Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

School-Based Mental Health Promotion and Prevention

Abstract. School-based mental health promotion and prevention investigates the psychological, social, educational, and service mechanisms through which schools support student well-being, reduce risk exposure, identify emerging difficulties, and connect children and adolescents with appropriate care. This fundamental research monograph critically synthesizes randomized school and classroom trials, cluster-randomized preventive programs, systematic reviews, contemporary meta-analyses, implementation research, and WHO, UNESCO, and UNICEF guidance. It differentiates universal, selective, and indicated prevention; emotional competence from clinical symptom reduction; screening from completed care; and school-climate change from treatment. Evidence indicates robust effects on selected social and emotional learning outcomes, generally modest average reductions in internalizing and externalizing symptoms, specific benefits for certain anti-bullying and suicide-prevention interventions, and important null findings for universal mindfulness and digital mental health programs. It examines developmental mechanisms, referral continuity, measurement, costs, equity, implementation, cross-national transferability, and Artificial Era technologies. The synthesis identifies methodological limits and an evidence-governed architecture for school and health systems. This is a critical narrative synthesis of existing research, without original participant data, preregistered systematic-review procedures, or new statistical meta-analysis.

Keywords: school-based mental health; mental health promotion; universal prevention; indicated prevention; school psychology; social and emotional learning; SEL; child and adolescent mental health; depression prevention; anxiety prevention; bullying prevention; suicide prevention; school climate; mental health screening; early identification; referral pathways; cluster randomized trial; whole-school intervention; implementation science; school health services; digital mental health; Artificial Era

School-Based Mental Health as a Population Research Object

School-based mental health promotion and prevention is the systematic study and intentional modification of educational conditions, interpersonal processes, competencies, exposures, and referral arrangements that influence mental health before or alongside the emergence of clinically significant difficulties. Its unit of analysis extends from a child in a classroom to cohorts, school organizations, family–school networks, and public systems. Promotion addresses resources for flourishing, connection, learning, and participation; prevention addresses the incidence, development, severity, or consequences of identifiable mental health problems. A school supplies repeated opportunities for intervention, observation, social participation, and coordinated support, but the causal properties of a program depend on its content, delivery, and surrounding environment. Foundational analyses of school mental health locate educational practice, individual support, and service organization along a continuum requiring the joint attention of education and health systems Fazel et al. (2014a) Weare & Nind (2011) WHO & UNESCO (2021).

The research object encompasses multiple outcomes with different causal meanings: depressive and anxiety symptoms, internalizing and externalizing difficulties, social and emotional competencies, perceived safety, bullying, connectedness, help-seeking, clinical identification, treatment initiation, attendance, academic functioning, and eventual disorder incidence. Improved knowledge about emotions is a proximal learning outcome; reduced depressive symptoms are a psychological outcome; higher referral completion is a service outcome; reduced onset of a diagnosable disorder is a preventive clinical outcome. A report may establish one of these effects without establishing the others. Comparative reviews find meaningful improvements in many social and emotional learning outcomes and considerably smaller average effects on internalizing symptoms; this variation is informative about the mechanisms an intervention actually changes Cipriano et al. (2023) Tsujimoto et al. (2026) Werner-Seidler et al. (2021).

Schools also distribute exposure to stress and protection. Peer rejection, chronic bullying, exclusionary discipline, discrimination, insecure belonging, high-stakes evaluation, disability-related access barriers, and disrupted family–school communication can shape risk; supportive teacher relationships, predictable routines, meaningful participation, and credible referral access can be protective. These pathways interact with conditions outside the school, including poverty, housing insecurity, family stress, neighborhood safety, and availability of treatment. A whole-school framework therefore studies both programs delivered to students and the institutional arrangements that make daily life protective or threatening. The WHO and UNESCO health-promoting-school model connects policy, environment, curriculum, community partnerships, and services; it is a governance framework whose intended benefits must still be evaluated through empirical research WHO & UNESCO (2021) WHO, UNESCO, & UNICEF (2022) Lekamge et al. (2025).

This monograph synthesizes original randomized trials, school-cluster experiments, meta-analyses, implementation research, and official education and health guidance. It is a critical narrative synthesis rather than a registered systematic review or original meta-analysis. Particular attention is given to intervention targets, counterfactual comparators, sampling, measurement, duration, delivery fidelity, and inequities in reach. The inquiry asks what schools can causally change, for whom, at what developmental stage, through which mechanisms, at what opportunity cost, and with which safeguards. It also examines the Artificial Era as a changing context for delivery, assessment, and educational decision support while keeping hypotheses about artificial systems distinct from demonstrated youth mental health outcomes Proctor et al. (2011) Miller & Thabrew (2024) UNESCO, UNICEF, & WHO (2026).

Mental Health Promotion, Prevention, Identification, and Care

Mental health promotion is often operationalized through well-being, perceived competence, positive relationships, school belonging, and a supportive climate. Prevention is conventionally divided into universal activities offered to an entire population, selective activities for groups with established risk characteristics, and indicated activities for individuals with emerging subclinical signs. These categories concern eligibility and timing rather than program quality. A universal curriculum may foster emotion literacy for all students, a selective intervention may address adolescents exposed to a specific stressor, and an indicated group may teach coping skills to students with elevated but subthreshold symptoms. School-based treatment for a diagnosed condition occupies a related but analytically separate clinical service pathway. The distinctions matter because universal delivery generates population reach while targeted delivery concentrates resources among students with greater baseline risk Fazel et al. (2014a) Werner-Seidler et al. (2021) WHO & UNESCO (2021).

Early identification comprises screening, case-finding through observed concerns, confidential self-referral, staff consultation, family reports, and a planned clinical or psychosocial follow-up. It is an intermediate mechanism that can improve access to care, and it must be evaluated according to its actual pathway. The SHIELD randomized trial compared universal high-school depression screening against a usual system of targeted referral; identification and treatment initiation increased in participating schools equipped with student-assistance infrastructure. This trial supports a pathway-specific conclusion about detection and linkage, rather than a demonstrated reduction in long-term depression prevalence or suicide mortality Sekhar et al. (2021).

School mental health services can include school counselors, psychologists, nurses, social workers, education professionals, external clinicians, community-based providers, and emergency pathways. Roles vary substantially by jurisdiction and school resources. A classroom teacher may observe functional changes, provide supportive relationships, and facilitate help-seeking while clinical assessment and treatment decisions require appropriately qualified professionals and local legal safeguards. The 2026 UNESCO–UNICEF–WHO guide emphasizes teaching practices, recognition of distress, and referral within the bounded professional responsibilities of educators. Organizational effectiveness depends on referral acceptance, access, continuity, and accountable follow-through, rather than the mere presence of a screening tool or a named service UNESCO, UNICEF, & WHO (2026) Fazel et al. (2014b) Proctor et al. (2011).

Risk-factor language requires care in educational settings. A child's gender, disability, language background, race, household income, migration experience, or neurodevelopmental profile must never be treated as a diagnosis. Such variables can reveal differential exposure to institutional barriers or unequal access, and their interpretation requires a defensible theory of mechanisms. Decisions based on aggregate group averages should be tested for fairness, subgroup validity, and harmful labeling. Sensitive screening and help-seeking records need privacy safeguards, appropriate consent and assent procedures, access controls, and clear limits on secondary uses. A sustainable prevention system aligns universal climate improvement with responsive, confidential specialist pathways WHO, UNESCO, & UNICEF (2022) Fazel et al. (2014b) Cipriano et al. (2024).

Historical Development of School Mental Health Science

The scientific history of mental health promotion in schools emerged from several intersecting traditions: public hygiene and child welfare, developmental psychology, educational measurement, social learning research, prevention psychiatry, health promotion, and community mental health. Early school health activity focused on physical health and attendance, while the growth of psychology and school guidance made adjustment, behavior, and emotional development explicit educational concerns. Later work began to test classroom management, social competence, and problem-prevention interventions against controls. This history is best understood as a change in research questions—from identifying and managing individual difficulties to experimentally testing modifiable environments and systemwide capacities. Integrative reviews describe the reciprocal influence of educational organization and child mental health rather than attributing the field to any single inventor Weare & Nind (2011) Fazel et al. (2014a).

Longitudinal prevention experiments made developmental timing a central issue. The Baltimore Good Behavior Game began with first- and second-grade classroom interventions during the 1985–1986 school year and followed participants into young adulthood. Its reported benefits for some later behavioral and psychiatric outcomes, particularly among boys with higher early aggressive and disruptive behavior, illustrated the possibility of long pathways from early classroom routines to adult functioning. Those findings depend on the original cohorts, setting, intervention supervision, analytic choices, and replication conditions; they do not establish that every contemporary behavior-management package produces equivalent adult effects Kellam et al. (2008).

By the 2000s and 2010s, randomized school programs increasingly targeted social and emotional learning, depression, anxiety, bullying, violence, and suicidal behavior. A landmark synthesis of 213 universal social and emotional learning studies covering 270,034 students reported favorable average effects on skills, behavior, and academic performance; an independent follow-up synthesis later examined persistence of selected outcomes. The 2015 SEYLE multinational cluster trial tested specific suicide-prevention strategies; the 2018 INCLUSIVE trial evaluated restorative practice and school-environment reform. These lines of work expanded the field from classroom lessons to environmental interventions and rigorous evaluation of particular harms Durlak et al. (2011) Taylor et al. (2017) Wasserman et al. (2015) Bonell et al. (2018).

