly of cereal and legume intake ⁵³. JHS students are also relatively younger and could be more influenced by family members who encourage them to eat specific foods. A systematic review of family influence on adolescent diets concluded that increased autonomy in older adolescence alongside a reduction in the amount of parenting time are important factors in food choice and intake ⁵⁹.

While dairy products are not a main component of traditional diets in Ghana, 62% of adolescents reported bringing milk and milk products to school. This tendency could be an indication of the success of marketing of dairy products toward children and a trend toward increasing dairy consumption in countries without strong dairy traditions^60,61. The high proportions of adolescents who reported bringing sweets, desserts, and sweetened drinks to school are consistent with a recent study of adolescent diets in the context of the nutrition transition, which found a steady increase in added sugar intake in multiple countries over an 8-year period⁶². A key aim of the national Nutrition Friendly Schools Initiative is to address the double burden of undernutrition and diet-related chronic disease affecting school children in Ghana, and thus there may be room for enhancing school-level guiding principles and action plans to discourage students from bringing unhealthy foods to campus.

Our results showed that the foods adolescents tended to purchase at school were: grains, cereals, white roots, and tubers; pulses; meat and poultry; eggs, and; fish and seafood. This finding indicates that when adolescents have the purchasing power to do so, they often favor protein-rich food items, and that the foods purchased on campus could make an important contribution to daily micronutrient intakes. The difference in foods brought to school compared to foods purchased on campus may be due in part to the convenience to parents of giving their children pre-packaged items rather than preparing and packing meals at home.

Despite greater awareness of anemia and iron-folic acid tablets and consumption of supplements containing iron and/or folic acid in the past week among girls compared to boys, anemia prevalence was still significantly higher among girls. This result reflects the physiologic differences between boys and girls, and especially the higher biological requirements for iron for girls. Overall, boys and girls had similar food group consumption patterns, a finding that remained true when food groups were aggregated into the 10 MDD-W food groups. Boys were more likely to consume pulses while girls were more likely to consume other vitamin A-rich fruits and vegetables and other fruits. These differences could be related to personal preferences as well as perceptions of the health attributes of certain foods^4,17,19. The high proportion of girls and boys consuming starchy foods, other vegetables, and fish and seafood is consistent with dietary practices and cultural norms in Ghana. The low proportion consuming mango and papaya could be driven by seasonality as data were collected in November when mangos were not in season. Research in Mozambique and Zambia has noted important differences in beta-carotene intake in and out of mango season^63,64.

The FES was a significant predictor of dietary diversity outcomes at the SHS level, but not at the JHS level, suggesting that adolescent dietary behavior at the JHS level is driven by factors other than the school food environment (e.g., wealth status). The FES was also inversely associated with anemia in both boys and girls. These results related to important program indicators (i.e., dietary diversity and anemia) suggest that the Nutrition Friendly Schools checklist—which we used to create the FES—appears to be well-designed to capture correlates of health and nutrition among school-aged children.

Maturation status was a significant positive predictor of all 3 dietary diversity outcomes at the SHS level and was also predictive of higher Hb in girls and boys, and lower prevalence of anemia in boys (but not girls), even after controlling for age. This finding could be an indication that, even as nutrient requirements increase during adolescence, the social changes that often accompany puberty, including heightened autonomy, access to spending money, and health consciousness, may result in improved dietary practices, especially related to the consumption of iron-rich foods^4,14,17. Various studies have used the Health Belief Model to understand adolescent dietary behavior and have found that perceived benefits and threats, self-efficacy, and cues to action are important predictors of diet and food-related behavior^65–68.

A recent study in Filipino school-aged children and adolescents found that socioeconomic status played an important role in dietary intake, and that children from poorer households were more at risk of consuming a less diverse diet⁵². Similarly, we found that household wealth status was a significant predictor of dietary diversity outcomes at both the JHS and SHS level. However, the impact on ADDS was more pronounced in students from JHS compared to SHS, with low wealth JHS students consuming nearly a whole food group less compared to their peers from high wealth households. This could be explained by the fact that many SHS are boarding schools, and thus more insulated, such that wealth does not have as strong an effect on ADDS compared to JHS. Wealth was not a significant predictor of Hb or anemia in boys or girls, highlighting the complex nature of the relationship between diet and nutrition outcomes, in particular in the case of anemia. While increased access to resources may enable adolescents to access and consume a wider range of foods, anemia has a complex etiology involving substantial increase in nutrient requirements as children age alongside the effects of malaria, inflammation, infection rates, and other factors^69–73.

The strengths of this research include the use of a large, nationally representative sample of in-school adolescent boys and girls, the inclusion of both dietary intake data and biological data, and the creation of a 23-point score to quantify the school food environment. However, there are several limitations to this study. Our data are from a single population of in-school adolescents attending JHS and SHS in Ghana and thus may not be generalizable to other populations. We applied both the MDD-W indicator, validated only for women of reproductive age (15–49 years), and the MDD-A indicator, which we developed using NHANES data for US adolescents, but which has not yet been validated in other adolescent populations. Although neither indicator is perfectly suited to our study population of adolescents aged 10–19 years in Ghana, we applied both cutoffs to provide a more complete picture of the potential inadequacy of diets. These indicators are subject to the limitations of reported dietary data (i.e., biases like recall and social desirability as well as measurement error) relative to the more objective micronutrient status indicated by biomarkers⁷⁴. Further, our single 24-hour recall data are based on what adolescents consumed in any environment, not only in school, and thus do not allow us to make inferences about usual intake or about intake during school only. While girls underwent venous blood draws, boys underwent capillary blood draws, which could introduce error to the comparison of anemia prevalence. Finally, the school FES we developed is intended only for use in Ghana and is not validated but could be used as a starting point for developing other country-specific and/or global school food environment scores.

Among in-school adolescents in Ghana, a higher school FES was positively associated with dietary diversity among SHS students, but not among JHS for whom factors beyond the school food environment (e.g., wealth) may have accounted for dietary diversity. For both girls and boys, high FES was inversely associated with anemia. The findings from this study help to fill key evidence gaps around reaching adolescents in Ghana with school-based health and nutrition programs to improve dietary diversity and reduce anemia. Our results suggest that a multi-pronged approach, which seeks to enhance the school food environment through policy, physical, and sociocultural improvements, could play a role in determining diet and nutrition outcomes among adolescents in Ghana.