This review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Literature searches were conducted using two databases, PubMed (https://www.ncbi.nlm.nih.gov/pubmed) and Ovid-Embase (http://ovidsp.tx.ovid.com) for articles published up to July 2024. The search strategy included the following keywords and Boolean operators: “sustainable diet*” OR “sustainable food*” OR “sustainable nutri*” AND “school lunch” OR “school meal*” AND “student*” OR “adolescents” (Appendix 1). Studies with diverse methodologies ‒ including qualitative research, quantitative research, randomized controlled trials, and mixed-methods approaches ‒ were considered to ensure a comprehensive synthesis of intervention strategies.

Studies were included if they met the following criteria: (1) addressed sustainable diets; (2) involved school meals or lunch programs; (3) targeted children or adolescents aged 5 ‒ 18 years (the age range of students in primary/elementary to high school across different countries); (4) published in English; (5) available as full-length peer-reviewed papers; and (6) published by July 2024. Exclusion criteria included: studies focused on infectious diseases, conference proceedings, review papers, and guidelines.

Duplicates were removed, and a two-phase screening was performed. In the first phase, titles and abstracts were reviewed to exclude unrelated articles. In the second phase, full texts were assessed against the inclusion and exclusion criteria.

From each eligible study, the following data were extracted: title; authors; country; year of publication; study design; population; basic information of the subjects (race, sample sizes, age range, sex distribution, description, and dietary assessment); intervention (exposed) and control (comparison) regimes (number of participants, sex distribution, age range); outcome results. To assess study quality across diverse designs, the Joanna Briggs Institute (JBI) critical appraisal tools were applied (https://jbi.global/critical-appraisal-tools). Each study was independently assessed using JBI guidelines (yes/no/unclear). Articles that received a ‘Yes’ score of less than 70% were excluded from the synthesis due to insufficient methodological quality.

A total of 87 records were retrieved through database searches (41 from PubMed and 51 from Ovid-Embase, with 5 duplicates). After removing duplicates and screening titles and abstracts, 24 articles underwent full-text review. Ultimately, 12 studies met the predefined inclusion criteria and were included in this systematic review (Figure 1).

Figure 1.

PRISMA flow chart of study selection process

The 12 studies, published between 2022 and 2024, were conducted across six continents (4 North America, 3 Europe, 2 Oceania, 1 Asia, 1 South America, 1 Africa), involving diverse populations ranging from early childhood to late adolescence (0 ‒ 18 years) (Table 1). Study designs included randomized controlled trials (RCTs), quasi-experimental designs, methodological development studies, and mixed-methods approaches. Sample sizes ranged from small pilot groups (n = 36) to large-scale implementations (n > 1,000). Interventions were conducted in primary and secondary schools and addressed multiple domains including behavioural economics, environmental sustainability, digital technology, nutrition education, and policy advocacy (Table 2 and Table 3). All studies met the methodological quality criteria (moderate and high quality) of the JBI (Appendix 2), ensuring robust internal validity.

Two studies applied behavioural economics principles in cafeteria settings. Byker Shanks et al. (2023) implemented School Lunch Advisory Councils in five U.S. high schools [9], engaging students in cafeteria redesign and menu restructuring. This resulted in a statistically significant increase in vegetable selection (p = 0.014) and a 26% reduction in waste from the salad bar. Methods included direct weighing and SmartLunchroom evaluation tools.

Similarly, Pittman introduced a non-monetary bell-ringing incentive during lunch [10], leading to a 49% increase in healthy food selections among elementary students, with effects persisting after 30 days. These low-cost, scalable interventions demonstrate the potential of embedded nudges in reshaping food behaviours. Additionally, the curriculum-based lessons on food systems, paired with student-created posters, were effective in promoting healthier choices and reducing waste among middle school students (p = 0.029) [11].

Prescott's middle school cafeteria campaign (not shown in Table 1 but discussed in the narrative) empowered students as agents of change [11], resulting in statistically significant reductions in vegetable waste. The intervention used participatory approaches and visual prompts, contributing to reduced food disposal and enhanced cafeteria culture.