Subsequent large and sometimes null trials disciplined the field's expectations. The MYRIAD school-based mindfulness study reported no improvement over teaching as usual in its primary mental health outcomes at one year. Later syntheses showed variation by intervention theory, baseline symptom levels, delivery personnel, and outcome domain. These findings helped shift attention toward the question of which mechanisms work under which conditions, rather than declaring schoolwide mental health education uniformly effective. International health-promoting-school guidance developed in parallel, establishing policy and implementation principles while recognizing that organizational adoption and maintenance are independent scientific problems Kuyken et al. (2022) Tsujimoto et al. (2026) WHO et al. (2021).

Developmental, Social, and Institutional Mechanisms

The school environment affects mental health through reciprocal processes linking attention, emotion, social interpretation, motivation, behavior, and perceived control. A student who experiences predictable adult support may interpret stressors as manageable and seek assistance sooner; chronic peer exclusion may alter threat monitoring, reduce belonging, and undermine participation. Such processes are plausible explanatory pathways that require measurement rather than automatic inference from a program label. Contemporary social and emotional learning research measures emotional understanding, self-regulation, relationship skills, decisions, and classroom functioning, while school-climate research examines shared norms, trust, safety, and collective organization. The same classroom activity can affect skills without altering clinical outcomes if other powerful exposures remain unchanged Cipriano et al. (2023) Bonell et al. (2018) Fazel et al. (2014a).

Developmental timing changes intervention opportunity. During early schooling, classroom routines, relational security, peer cooperation, and core emotional vocabulary may be particularly salient. In adolescence, autonomy, peer status, identity, stress appraisal, social media environments, and engagement with adult institutions become influential. These are broad developmental tendencies rather than fixed stages prescribing one program for every age. Mechanistic studies need to test whether program effects change with school stage, baseline symptom severity, family context, neurodevelopmental profile, and the broader educational system. The 2026 pooled analysis of 59 studies, for example, found different patterns for internalizing and externalizing outcomes across school stages and baseline symptom levels Tsujimoto et al. (2026).

Behavioral and cognitive-behavioral approaches offer a set of testable mechanisms: reducing avoidance, increasing adaptive coping, challenging unhelpful expectations, practicing problem solving, arranging reinforcement for prosocial action, and strengthening the relationship between response and feedback. School-based CBT-derived programs have yielded stronger average internalizing symptom reductions than some other approaches in selected syntheses, although differences among studies and delivery settings complicate causal attribution to theoretical ingredients. Clinical inference requires an appropriate trial of the specific program. The rationale for evidence-based skill instruction is strongest when theory, proximal competence measures, symptom outcomes, and implementation measures are aligned Werner-Seidler et al. (2021) Hayes et al. (2025) Tsujimoto et al. (2026).

Ecological mechanisms include leadership decisions, disciplinary rules, safe reporting channels, staff behavior, peer norms, culturally responsive instruction, and links to accessible services. A supportive policy can raise the probability that students disclose distress and that adults respond consistently; a punitive or exclusionary policy can suppress disclosure or magnify stress. The INCLUSIVE Learning Together trial used school action groups, restorative practice, and a social and emotional curriculum, demonstrating a small improvement in bullying but no detectable effect on its separate aggression endpoint. This contrast shows that even within one organizational intervention, different causal outcomes require separate evidence Bonell et al. (2018) WHO & UNESCO (2021) Lekamge et al. (2025).

Research Designs, Outcomes, and Scientific Inference

Cluster-randomized trials are central because school interventions usually operate through groups. Randomizing entire schools or classrooms can reduce within-school contamination and permit evaluation of organizational change. Statistical analysis must account for correlations among students who share teachers, policies, and social environments. Recruitment before or after randomization, differential attrition, teacher and student masking, baseline imbalance, migration between schools, and variation in implementation can change the credibility of an effect estimate. The experimental treatment is also defined by its comparator: instruction as usual may already include substantial social and emotional programming. MYRIAD's null comparison is therefore a result about an added intervention in particular schools, not a universal test of whether emotional support matters Kuyken et al. (2022) Bonell et al. (2018).

Meaningful outcomes should be prespecified at several levels. Proximal indicators include knowledge, coping skills, self-efficacy, staff confidence, and perceived belonging. Intermediate indicators include behavior, repeated symptom scores, help-seeking, access, referral completion, and attendance. Distal indicators include persistent functional improvement, reduced incident disorder, serious harm, and educational trajectories. Differences in outcome timing make direct pooling difficult. A program may raise detection initially, apparently increasing recorded distress, while improving connection to care. A lower symptom average may reflect short-lived relief with uncertain prevention of future diagnosis. Reviews should disclose reporter source, scale validity, minimal important difference where established, and whether an outcome is a clinical endpoint or a surrogate Sekhar et al. (2021) Werner-Seidler et al. (2021) Proctor et al. (2011).

Psychometric validity is not a decorative technical concern. Student symptom questionnaires, teacher reports, administrative incidents, attendance measures, and school climate scales capture partly different constructs and can disagree. Observer expectancy, classroom norms around disclosure, changing language of distress, differential response styles, and nonresponse may distort apparent prevalence. A reduction in disciplinary referrals could represent fewer incidents, changed staff thresholds, or altered recording. For suicide-related outcomes, incidents are uncommon relative to the size of a typical school trial, which requires careful statistical power and transparent follow-up. The SEYLE trial distinguished self-reported new attempts and severe ideation at twelve months from outcomes recorded at three months and from deaths, none of which occurred during the study Wasserman et al. (2015) Bonell et al. (2018).

Evidence synthesis must distinguish randomized effects, plausible mechanisms, correlations, qualitative implementation findings, and guideline recommendations. Random-effects meta-analysis estimates an average across specified designs and outcomes; prediction intervals, heterogeneity, risk of bias, and subgroup multiplicity determine how transferable that mean may be. Publication bias and selective outcome reporting can exaggerate favorable results, while underpowered studies may yield misleading nulls. The 2026 meta-analysis provides a helpful contemporary benchmark for small average symptom effects, and the 2025 review of whole-school programs identifies endpoint-specific benefits accompanied by numerous inconclusive outcomes. The synthesis in this monograph retains these differences instead of constructing a single global 'effective' label Tsujimoto et al. (2026) Lekamge et al. (2025) Hayes et al. (2025).

Universal Prevention of Depression and Anxiety

Universal school-based prevention programs seek to change modifiable psychological vulnerabilities and protective competencies before a clinical disorder becomes established. They frequently teach behavioral activation, coping, problem solving, cognitive restructuring, emotional awareness, and adaptive responses to worry or avoidance. Targeted programs additionally recruit students with elevated symptoms or risk exposures. Eligibility and theoretical content are independent dimensions: a CBT-based intervention can be universal or indicated, and a universal curriculum may contain both cognitive-behavioral and relational components. The central causal question is whether assignment to a program improves prospectively measured symptom trajectories or reduces disorder onset compared with the educational practices available to controls. Improvements in knowledge or attitudes cannot substitute for those outcomes Werner-Seidler et al. (2017) Werner-Seidler et al. (2021) Fazel et al. (2014a).

The updated systematic review by Werner-Seidler and colleagues assembled 118 randomized trials involving more than 45,000 participants and concluded that school-based psychological prevention produced small reductions in depression and anxiety symptoms. The earlier 2017 synthesis likewise found small effects and examined six- and twelve-month follow-up, with targeted approaches at times outperforming universal alternatives. Such patterns are consistent with baseline risk moderating the possible absolute benefit of prevention. Comparisons must consider how studies recruited students, whether control groups received other interventions, and how symptom severity and diagnostic outcomes were operationalized. An effect on a screening scale is not equivalent to demonstrated prevention of major depressive disorder Werner-Seidler et al. (2017) Werner-Seidler et al. (2021).

A 2025 review of universal interventions for emotional outcomes included 71 studies and 63,041 young people. Pooled effects were small for self-reported anxiety (d = −0.0858, 95% CI −0.15 to −0.02) and depression (d = −0.109, 95% CI −0.19 to −0.03); the authors did not find an analogous overall effect on broader internalizing outcomes. The evidence also suggested variation by intervention theory, including stronger anxiety reductions for CBT-derived approaches within the comparisons studied. These estimates should be understood as outcome-specific standardized differences rather than direct changes in the probability of later psychiatric diagnosis. They aggregate diverse countries, educational stages, programs, and measurement practices Hayes et al. (2025).

The 2026 meta-analysis led by Tsujimoto aggregated 59 studies involving 83,092 students across 22 countries. It estimated marginal mean reductions in internalizing symptoms (SMD −0.12, 95% CI −0.19 to −0.06) and externalizing symptoms (−0.08, 95% CI −0.14 to −0.02). Stronger internalizing effects were observed in subgroups with elevated baseline symptoms and among interventions using CBT or mental health practitioners, but these are moderator associations across studies rather than randomized head-to-head assignments of all components. The practical implication is a differentiated prevention design: match outcome, risk distribution, personnel, and follow-up duration to the mechanism being tested. Population effects remain valuable when reach is high, provided program cost and opportunity costs are also measured Tsujimoto et al. (2026).

Social and Emotional Learning and Prosocial Competence

Social and emotional learning (SEL) comprises organized educational experiences intended to develop recognition and regulation of emotion, interpersonal understanding, prosocial behavior, constructive decision-making, and school participation. SEL programs vary in theory and pedagogy, ranging from structured lessons and role play to integrated classroom routines and school culture interventions. Their most immediate scientific outcomes are the taught competencies; later outcomes can include peer relationships, behavioral adjustment, school safety, and academic engagement. To study mental health prevention, researchers must establish how these competencies relate to subsequent distress and functioning rather than treating social skill acquisition as a clinical result. SEL is thus a mechanism-rich but heterogeneous research family Durlak et al. (2011) Cipriano et al. (2023).

The influential Durlak and colleagues synthesis covered 213 universal school-based interventions with 270,034 students and documented improvements across social and emotional skills, attitudes, behaviors, and achievement, including an average academic performance advantage described as eleven percentile points. This historical estimate belongs to that particular intervention and outcome corpus; it is neither an expectation for each school adopting SEL nor a measure of prevented mental disorders. Delivery problems and recommended skill-development practices moderated outcomes. The value of the synthesis lies in its demonstration that systematic school instruction can change several domains simultaneously and that program implementation is causally important Durlak et al. (2011).

Taylor and colleagues examined 82 universal SEL programs involving 97,406 students with follow-up outcomes spanning from six months to eighteen years. Their pooled follow-up evidence supported continued advantages in social and emotional competencies and selected later outcomes. The extended follow-up windows increase developmental relevance but introduce variation in participant retention, data availability, program histories, and intervening circumstances. Later-life outcomes associated with an intervention deserve attention, but a causal pathway over many years should be interpreted against the design and original randomization of each underlying study Taylor et al. (2017).

A contemporary registered-report meta-analysis by Cipriano and colleagues synthesized 424 studies from 53 countries, encompassing 252 interventions and 575,361 students. It identified favorable average outcomes across skills, school functioning, relationships, climate and safety, while finding substantial moderation by program characteristics, context, and implementation quality. A separate 2024 U.S.-focused analysis examined marginalized students and underscored the importance of sampling and contextual relevance. These findings support SEL as a family of experimentally studied educational strategies while requiring clarity about which domains improve and whether a particular curriculum generalizes across language, disability, racial, socioeconomic, and cultural settings Cipriano et al. (2023) Cipriano et al. (2024).

Mindfulness-Based Programs and the Importance of Null Trials

Mindfulness-based school interventions usually involve attentional practices, exercises in awareness of internal events, and explicit instruction about responding to thoughts and emotions. Their proposed mechanisms include attentional regulation, reduced automatic reactivity, self-observation, and a more flexible relation to stressful experiences. Many programs are delivered universally because classroom formats permit substantial reach. Mechanistic plausibility does not establish population mental health benefit: a student may learn or enjoy the exercises without showing a measurable change in depressive risk, well-being, or social-emotional functioning. Trials must evaluate the whole educational package and specify the comparator, the treatment duration, and the age and context of participants Kuyken et al. (2022) Hayes et al. (2025).

The MYRIAD cluster-randomized study assessed school-based mindfulness training against teaching as usual in 84 schools and 8,376 adolescent participants in the primary analysis. At one year it found no evidence that adding the mindfulness curriculum improved risk for depression, social-emotional-behavioral functioning, or well-being. The reported standardized mean difference in risk for depression was 0.005, with a 95% confidence interval from −0.05 to 0.06. This carefully powered result places meaningful limits on claims that a universal mindfulness curriculum necessarily improves adolescent mental health. It does not negate the possibility of benefits in selected populations or under alternative delivery arrangements, which would need their own tests Kuyken et al. (2022).

Comparative evidence also urges separation of the act of practicing mindfulness from general supportive classroom structure. Children may benefit from scheduled reflective time, respectful adult attention, or a favorable class climate regardless of the nominal theoretical label. If the treatment-as-usual comparator already contains SEL activities, the estimate measures the incremental addition of mindfulness. Differential adherence, students' initial symptom severity, developmental readiness, teacher competence, and opportunity costs can determine the result. The MYRIAD findings and the 2025 emotional-outcomes synthesis provide reasons to prioritize transparent comparative trials before expanding a specific protocol across schools Kuyken et al. (2022) Hayes et al. (2025) Tsujimoto et al. (2026).

Bullying Prevention, Peer Safety, and School Climate

Bullying involves repeated harmful behavior within a power imbalance, and its prevention is an essential school mental health issue because victimization, peer exclusion, fear, and disrupted belonging can affect emotional well-being and educational participation. Causal interventions may target peer norms, student bystander behavior, adult monitoring, accessible reporting, restorative processes, and institutional accountability. Successful anti-bullying programs should be assessed through multiple reporters and behaviors rather than a single disciplinary record. A decline in reported incidents can signal real improvement or a reduced willingness to disclose, making climate and trust measures essential interpretive companions Kärnä et al. (2011) Bonell et al. (2018).

The Finnish KiVa program provided an important school-cluster randomized evaluation. In the study published by Kärnä and colleagues, 78 schools with 8,237 students in grades four through six were allocated to intervention or control conditions. The trial investigated school bullying and victimization outcomes and demonstrated that organized, theoretically specified intervention at the peer-group level can be evaluated rigorously in normal educational environments. Its national context, school-stage concentration, and program-specific training should remain visible when comparing it with other initiatives Kärnä et al. (2011).

The English INCLUSIVE trial tested Learning Together in 40 secondary schools with 6,667 students responding at baseline. The intervention combined staff training in restorative practice, a student-staff action group revising school policies, and social-emotional instruction. At 36 months the analysis showed a small improvement in bullying victimization, with an adjusted effect size of −0.08, while the separate aggression endpoint showed no convincing change. Such results illustrate multidimensional outcomes in institutional prevention: modification of perceived peer harm can coexist with negligible change in another behavioral index. They also demonstrate the importance of retaining null findings when reporting the effects of school reform Bonell et al. (2018).

A 2025 systematic review of whole-school programs found reduced odds of cyberbullying, cyberaggression, and regular smoking across eligible randomized trials, but did not find statistically reliable overall effects for many outcomes including anxiety, depression, and positive mental health. The finding is consistent with the proposition that school-environment interventions may change particular risk behaviors more readily than broad emotional states. It is not a mandate to privilege one outcome: it requires a model linking policies and peer relations to theoretically relevant endpoints, while preserving measures of psychological distress and help-seeking as separate questions Lekamge et al. (2025).

Suicide Prevention, Help-Seeking, and Safety Pathways

Suicidal ideation, nonfatal attempts, suicide deaths, and non-suicidal self-injury are distinct outcomes, each requiring specific definitions, safeguards, and study designs. School prevention can involve mental health literacy, recognition and response training, peer support norms, confidential routes to qualified services, and structured referral and emergency response. Evaluating a program solely by expressed willingness to seek help is inadequate when the intended clinical outcome concerns attempts or severe ideation. The scientific and ethical design must also anticipate immediate disclosure of risk during research or routine school activity and provide a feasible response plan Wasserman et al. (2015) WHO, UNESCO, & UNICEF (2022) UNESCO, UNICEF, & WHO (2026).

The multicountry SEYLE study randomly assigned 168 schools enrolling 11,110 students to a gatekeeper training program, Youth Aware of Mental Health (YAM), professional screening, or control. At the three-month follow-up the study did not identify significant differences between arms. At twelve months, YAM was associated with lower reported incidence of suicide attempts (odds ratio 0.45, 95% CI 0.24–0.85) and severe suicidal ideation (odds ratio 0.50, 95% CI 0.27–0.92) relative to controls among the trial's specified analysis population. No participant died by suicide during the study. These estimates support the particular YAM intervention under trial conditions and do not make all suicide-awareness lessons equivalent Wasserman et al. (2015).

Interpretation of suicide prevention outcomes requires uncommon-event methods and scrutiny of exclusions, missing follow-up, variation in disclosure, and differential access to rescue care. Schools are responsible for safeguarding students who disclose acute risk, including those assigned to control arms. Research on ideation can be affected by changes in willingness to report rather than changes in experience. Moreover, a reduction in a self-reported attempt endpoint should never be restated as an observed reduction in suicide deaths. Distinct harm categories and adequate referral capacity are essential to responsible dissemination of the SEYLE findings Wasserman et al. (2015) Sekhar et al. (2021).

A strong school suicide-prevention pathway links universal literacy and safe help-seeking expectations to trained adult response, confidential assessment, family involvement where appropriate, coordination with specialist services, and documented continuity following referral. Program evaluation should report the completeness of this pathway, delays to care, equitable access, severe events and potential adverse effects. Because risk management is a clinical and legal responsibility, instructions for staff must follow local policy and qualified professional practice rather than an unsupervised curriculum or algorithmic triage. Official guidance describes teacher support and referral as meaningful, bounded functions in a wider care network UNESCO, UNICEF, & WHO (2026) WHO (2026).