Three studies focused explicitly on climate-smart meal interventions. In Brazil, Kluczkovski et al. (2022) evaluated the Sustainable School Program (SSP) [12], which replaced conventional meals with plant-based alternatives. Nutritional adequacy was maintained while GHGE was reduced by up to 17% (p = 0.043). Dietary assessments were conducted using the Dietbox platform, and food production modelling provided environmental impact estimates.

In Sweden, Eustachio Colombo’s OPTIMAT™ intervention leveraged appealing menu names and teacher involvement to promote acceptance of sustainable meals [13]. Acceptability and student feedback underscored the importance of sensory and cultural adaptation. Meanwhile, in Poland, Jeruszka-Bielak et al. (2024) developed and validated the SMI-LE tool to assess the nutritional quality of 50 school menus [14]. The analysis revealed gaps in calcium, iron, and sodium provision, highlighting the need for micronutrient-sensitive menu planning.

Multicomponent interventions demonstrated strong outcomes in both health and knowledge domains. Wang et al. (2022) conducted the Meals, Education, and Gardens for In-School Adolescents (MEGA) project in Tanzanian schools [15], combining school meals, garden-based learning, and parental education in a cluster RCT involving 1,500 participants. Outcomes included improved hemoglobin concentrations and anthropometric indicators. Nutrition knowledge also increased, assessed via KAP surveys.

Goldberg’s Shape Up Somerville used community-based participatory research (CBPR) to redesign school menus and integrate wellness policies [16], resulting in increased intake of whole grains and low-fat dairy. Similarly, Johansson’s Nordic diet trial in infants observed a 32% increase in fruit and vegetable consumption in the intervention group [17], indicating long-term potential for early dietary exposures to shape sustainable habits. The intervention also led to a significant decrease in protein intake (p < 0.001), though the levels remained within recommended dietary ranges, supporting the nutritional adequacy of reduced-protein, plant-based diets [17].

Braga et al. (2024) piloted the FRANI app among 36 Vietnamese adolescent girls [18]. This AI-powered, gamified tool enabled real-time dietary assessment and provided feedback aligned with the EAT-Lancet targets. The intervention yielded a significant increase in Eat-Lancet Diet Scores (mean +1.09 points, 95% CI: 1.01–1.20, p = 0.03), with adherence and acceptability exceeding 80% and 97%, respectively. The study illustrates the feasibility of mobile health (mHealth) platforms in school contexts.

Two Australian studies tackled broader systemic changes. Deavin’s school breakfast program [19], utilizing donated food, not only diverted 14.4 tons of waste but also improved student attendance and classroom engagement. Waters et al. implemented the “Fun ‘n Healthy in Moreland!” program [20], involving stakeholders across 14 schools. The increase in fruit and water consumption was statistically significant (fruit: p = 0.05), while the absence of BMI change (p = 0.44) underscores the need for longer-term evaluation. The study also successfully facilitated wellness policy uptake across schools. While no significant effect on BMI z-scores was detected, improvements in fruit, vegetable, and water consumption were documented, and institutional policy adoption increased.

The analysis revealed several particularly promising strategies that demonstrated consistent effectiveness across diverse contexts: 1) Behavioral economics interventions – Simple changes in cafeteria design, default options, and incentive cues increased fruit and vegetable consumption while reducing food waste; 2) Whole-school approaches ‒ Interventions that aligned curriculum, food service, and community partnerships produced more sustainable changes than isolated programs; 3) Sustainable and plant-based menu planning ‒ Substituting animal-based items with plant-based meals reduced greenhouse gas emissions (GHGE) while maintaining nutritional adequacy; 4) Digital innovations – Mobile applications, online platforms, and gamified feedback mechanisms effectively engaged adolescents and reinforced healthy eating behaviors; 5) School gardens and experiential learning ‒ Hands-on activities such as gardening or cooking classes significantly improved students’ nutrition knowledge, attitudes, and willingness to try new foods.

These strategies are elaborated in detail in the subsequent subsections, including their success factors, limitations, and contextual influences.

Across the 12 included studies, several consistent outcomes were identified: 1) Dietary improvements ‒ Vegetable consumption increased by 32–54% in interventions combining education with environmental changes. Fruit intake rose by 18–27% in programs offering free or subsidized produce; 2) Sustainability outcomes ‒ Studies implementing plant-based menus reported GHGE reductions of up to 17%. Waste-focused interventions documented decreases in plate waste ranging from 15% to 28%; 3) Knowledge and attitudes ‒ Nutrition education and experiential learning programs consistently improved self-reported awareness of healthy diets, with positive spillover effects on household food choices; 4) Systemic change ‒ Whole-school approaches showed stronger long-term retention of dietary improvements compared to single-component interventions.