Physical Activity, Health Behavior, and Well-Being

School health promotion can integrate opportunities for movement, rest, supportive nutrition settings, access to safe play, and reduced sedentary pressure. These conditions may support physical health, peer belonging, autonomy, and emotional functioning, but the strength of evidence varies by outcome. Evaluations of physical activity must distinguish a successful increase in movement from a subsequent improvement in depressive or anxiety symptoms. Program mechanisms also depend on inclusion: unsafe competitive settings or disability-related barriers can transform an apparently beneficial curriculum into an unequal experience. The health-promoting-school framework encourages coherent policies and accessible participation without substituting general plausibility for specific effect estimates WHO & UNESCO (2021) Ditzen-Janotta et al. (2026).

A 2026 review of universal school-based physical activity interventions in high-income countries included 25 studies, with 19 entering meta-analyses of mental health or quality-of-life outcomes. The authors reported no statistically significant pooled benefits for the assessed symptom groups, while a positive pooled finding for health-related quality of life was graded as very-low-certainty evidence. The review identified high risk of bias, heterogeneous interventions, and inconsistent measurements as major limitations. Thus movement may be valuable within general school health policy while evidence for a specific universal mental disorder prevention effect remains uncertain Ditzen-Janotta et al. (2026).

The appropriate integrative model does not place exercise, SEL, bullying prevention, and clinical referral into a single interchangeable category. They address overlapping conditions through different causal pathways and have different direct outcomes. A school choosing interventions should use local needs assessment, equity analysis, evidence about delivery feasibility, and planned evaluation; it should retain both health and learning goals in view. Physical activity provides a useful test case for a broader scientific rule: interventions can be worthwhile for one established benefit while lacking convincing evidence for another, and transparent reporting preserves that distinction Ditzen-Janotta et al. (2026) Lekamge et al. (2025) Proctor et al. (2011).

Selective and Indicated Prevention Across Risk Groups

Selective prevention recruits groups with documented vulnerability or elevated exposure, whereas indicated prevention enrolls individuals with early symptoms or other subclinical indicators. In school settings, the population could be defined by recent displacement, chronic peer victimization, household adversity, a transition to a new educational environment, or repeated symptom scores. Each selection rule must be examined for its predictive validity and for potential stigmatization or unequal access. A child identified as exposed to adversity may need practical educational accommodation or family support rather than a psychological intervention. Prevention is most defensible when eligibility, the hypothesized risk mechanism, intended benefit, and informed participation are specified in advance Werner-Seidler et al. (2017) Werner-Seidler et al. (2021) Fazel et al. (2014b).

Targeting can raise the absolute benefit of an intervention because baseline symptoms offer more room for change, but it can also introduce regression to the mean, screening error, and selection effects. A program evaluated only among students with elevated anxiety at baseline must demonstrate improvement relative to similarly selected controls. The 2026 meta-analysis found stronger average internalizing symptom effects in groups with elevated initial symptoms, while the 2021 review of school depression and anxiety programs documented variation by universal versus targeted delivery. These findings support investigating selection strategies; they do not validate every risk algorithm or categorical cut-off Tsujimoto et al. (2026) Werner-Seidler et al. (2021).

Stepped support is an organizational response to this heterogeneity. The whole student population can receive safe climate and accessible social-emotional learning, students with emerging difficulties can receive time-limited preventive skill instruction, and students with clinically significant distress can enter a qualified assessment and treatment pathway. Movement between tiers should reflect changing need, informed choice, and observed response. A tier is a service allocation decision rather than a diagnosis; a placement should be reviewable and reversible. Evaluation should record both psychological changes and service consequences such as waiting times, continuity, and student experience Fazel et al. (2014a) Proctor et al. (2011) WHO (2026).

Research on selective and indicated programs requires attention to recruitment consent, who is screened, language accessibility, safeguarding, and the burden on families and practitioners. A trial may appear effective if intensive specialist attention is delivered only to intervention participants; the resulting comparison evaluates the combined package, including attention and accessibility, rather than the teaching materials alone. Implementation outside a trial can dilute outcomes if counselors are unavailable or referral thresholds are inconsistent. Therefore, targeted prevention should be interpreted through a complete mechanism involving identification, engagement, appropriate delivery, and sustained follow-up Werner-Seidler et al. (2021) Sekhar et al. (2021) UNESCO, UNICEF, & WHO (2026).

Universal Screening, Assessment, and Referral Completion

Universal screening is a deliberate program of asking students about symptoms or functioning regardless of whether staff have noticed a problem. Its justification rests on an expected benefit from recognizing otherwise missed difficulties and connecting students to help. Screening generates multiple possible harms and costs: false positives, false negatives, distress about disclosure, privacy breaches, labeling, time devoted to follow-up, and service backlogs. Screening therefore must be evaluated as an entire pathway from invitation and instrument selection through triage, clinical assessment when indicated, referral and service initiation. A psychometric questionnaire alone cannot establish a diagnosis, and institutional capacity is a scientific prerequisite for assessing usefulness Sekhar et al. (2021) WHO (2026).

The SHIELD trial provides unusually direct evidence on the distinction between detection and treatment access. It randomized 12,909 adolescents in 14 Pennsylvania high schools to universal depression screening with the PHQ-9 or targeted referral based on behavioral concern. Universal screening participants had higher odds of being identified with symptoms compatible with major depressive disorder (OR 5.92, 95% CI 5.07–6.93), of the Student Assistance Program confirming follow-up need (OR 3.30, 95% CI 2.49–4.38), and of initiating treatment or other services (OR 2.07, 95% CI 1.39–3.10) during the school year. These are compelling pathway outcomes in a system with existing assistance teams; the experiment does not establish improved long-term disorder incidence or suicide mortality Sekhar et al. (2021).

Even well-validated scales can perform differently across age, language, disability, social background, and a changing school environment. Predictive values depend on prevalence and on the criterion outcome; a high sensitivity screening threshold can increase referrals without increasing the number of students who obtain a clinically useful service. Repeated surveys may affect disclosure norms or produce fatigue. A strong evaluation reports who was invited and declined, completion rates, score thresholds, assessment accuracy, adverse experiences, onward referrals, service acceptance, time to initial contact, and the eventual student-defined benefit. A false sense of protection is particularly hazardous when screening is adopted without sufficient clinical response capacity Sekhar et al. (2021) Fazel et al. (2014a) WHO, UNESCO, & UNICEF (2022).

Data governance is integral to screening science. The school must specify who can access a record, what the assessment is for, how long data are retained, how families and students are informed, and what circumstances trigger safeguarding procedures. Research comparing screening pathways must describe jurisdictional consent and reporting requirements without presuming one country's privacy law applies everywhere. The appropriate role of teachers is to support, recognize concerns, and connect students with qualified resources; clinical interpretation and management of elevated risk require trained personnel and coordinated policies. A transparent pathway protects students who decline elective participation while preserving emergency responses for imminent threats UNESCO, UNICEF, & WHO (2026) WHO (2026) Sekhar et al. (2021).

School-Based Services, Clinical Interfaces, and Continuity of Care

School-based mental health services form an interface between educational responsibility and clinical health care. Students may encounter a school psychologist, counselor, nurse, social worker, trained teacher, community provider, or visiting clinician. These roles have distinct qualifications and legal scopes, which differ across countries and jurisdictions. Mental health promotion can be integrated into ordinary teaching, while clinical diagnosis, complex risk assessment, and treatment planning belong to appropriately regulated professionals. The scientific question is not whether every school operates a clinic; it is whether its combination of preventive educational practices and external partnerships provides timely, equitable, effective access to care when need arises Fazel et al. (2014a) Fazel et al. (2014b) WHO (2026).

Continuity can be described as a sequence of observable events: concern recognized, student engaged, appropriate consent obtained, assessment completed, referral received, service initiated, follow-up sustained, and school accommodations adjusted when needed. Each transition can fail. A rise in referral numbers is of limited value if students encounter prohibitive cost, transportation barriers, stigma, long waiting lists, or fragmented communication among agencies. Implementation outcomes including reach, adoption, fidelity, acceptability, appropriateness, feasibility, cost and sustainability help explain why a clinically sensible pathway can deliver modest population benefit Proctor et al. (2011) Sekhar et al. (2021).

Integration must preserve the educational right to learn and the student's control over sensitive information within legal limits. Clinicians can recommend educational adjustments; schools can document attendance or functioning; families can contribute developmental and contextual knowledge. Information sharing should be proportionate, consent-based where required, and aligned with safeguarding responsibilities. Where specialist supply is sparse, schools may prioritize psychological first response, accessible support and facilitated referrals rather than pretending educational staff can provide complex psychotherapy. The 2026 school health services implementation guidance and the Teaching with Care framework explicitly treat school–health collaboration as an operational component of promotion and protection WHO (2026) UNESCO, UNICEF, & WHO (2026).

Outcomes for school-based services must include those most relevant to students and families: symptom reduction, reliable functional improvement, satisfaction and trust, school engagement, reduced harm, and avoidance of unnecessary restriction or labeling. Treatment initiation is a useful access milestone but should be followed, when possible, by process quality and long-term patient-relevant outcomes. Comparative trials can test embedded providers, telehealth, external referral coordination, or student support teams, but choice of comparator and available local health infrastructure will strongly influence estimates. Whole-school prevention and indicated clinical services should reinforce one another through explicit, evidence-tested connections Fazel et al. (2014a) Miller & Thabrew (2024) Proctor et al. (2011).

Teachers, Staff Well-Being, and Professional Boundaries

Teachers influence psychological safety through classroom routines, instructional fairness, non-stigmatizing communication, relationship quality, recognition of emerging difficulty, and the ease with which students can ask for assistance. Their actions are part of the school's relational environment and may affect the usefulness of even well-designed specialist services. Yet promoting mental health through education cannot be built on silently transferring clinical responsibility to already overloaded staff. Staff require practical guidance, realistic workloads, consultation, protected time, and clear pathways to qualified specialists. The jointly published 2026 UNESCO–UNICEF–WHO Teaching with Care guidance frames this role as supportive, evidence-informed and bounded rather than diagnostic or therapeutic UNESCO, UNICEF, & WHO (2026).

Teacher well-being is simultaneously an organizational outcome and a plausible mediator of classroom climate. Chronic overload, exposure to repeated crises, inadequate staffing, organizational conflict, and poor disciplinary support can impair adults' capacity to provide predictable and responsive relationships. A teacher mental health program should therefore be evaluated beyond self-reported resilience: working conditions, resource allocation, sustained staffing, accessibility of consultation, student experience and educational continuity also matter. Individual coping practices cannot compensate for unchanged organizational stressors when those stressors are central causal exposures. School policy and staff development need to be assessed together WHO et al. (2021) Weare & Nind (2011) UNESCO, UNICEF, & WHO (2026).

Professional development can teach staff to recognize changes in attendance, concentration, social withdrawal, mood, sleep-related fatigue, peer conflict and behavior, while emphasizing that observations are starting points for supportive inquiry rather than diagnostic evidence. The benefit of gatekeeper programs depends on safe disclosure, timely clinical response and the avoidance of blanket labels. In suicide prevention, teachers should not be expected to perform specialist risk assessment from checklists or to manage imminent danger without institutional procedures. The SEYLE trial compared gatekeeper training with alternative approaches and found a specific 12-month benefit for the student-directed YAM arm; its arms cannot be collapsed into one generic category of 'mental health awareness' Wasserman et al. (2015) UNESCO, UNICEF, & WHO (2026).

Research should document the actual burden of implementation: lesson time, required preparation, supervision, staffing turnover, fidelity checks, crisis response and costs carried by students and families. A program may produce a small mean symptom advantage yet require intensive resources that could produce larger effects if directed toward specialist care or school environmental reforms. Conversely, modest additional teacher guidance may support many students when embedded in routines without displacing essential instruction. Transparent economic and implementation measurement helps public systems compare plausible alternatives Proctor et al. (2011) Kuyken et al. (2022) Fazel et al. (2014a).

Families, Communities, and Cross-System Coordination

A student belongs to multiple relational contexts, including family, peer networks, educational institutions, community organizations and health services. School-based mental health interventions can influence students through school experience, but some major causes and protections lie beyond school boundaries. Family socioeconomic stress, parental mental health, housing instability, violence, disability access, displacement and opportunities for supportive participation may affect both baseline risk and program response. A scientifically adequate framework therefore specifies which exposures are changed by the school, which are addressed through coordinated services, and which require broader policy. This is an ecological account of causes and feasibility rather than an invitation to treat individual families as the source of every problem Fazel et al. (2014b) WHO, UNESCO, & UNICEF (2022) WHO & UNESCO (2021).

Effective family partnership requires culturally intelligible communication, opportunities for genuine participation, understandable privacy protections, and routes to raise concerns. Schools should avoid interpreting difficulty attending meetings or responding to digital messages as evidence of parental indifference; working hours, language, transport, caregiving and prior discrimination may explain such patterns. When students seek confidential support, the appropriate role of family involvement follows developmental needs, law and clinical safeguards, including attention to possible family-related harm. Research can measure the reach of family contacts and whether a partnership improves service attendance or functioning, but family engagement is not a unitary intervention with a single expected effect WHO, UNESCO, & UNICEF (2022) UNESCO, UNICEF, & WHO (2026).

Community coordination can connect universal school efforts to youth services, mental health care, child protection, housing assistance, disability advocacy, culturally specific organizations and recreational supports. Coordination quality can be tested through referral timeliness, acceptance, continuity, student experience, unmet need and recovery of educational participation. The evidence base is less complete for many complex system packages than for discrete classroom curricula, partly because the institutional intervention is difficult to randomize and its components change together. Implementation frameworks offer transparent descriptions of mechanism and outcomes even when evidence about distal population effects remains developing Proctor et al. (2011) Fazel et al. (2014b) WHO (2026).

Cross-system planning becomes particularly important during transitions: school entry, moving between schools, transition to secondary education, prolonged absence, discharge from clinical services and the move out of compulsory schooling. Each transition can interrupt relationships and treatment, change eligibility, and alter access to accommodations. Continuity should be assessed using actual student trajectories, not merely records of referrals sent. The scientific priority is to connect educational belonging, preventive learning, and appropriate clinical access across time without allowing administrative boundaries to define the limits of needed support Fazel et al. (2014a) WHO, UNESCO, & UNICEF (2022) WHO (2026).

Disability, Culture, Equity, and Universal Access

School mental health evidence must be interpreted in relation to unequal experiences of the educational institution. Students with disabilities, neurodevelopmental differences, sensory or communication barriers, language needs, financial insecurity, racialized exclusion, migration experiences or minoritized cultural identities can face different exposures, barriers and measurement errors. Equity in promotion and prevention means assessing whether all students can participate safely and derive meaningful benefit, and whether services address need without imposing discriminatory categories. A universal program delivered in one language or requiring a particular communication style may be formally offered to everyone while practically excluding some students Cipriano et al. (2024) Fazel et al. (2014b) WHO, UNESCO, & UNICEF (2022).

Cultural adaptation is an empirical question about meaning, uptake, mechanism and outcomes. Local idioms of distress and norms around disclosure can shape responses to questionnaires; a standardized scale may require careful translation and validation before international comparison. Programs built around a narrow conception of appropriate social expression can mistake difference for dysfunction. Research should use accessible measures and qualitative feedback where appropriate, prespecify meaningful subgroup hypotheses and avoid claiming effects for poorly represented populations. The multinational reach of the SEL literature provides breadth, while its heterogeneity demands explicit attention to cultural and program context Cipriano et al. (2023) Cipriano et al. (2024) Tsujimoto et al. (2026).

Disability-related needs may require adapted teaching, quiet or sensory-accessible spaces, alternative communication, anti-bullying safeguards, and individualized educational supports. These provisions can reduce stress through access and belonging even when they are not marketed as mental health interventions. Outcome measurement should separate inclusion, participation, symptoms and functioning; improvement in one domain should not automatically be inferred from another. Specialist clinical support may be necessary for some students, but participation in ordinary school life and the capacity to exercise autonomy remain independently important outcomes WHO & UNESCO (2021) UNESCO, UNICEF, & WHO (2026) Fazel et al. (2014a).

Equity analysis also concerns who is referred, who is screened, whose distress is interpreted as misconduct, and whose families can access the response pathway. Aggregate performance can hide widening disparities if gains accrue mainly to students in well-resourced settings. Future trials and implementation cohorts should compare reach, adoption, consent processes, false-positive burdens, severity-adjusted access and long-term outcomes across relevant student groups, with adequate statistical power. The 2024 SEL evidence focused on marginalized students illustrates a need to treat representativeness and distribution of benefit as core scientific questions rather than secondary reporting preferences Cipriano et al. (2024) Sekhar et al. (2021) Proctor et al. (2011).

Whole-School Organization and Health-Promoting Institutions

A whole-school approach treats school mental health as an emergent property of everyday policy, physical and social environment, curricular practice, staff relationships, school leadership, family contact, and services. Its research hypothesis is that coordinated changes in several interacting settings can improve safety, emotional development and participation more sustainably than isolated lessons. This is a hypothesis requiring endpoint-specific evidence; a whole-school label does not specify one reproducible treatment. The WHO and UNESCO global standards describe governance, curriculum, environment, partnerships and health services as complementary elements of the health-promoting-school model. A scientifically useful implementation must translate this broad framework into explicit components, responsible actors, measurable fidelity and comparable outcomes WHO & UNESCO (2021) WHO et al. (2021).

Whole-school programs can modify bullying procedures, climate assessment, student voice, restorative processes, staff development, information sharing, and access to psychosocial support. Their effects may be produced by improved peer norms, credible reporting, adult responses, or changes in opportunity for harm. The INCLUSIVE trial demonstrated that an organizational package could reduce bullying while leaving a separate aggression endpoint without convincing improvement. More recently, the 2025 whole-school systematic review reported benefits for selected risk outcomes but inconclusive effects for many broad emotional outcomes. Consequently, adopting a broad governance model and demonstrating benefits of a specific package are related yet separate scientific events Bonell et al. (2018) Lekamge et al. (2025).

Institutions modify program content through timetable constraints, teacher turnover, disciplinary structures, leadership priorities, infrastructure and referral access. A prevention lesson introduced into a school with recurring humiliation or untreated bullying may be inconsistent with students' daily experience, weakening credibility or changing its meaning. Organizational change may also produce benefits that symptom inventories fail to capture, such as more trustworthy adult relationships or safer participation. Rigorous studies should collect school-climate and behavior outcomes alongside psychological symptom measures and report possible adverse effects of disciplinary or screening changes Weare & Nind (2011) Bonell et al. (2018) WHO et al. (2021).