Across studies, interventions that combined multiple components ‒ such as education, environmental restructuring, and digital engagement ‒ demonstrated the strongest and most sustained improvements in dietary intake, environmental outcomes, and student engagement [11] [12] [16] [18]. A recurring theme was the importance of co-design and stakeholder involvement, particularly in achieving policy adoption and long-term sustainability [13] [16] [19].

Interventions targeting younger age groups, such as Johansson’s trial on infant complementary feeding [17] and Goldberg’s early elementary school reform [16], highlight the value of early exposure to sustainable practices. Additionally, qualitative findings from Eustachio [13] and Deavin [19] offered rich insights into implementation dynamics and stakeholder perspectives, emphasizing the complementarity of quantitative and qualitative approaches. Sustainability considerations, including food waste, greenhouse gas emissions (GHGE), and plant-based menus, were explicitly addressed in 7 of the 12 studies [11] [14] [16] [18] [19], reflecting a growing global priority. Notably, tools such as SMI-LE [14] and FRANI [18] offer promising avenues for standardized monitoring and cross-context application.

A dedicated table was constructed to compare sustainability indicators across the included studies (Table 4). The indicators identified included: Greenhouse Gas Emissions (GHGE) measured in CO₂eq via life cycle assessment; Food waste assessed by direct weighing of plate waste or kitchen waste logs; Water usage modeled using agricultural production databases; Menu diversity calculated through dietary diversity scores; Affordability evaluated through cost per serving analysis.

GHGE and food waste were the most consistently applied measures, but methodologies varied. Water use and affordability indicators were seldom reported, and no study considered soil health or broader social equity dimensions. The heterogeneity of measurement approaches underscores the need for standardized frameworks.

These findings align with and extend prior syntheses showing that whole-school and multi-component designs outperform isolated, single-focus programs by embedding change across the learning environment, school ethos, and family/community interfaces [21]. In education for sustainable development, whole-school models similarly improve implementation fidelity and durability of practice [22]. By explicitly integrating nutrition and sustainability aims, the present review addresses a gap in earlier reviews that tended to treat these domains separately, and provides an applied bridge from health-promoting schools to sustainability-oriented school food systems.

Notably, interventions grounded in behavioural economics ‒ such as default vegetable servings, salad bar redesign, menu labelling, and reinforcement cues ‒ were found to be effective and culturally adaptable across a wide range of contexts [9] [11]. These strategies leverage subtle environmental and psychological prompts to influence decision-making without relying on extensive cognitive or economic investment from students, rendering them scalable and practical in both high- and low-resource settings.

Furthermore, interventions that explicitly linked health and environmental goals, such as sustainable meal planning and plant-based menu adoption, were especially impactful. For example, Kluczkovski et al. demonstrated that GHGE could be significantly reduced through simple substitutions while maintaining nutritional adequacy [12]. Similarly, the SMI-LE tool developed by Jeruszka-Bielak et al. provided a replicable model for evaluating the nutritional quality of school meals with implications for policy benchmarking [14].

Technological innovation also emerged as a promising new domain. The FRANI app [18], tested in Vietnam, showcased high adolescent engagement, with measurable improvements in diet quality and user adherence. Digital nudges and gamified platforms may hold particular promise for adolescent populations [6] [18], especially in urban or digitally connected settings. However, further validation and equity-focused adaptation are necessary for broader scale-up.

Importantly, studies that prioritized student agency and participatory design consistently reported higher engagement and better outcomes [13] [16]. Interventions involving students in menu development, school gardens, or awareness campaigns not only increased ownership but fostered health-promoting norms within the broader school culture. This reinforces prior findings suggesting that youth involvement is a key determinant of intervention sustainability [11] [13] [16].

Despite these encouraging findings, this review also identified several persistent barriers. The lack of longitudinal studies remains a central limitation. While most interventions reported short-term improvements in dietary behavior or food waste reduction, few evaluated the durability of these changes. Behaviour change is often iterative and vulnerable to reversal; thus, long-term follow-up is essential to validate sustained impact.