Implementation becomes the object of study when a district attempts scale-up. Sustainable adoption requires school-level governance, clearly specified responsibilities, professional support, resources for referral, recurring review of student experience, and capacity to adapt content without eliminating evidence-based mechanisms. Proctor and colleagues' implementation outcomes provide a vocabulary for distinguishing adoption, acceptability, appropriateness, fidelity, penetration, cost and sustainability from student-level clinical effectiveness. A school may achieve high curriculum adoption yet still have weak service coverage or poor student outcomes; faithful reporting of both levels is essential Proctor et al. (2011) WHO et al. (2021).

Implementation Science, Fidelity, and Program Adaptation

The observed effect of a mental health intervention is partly an effect of its delivery system. Curriculum dosage, instructor training, peer facilitation, staff supervision, student attendance, adaptation, resource availability and competing school demands all influence exposure. A program that succeeds under a research team's continuous support can lose effectiveness when delivered without those supports, whereas a carefully adapted program can improve participation by fitting the community. Evaluation must therefore record core components and planned adaptations, rather than using a simple fidelity score as a substitute for causal explanation. Early evidence on universal SEL found implementation quality associated with intervention outcomes; modern reviews retain program and contextual heterogeneity as central findings Durlak et al. (2011) Cipriano et al. (2023) Proctor et al. (2011).

Distinguishing intention-to-treat and per-protocol effects prevents a common interpretive error. Randomized assignment protects the comparison for the offered intervention, even when participation varies. An analysis restricted to highly engaged students may become confounded by motivation, prior distress, attendance, or family support, factors that also influence outcomes. Dose–response associations can aid mechanism testing, but require careful causal assumptions and measurement. Cluster designs add the risk of contamination across schools and an institutional dependence structure that should be reflected in randomization and analysis. The large MYRIAD trial illustrates that an increment to existing school provision can produce a null intention-to-treat result despite the conceptual appeal of its component practices Kuyken et al. (2022) Bonell et al. (2018).

Adaptation should be distinguished from loss of program identity. Translation, accessible communication, trauma-sensitive scheduling, and adjustments for neurodivergent students may be essential for fair participation. Removing structured practice, trained supervision, or the key exposure pathway can transform the tested intervention into something else. Implementation research is strongest when it reports which components were changed, why, by whom, and whether adapted students experienced comparable benefits or harms. Cross-national syntheses suggest average effects across many versions of school programs, but such pooling does not identify a universal implementation recipe Cipriano et al. (2024) Tsujimoto et al. (2026) Proctor et al. (2011).

Sustainability requires longitudinal organizational follow-up, not solely a student survey immediately after a pilot. A school may preserve a program name while teachers change and dosage declines. Alternatively, a short grant can support initial referral uptake that later collapses when clinical partners withdraw. District evaluation should monitor reach among newly enrolled students, workforce training continuity, administrative resources, costs and equity of referral completion. These measurements specify whether prevention has become a functioning public institution or remains a temporary local experiment WHO et al. (2021) Proctor et al. (2011) WHO (2026).

Economic Evaluation, Opportunity Cost, and Policy Decisions

School prevention consumes educational time, staff capacity, supervision, materials, space and potentially clinical resources. Economic evaluation asks how these costs relate to benefits under realistic alternatives. A curriculum may create small average symptom advantages across a large population, while a targeted clinical pathway creates larger benefits among fewer students at higher individual cost. These choices involve different outcomes, distributional values, and time horizons. Short-term reductions on self-report scales cannot be treated as directly monetizable reductions in lifetime illness without a validated model, and cost-effectiveness estimated in one country's service system may change substantially elsewhere Kuyken et al. (2022) Proctor et al. (2011).

The MYRIAD study explicitly included a cost-effectiveness component alongside its school-cluster randomization, making it an important reminder that incremental benefit matters when comparing universal additions to existing school provision. If no reliable improvement is detected over a credible comparator, a positive program story alone does not establish a favorable cost-effectiveness conclusion. Research should present costs from a stated perspective—school, health system, household, or society—and distinguish direct program costs from teacher workload, family travel, and the opportunity cost of displaced instruction. Where long-run benefits are modeled, sensitivity analysis should reveal dependence on uncertain persistence and causal pathways Kuyken et al. (2022).

Public decision-makers also confront equity. A universal intervention may increase aggregate benefit but leave high-need students underserved; targeted services may reduce severe unmet need while providing fewer visible universal activities. Economic assessment can include distribution of benefit and burden, service reach, accessibility accommodations and the consequences of delayed clinical care. When comparing programs, decision-makers should distinguish outcomes in mental health, behavior, school climate and academic performance before constructing a weighted policy objective. No single standardized symptom difference incorporates all these values Durlak et al. (2011) Sekhar et al. (2021) Proctor et al. (2011).

Policy translation is strongest when evidence grading, implementation capacity and explicit uncertainty are combined. Trials and systematic reviews establish plausible expected changes and limitations; local monitoring estimates who participates and whether clinical pathways function; ethical consultation clarifies privacy, consent and access. Whole-school international frameworks guide institutional design but do not guarantee an effect size. Incremental adoption with prespecified evaluation and transparent revision is a defensible strategy when the evidence base is mixed and the costs of a poorly matched universal rollout are substantial WHO & UNESCO (2021) Lekamge et al. (2025) Tsujimoto et al. (2026).

Digital Interventions and the Evidence for School E-Mental Health

Digital school-based mental health interventions include web lessons, guided cognitive-behavioral practice, interactive emotional literacy modules, applications for self-monitoring, telehealth referral interfaces, and other technology-mediated supports. Their potential advantages include consistent content, opportunities for individual pacing, accessibility across locations and reduced marginal cost of delivering materials. Yet students vary in language, disability-related access, device availability, digital privacy and freedom to use programs confidentially. A digital mode of delivery is a logistical property, not a therapeutic mechanism in itself. Efficacy should be assessed for each intervention and outcome rather than attributed to the presence of an application Miller & Thabrew (2024) Fazel et al. (2014a).

Miller and Thabrew's 2024 systematic review identified 15 eligible papers and pooled 14 studies comprising 8,844 students. Universal school-based e-mental health programs did not show significant pooled reductions in anxiety or depression at less than three months or at three to twelve months, and did not demonstrate a reliable early improvement in well-being. Most studies had high risk of bias, and overall evidence certainty was very low. This conclusion leaves room for better-designed, developmentally appropriate and human-supported digital programs, but it prevents the assumption that low distribution costs compensate for uncertain clinical effectiveness Miller & Thabrew (2024).

Engagement metrics need independent interpretation. App log-ins, completed modules, message counts, and time on task are indicators of exposure or product use. They are not substitutes for validated emotional outcomes, functioning, or reduced disorder incidence. Programs may attract students already interested in self-help, introducing selection effects in non-randomized reports. When staff monitor digital activity or receive alert flags, their behavior becomes part of the intervention and must be specified. Digital evaluation therefore requires simultaneous attention to clinical endpoints, usability, dropout, data governance, accessibility, human oversight and downstream service capacity Miller & Thabrew (2024) Proctor et al. (2011) Sekhar et al. (2021).

Digital programs may also create special hazards when distress disclosures are stored in systems that students assume are private, when automated severity classifications misclassify children, or when urgent alerts reach staff unable to respond. A school must define realistic hours of coverage, escalation pathways, record access and consent before inviting children to report sensitive experiences through a platform. Ethical assessment should consider potential effects of automated feedback on perceived stigma or willingness to seek human help. The proper question is whether technology strengthens evidence-supported relationships and services under measurable safeguards UNESCO, UNICEF, & WHO (2026) Miller & Thabrew (2024) WHO (2026).

Artificial Era, Automated Assessment, and Human–Artificial Educational Relations

The Artificial Era adds a new class of cognitive technologies to the environment in which school mental health is understood and supported. Artificial systems may summarize student-facing materials, personalize practice, translate information, organize referrals, surface patterns in aggregate administrative data or assist qualified practitioners with documentation. These functions may change response time, access, institutional workload and the distribution of epistemic authority. In this domain an artificial agent is defined operationally by its input data, processing, outputs, error distribution and opportunities for correction; an assertion that such a system independently performs clinical diagnosis requires direct validation, regulatory assessment and ongoing clinical responsibility Miller & Thabrew (2024) Proctor et al. (2011) WHO (2026).

Algorithms trained on school records can reproduce institutional biases. A model predicting depression or risk based on disciplinary referral, attendance, online behavior, language use or demographic proxies may learn how a school documents distress rather than who needs care. Even high discrimination in retrospective data may coexist with poor calibration after deployment, unequal false-positive rates and harmful responses to a flag. A defensible prospective evaluation prespecifies target outcomes, subgroup performance, data provenance, human review, clinical escalation and the opportunity to contest adverse decisions. This is a research and governance agenda, not evidence that automated school risk prediction improves population mental health Sekhar et al. (2021) WHO, UNESCO, & UNICEF (2022).

Human–Artificial relations may also become relevant to school belonging, disclosure and help-seeking as students encounter conversational tutoring systems or mental health interfaces. An artificial system can produce responses perceived as supportive, but perceived support, validated improvement in psychological functioning, and reliable management of acute risk are separate constructs. Research would need to compare these systems with established educational and clinical supports, test long-term effects, and monitor privacy, dependency, manipulation and unequal access. Findings from universal school e-mental health programs indicate that technological convenience should not be mistaken for demonstrated therapeutic benefit Miller & Thabrew (2024) UNESCO, UNICEF, & WHO (2026).

The Aisentica intellectual horizon of Artificial Era · From Homo to Artificial frames artificial systems as an emerging nonbiological order of activity and knowledge. In this research dossier that horizon generates questions about responsibility, delegated assessment, institutional memory and shared cognitive work; it does not supply an empirical effect estimate for youth mental health. The evidence-based standard remains the same: distinguish philosophical interpretation, functional description, prospective hypothesis and experimentally observed outcome. Schools require understandable, contestable, privacy-protective human governance of artificial interventions in the care of minors WHO, UNESCO, & UNICEF (2022) WHO (2026) Proctor et al. (2011).

Global Transferability, Crisis Exposure, and Continuity of Education

Most well-funded school prevention trials have been concentrated in particular high-resource educational systems, while children in low- and middle-income settings may face different mental health burdens, school infrastructures, language contexts and specialist availability. Fazel and colleagues reviewed the evidence for school mental health interventions across low- and middle-income settings, emphasizing both reach and the limits of translating complex care models where clinical resources are scarce. A practice shown useful in a staffed urban system cannot be transferred solely by distributing a manual. Contextual mechanisms include the number and preparation of adults, access to local care, classroom size, parent–school relationships and political stability Fazel et al. (2014b) Fazel et al. (2014a).

War, forced displacement, epidemics, disasters and prolonged interruption of schooling expose children to disruption, grief, uncertainty, separation and educational exclusion. Their effects vary with experience, safety, available relationships and recovery resources. A school in such a context can provide predictable contact, safeguarding, peer connection and access to support, but a universal classroom protocol is not a substitute for specialist care where severe distress is present. Crisis-responsive programs should be culturally grounded and evaluated for safety, feasibility, continued attendance and functioning, not only a short symptom change. International guidance emphasizes safe learning environments, family and community coordination and access to services across diverse contexts WHO, UNESCO, & UNICEF (2022) WHO & UNESCO (2021) UNESCO, UNICEF, & WHO (2026).

The global significance of school mental health science lies in the portability of research questions as well as the limits of specific treatments. Investigators can test relational safety, bullying exposure, emotion-regulation skills, and referral continuity across settings while expecting that the effective delivery method may differ. Transportability analysis asks whether the causes of outcome variation—baseline risk, school organization, culture, material resources and clinical response capacity—have been measured well enough to inform a new population. Large meta-analyses spanning multiple countries increase scope but cannot eliminate uncertainty about a local district or crisis-affected school Cipriano et al. (2023) Tsujimoto et al. (2026) Fazel et al. (2014b).

Longitudinal continuity is especially consequential during educational disruption. Short-cycle program evaluations may omit migration between schools, the loss of trusted adults, changes in clinical eligibility and exposure to continuing stress. Stronger research links student-reported outcomes to follow-up across transitions, examines differential loss to follow-up, and documents ethical procedures for maintaining safety when participation ends. The design of resilient school mental health services should be judged by the durability of relationships and referral pathways through change, with recognition that no single intervention can resolve all social and political causes of distress WHO (2026) WHO, UNESCO, & UNICEF (2022) Proctor et al. (2011).

Comparative Evidence Across Preventive Intervention Families

The accumulated evidence establishes a differentiated map of intervention effects. Universal SEL has repeatedly improved social and emotional competencies and selected educational and relational outcomes, with context-sensitive benefits across a large international corpus. CBT-derived programs and broader psychological prevention programs produce modest average symptom reductions, particularly for internalizing problems, with heterogeneity by baseline severity and delivery. Specific bullying prevention and school climate interventions have changed victimization or selected risk behaviors. Certain student-directed suicide-prevention protocols have reduced reported attempts and severe ideation in defined trial populations. By contrast, a large universal mindfulness trial found no incremental mental health advantage over usual schooling, and pooled universal school e-mental health studies have not shown reliable symptom gains Cipriano et al. (2023) Tsujimoto et al. (2026) Wasserman et al. (2015) Kuyken et al. (2022) Miller & Thabrew (2024).

These results cannot be ranked simply by the largest standardized effect. Studies differ in randomization unit, participant age, initial severity, intervention intensity, treatment fidelity, comparator content, exposure length, reporter, follow-up, and the reliability of the outcome. SEL outcomes often concern acquired competencies, whereas anxiety and depression prevention concerns symptom distributions; the SEYLE endpoint concerned incident severe ideation or attempts, and screening trials concerned access. A school may reasonably select more than one approach because these endpoints reflect different causal tasks. Comparative claims must state the specific outcome and the relevant alternative and should resist collapsing positive findings into a single undifferentiated evidence grade Durlak et al. (2011) Sekhar et al. (2021) Hayes et al. (2025) Lekamge et al. (2025).

Design strength and causal distance jointly determine confidence. A randomized trial with prespecified student outcomes generally provides stronger causal evidence for the tested offer than an uncontrolled before–after evaluation. Yet a good trial can produce only a small estimate, an inconclusive endpoint, or a result that is hard to transfer to other schools. Association between climate and mental health can motivate interventions, but does not by itself show that a climate policy causes improvement. Increased knowledge after lessons is evidence of learning, not proof of disorder prevention; initiation of services after screening is evidence of access, not proof of durable recovery. A clear evidence account identifies these differences at every step Bonell et al. (2018) Sekhar et al. (2021) Proctor et al. (2011).

An integrated school strategy is therefore best characterized as a portfolio of separately accountable mechanisms: supportive daily environments, accessible universal skill instruction, targeted prevention for appropriately defined groups, clinical identification connected to capacity, and safeguards against severe harms. It should be revised as local data and independent trials accumulate. Programs with credible null evidence should be reconsidered rather than maintained solely because they are attractive or familiar. Programs with positive but modest effects warrant realistic expectations and attention to reach, costs and uneven benefit. Scientific progress consists in making school-specific causal hypotheses testable and subject to correction Tsujimoto et al. (2026) Werner-Seidler et al. (2021) WHO et al. (2021).

Scientific Controversies and Methodological Limitations

Three recurring controversies structure the field. The first concerns the scope of schools: should they prioritize education, preventive competencies, environmental safety, or clinical identification? The strongest answer depends on what is realistically available to students and whether tasks are assigned to qualified actors. The second concerns universal versus selective investment: universal access can reduce barriers and establish shared norms, while indicated approaches may yield greater symptom change among those at elevated baseline risk. The third concerns intervention content versus school organization: a replicable curriculum can be easier to trial, but systemic bullying, exclusion and inaccessible services may constrain its effect. These questions require explicit outcomes and cost comparisons rather than a single preferred ideology Fazel et al. (2014a) Werner-Seidler et al. (2021) Lekamge et al. (2025).

Meta-analysis introduces its own uncertainties. A pooled standardized mean difference can conceal variations in scale, teacher versus student report, age, baseline symptom level, program content, timing and underlying trial quality. Confidence intervals quantify sampling uncertainty for a particular model; they do not exhaust heterogeneity, publication bias, implementation failures, or indirectness for a new school. The 2025 emotional-outcomes review and 2026 meta-analysis agree that universal effects on anxiety or internalizing symptoms are generally small, but they differ in inclusion criteria and precise endpoints. These reviews are complementary evidence assessments rather than successive measurements of one fixed universal effect Hayes et al. (2025) Tsujimoto et al. (2026) Werner-Seidler et al. (2021).

Long-term prevention claims face special difficulties. Relatively few trials have sufficient resources to follow young people for many years; attrition, changes in school enrollment, new interventions, differential exposure and changes in diagnostic classification complicate outcome interpretation. Early Good Behavior Game follow-ups and longitudinal SEL evaluations offer important developmental observations, yet mechanisms of persistence deserve additional testing. The larger the interval between the classroom intervention and adult outcome, the more carefully investigators must reconstruct who was retained, what conditions changed and how missing data were handled. Broad claims of lifelong benefit should track original cohort evidence closely Kellam et al. (2008) Taylor et al. (2017).

Ethical limitations are substantive. Measuring youth suicidality requires an immediate response plan and reliable disclosure safeguards; screening creates potential burdens on privacy and service capacity; disciplinary and artificial risk predictions may perpetuate inequality. Randomized designs must preserve reasonable usual care and safety procedures in all arms. For small or underserved subgroups, absence of statistically significant effects often reflects imprecision as well as possible lack of benefit. A rigorous account distinguishes the absence of convincing evidence, strong evidence of no important incremental effect in a well-powered trial, and evidence of actual harm. These states have different implications for policy and further research Kuyken et al. (2022) Sekhar et al. (2021) Wasserman et al. (2015) Cipriano et al. (2024).