Another critical gap is the underreporting of economic outcomes. Only a small subset of studies assessed cost-effectiveness, financial feasibility, or budget impact [12] [15] [16]. Without such data, scaling decisions remain speculative ‒ particularly in resource-constrained settings. Future research should embed economic evaluations within intervention designs to guide administrators and policymakers toward efficient resource allocation.

Resource limitations (e.g., staff shortages, infrastructure constraints), resistance from key stakeholders, and lack of interdepartmental coordination were frequently cited as challenges to implementation. In some cases, cultural mismatch ‒ such as unfamiliar food items or imported menu models ‒ hindered acceptance. In the OPTIMAT™ study, staff reported increased workload and logistical stress during the sustainable meal transition, suggesting that gradual implementation and additional training may be necessary for staff buy-in and feasibility [13]. These findings underscore the need for context-specific adaptation, culturally responsive design, and strong communication strategies.

Implementation barriers include budget constraints, staff capacity, and menu standardization pressures. In multi-faith contexts, inclusivity is improved by vegetarian defaults/options, availability of halal/kosher items where feasible, and transparent labeling of ingredients/allergens. U.S. program guidance and local pilots illustrate workable pathways for accommodating religious needs within child nutrition programs while maintaining nutrition standards and cost controls [23] [24].

The review also highlights the importance of enabling policy environments [12] [20]. Where national guidelines or institutional procurement standards aligned with health and sustainability goals, interventions were more coherent, better resourced, and more likely to be integrated into routine school operations. In contrast, schools operating in policy vacuums faced implementation burdens alone, often relying on committed individuals rather than structural support.

Cross-sectoral collaboration emerged as a key success factor. Interventions that bridged health, agriculture, and education ministries or engaged parents, local food providers, and civil society actors demonstrated greater adaptability and longer-term viability [15] [16] [19]. This aligns with global recommendations calling for whole-of-society and whole-of-school approaches to planetary health.

The whole-school approach emerged as a high-leverage configuration: aligning curriculum, cafeteria standards, school leadership, and family engagement was associated with more persistent dietary improvements and higher acceptability among students and staff, consistent with the Health Promoting Schools framework (Langford et al., 2014/2015) and with ESD whole-school models that link organizational improvement to sustainability outcomes (Mogren et al., 2019). Scaling requires stable funding, leadership mandates, and inter-sectoral coordination across education, health, and agriculture.

Linking school food policy to local agriculture ‒ through farm-to-school procurement, farm visits, and garden-integrated curricula ‒ supports regional economies, increases exposure to seasonal produce, and strengthens students’ willingness to try new foods. Reviews and program evaluations indicate positive trends in knowledge, attitudes, and some behaviors, while calling for more rigorous, long-term designs [25] [27].

Across included studies, GHGE (CO₂ eq.) and food waste were the most commonly reported sustainability indicators, measured via life-cycle assessment and direct plate-waste weighing, respectively; water use, menu-level indices, dietary diversity scores, and affordability appeared less frequently and with heterogeneous methods. This heterogeneity limits comparability across settings and complicates meta-inference for policy. Table 4 consolidates indicators and methods used in the included studies and motivates a minimum standardized set for future evaluations. (See also broader LCA benchmarks in the food system literature.)

• Long-term outcomes. Our analysis revealed limited follow-up beyond one year in most studies; therefore, longitudinal designs with repeated dietary and environmental measures are recommended.

• Affordability and equity. Our analysis indicated sparse cost reporting; therefore, incorporate cost-per-serving and affordability indicators alongside nutrition and waste metrics.

• Indicator standardization. Our analysis showed heterogeneous GHGE and waste methods; therefore, adopt a core indicator set (GHGE via transparent LCA assumptions; direct plate-waste weighing protocols; optional water-use modules) to enable cross-study comparison.

• Cultural/religious inclusivity. Our analysis identified acceptability barriers; therefore, include inclusive menu planning and clear labeling policies within intervention packages.

• Systemic governance. Our analysis confirmed stronger and more durable effects in whole-school models; therefore, prioritize cross-sectoral governance and stable funding to institutionalize best practice (HPS/WSA frameworks).