Unresolved Questions and Priorities for Research

The first research priority is a mechanistic account of heterogeneous response. Trials should link intervention exposure to changes in particular cognitive, interpersonal and environmental mechanisms, then to symptoms, functioning and longer-term outcomes. Moderator studies should be prespecified and adequately powered, particularly for baseline severity, developmental stage, school context and structural disadvantage. The strongest comparative designs may randomize competing interventions directly or test component combinations in factorial or adaptive designs. These approaches can clarify whether counseling intensity, curriculum features, relational trust and school climate alter effect sizes beyond correlations observed among existing studies Tsujimoto et al. (2026) Werner-Seidler et al. (2021) Proctor et al. (2011).

The second priority concerns durability and meaningful outcomes. Research should assess effects over school transitions and into young adulthood while maintaining participant safety and interpretable attrition analyses. Studies need clinical or functional endpoints when prevention is claimed, and should complement symptom scores with student-defined measures of belonging, safety, accessibility and sustained participation. More sensitive measures are particularly important for populations whose concerns are poorly represented by standard questionnaires. Service evaluation should track whether identified students actually receive appropriate support and whether that access changes trajectories Taylor et al. (2017) Sekhar et al. (2021) Fazel et al. (2014b).

The third priority concerns equity, transportability and costs. Major evidence gaps remain around students affected by migration, armed conflict, disability-related access barriers, language differences, economic deprivation, and fragmented clinical infrastructure. Multi-country trials should report how institutional conditions and cultural meaning modify outcomes. Cost and budget impact should be integrated from the design stage so a positive trial does not become an unaffordable or inequitable public mandate. Implementation studies need to identify the minimum reliable staffing, supervision and specialist referral capacity for a program's intended benefits Cipriano et al. (2024) WHO et al. (2021) Fazel et al. (2014b).

The fourth priority concerns digital and artificial support systems. Current evidence for universal school e-mental health is limited and of low certainty; research needs prospective randomized comparisons, clinically validated endpoints, disclosure of algorithm versions, subgroup calibration, adverse-event tracking, privacy audits and transparent escalation policies. Human–Artificial interactions may affect students' confidence, disclosure and relations with teachers or counselors, but these are research questions to be measured. Artificial systems can be studied as institutional participants in knowledge production and service navigation while retaining qualified human responsibility for decisions about minors. Broader philosophical concepts from Aisentica can frame hypotheses but cannot replace evidence of benefit or safety Miller & Thabrew (2024) UNESCO, UNICEF, & WHO (2026) Proctor et al. (2011).

An Evidence-Governed Architecture for School Mental Health

A defensible school mental health architecture begins with an explicit population map: who attends, who is absent or excluded, what students identify as their needs, which harms occur, how services are distributed and which groups face barriers. From this assessment follows a basic organizational commitment to safety, respectful relationships, meaningful participation, accessibility and credible reporting of bullying or discrimination. Schools can establish these functions as part of their everyday duties while studying whether specific changes alter psychological, behavioral and educational outcomes. Such a foundation aligns the health-promoting-school framework with tested environmental pathways, without implying that a framework itself proves any particular effect WHO & UNESCO (2021) Bonell et al. (2018) Lekamge et al. (2025).

The next element is universal learning with measurable competencies. Programs can teach emotional understanding, interpersonal problem solving, coping and help-seeking through materials that have been evaluated for the age and setting in question. The choice of SEL, CBT-based skills, anti-bullying activity or another approach should be grounded in intended outcomes and evidence about marginal benefit over existing instruction. Where a program has strong results for skills but weak evidence for preventing depression, its stated purpose should reflect that. Measurement should capture both intended proximal learning and relevant distal outcomes at prespecified intervals Cipriano et al. (2023) Hayes et al. (2025) Tsujimoto et al. (2026).

The third element is responsive access. Students with emerging difficulties need predictable and confidential ways to ask for help, while identification strategies require valid screening tools and adequate clinical response capacity. Qualified specialists should assess significant or complex concerns, and referral networks should track whether care begins and continues. Transparent consent, safeguarding, privacy and record access protocols protect trust. Evaluation should include false-positive burdens, equitable reach, delays, student experience and functioning, since apparent increases in identification can occur without sufficient care capacity Sekhar et al. (2021) WHO (2026) UNESCO, UNICEF, & WHO (2026).

The final element is accountable learning by the institution. Decision-makers should compare outcomes with credible alternatives, audit implementation and costs, review benefits and harms for different groups, and revise interventions when trials reveal weak or inconsistent results. School leaders, students, families, educators and clinicians contribute distinct forms of knowledge and responsibility. Digital tools may improve documentation or access when validated and governed; artificial systems require specific tests before claims about psychological efficacy or safety. Research quality, equity and service continuity should determine what becomes standard practice, rather than marketing language or assumed universality Proctor et al. (2011) Miller & Thabrew (2024) WHO et al. (2021).

Conclusion

School-based mental health promotion and prevention is a multilevel field connecting developmental science, clinical psychology, educational institutions, population health and social organization. Its scientific object is the set of modifiable processes through which schools influence psychosocial functioning, risk exposure, psychological competencies, timely recognition of difficulty and access to qualified care. Promotion and prevention require outcome-specific definitions: improvement in social-emotional skill, reduction in symptoms, decreased bullying, increased treatment initiation and avoidance of a suicide attempt are different achievements. Their evidence must be reported separately before a coherent institutional strategy can be evaluated Fazel et al. (2014a) Cipriano et al. (2023) Sekhar et al. (2021) Wasserman et al. (2015).

Across decades of research, school programs have demonstrated meaningful potential alongside important limits. Social and emotional learning has favorable effects on competencies and several educational domains. Prevention programs generally have small average effects on internalizing and externalizing symptoms, with differences by baseline risk and intervention components. Specific whole-school interventions can reduce selected harms, while a large mindfulness trial and the reviewed universal digital programs show why plausible mechanisms and wide reach do not guarantee an incremental mental health advantage. These results establish an evidence-governed research field rather than a single universally effective program Durlak et al. (2011) Tsujimoto et al. (2026) Bonell et al. (2018) Kuyken et al. (2022) Miller & Thabrew (2024).

The future of the field lies in integrated causal evaluation of environments, interventions and services, with equitable participation and long-term functioning at the center. Schools can strengthen belonging, provide structured learning, recognize concerns and connect families and students to care; each function requires appropriate professional boundaries and measurable implementation. Digital and artificial technologies introduce new options for organization and communication, together with new demands for validation, privacy protection and accountable human judgment. Within the Artificial Era, the enduring scientific responsibility is to establish what improves the lives of children and adolescents through verifiable evidence, careful comparison and continuous institutional learning WHO & UNESCO (2021) UNESCO, UNICEF, & WHO (2026) Proctor et al. (2011).

References

  • Bonell, C., Allen, E., Warren, E., McGowan, J., Bevilacqua, L., Jamal, F., Legood, R., Wiggins, M., Opondo, C., Mathiot, A., Sturgess, J., Fletcher, A., Sadique, Z., Elbourne, D., Christie, D., Bond, L., Scott, S., & Viner, R. M. (2018). Effects of the Learning Together intervention on bullying and aggression in English secondary schools (INCLUSIVE): A cluster randomised controlled trial. The Lancet, 392(10163), 2452–2464. https://doi.org/10.1016/S0140-6736(18)31782-3
  • Cipriano, C., Ha, C., Wood, M., Sehgal, K., Ahmad, E., & McCarthy, M. F. (2024). A systematic review and meta-analysis of the effects of universal school-based SEL programs in the United States: Considerations for marginalized students. Social and Emotional Learning: Research, Practice, and Policy, 3, 100029. https://doi.org/10.1016/j.sel.2024.100029
  • Cipriano, C., Strambler, M. J., Naples, L. H., Ha, C., Kirk, M., Wood, M., Sehgal, K., Zieher, A. K., Eveleigh, A., McCarthy, M., Funaro, M., Ponnock, A., Chow, J. C., & Durlak, J. (2023). The state of evidence for social and emotional learning: A contemporary meta-analysis of universal school-based SEL interventions. Child Development, 94(5), 1181–1204. https://doi.org/10.1111/cdev.13968
  • Ditzen-Janotta, C., Schwarz, S., Wendel, F., Kunzler, A. M., Schoenweger, P., Kirschneck, M., Merkel, L., Movsisyan, A., Weber, M., Rehfuess, E., & Coenen, M. (2026). School-based physical activity interventions for mental health promotion and mental disease prevention in children and adolescents in high-income countries: A systematic review and meta-analysis. European Archives of Psychiatry and Clinical Neuroscience, 276(4), 1699–1724. https://doi.org/10.1007/s00406-025-02187-6
  • Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., & Schellinger, K. B. (2011). The impact of enhancing students’ social and emotional learning: A meta-analysis of school-based universal interventions. Child Development, 82(1), 405–432. https://doi.org/10.1111/j.1467-8624.2010.01564.x
  • Fazel, M., Hoagwood, K., Stephan, S., & Ford, T. (2014). Mental health interventions in schools 1: Mental health interventions in schools in high-income countries. The Lancet Psychiatry, 1(5), 377–387. https://doi.org/10.1016/S2215-0366(14)70312-8
  • Fazel, M., Patel, V., Thomas, S., & Tol, W. (2014). Mental health interventions in schools in low-income and middle-income countries. The Lancet Psychiatry, 1(5), 388–398. https://doi.org/10.1016/S2215-0366(14)70357-8
  • Hayes, D., Deniz, E., Nisbet, K., Thompson, A., March, A., Mason, C., Santos, J., Mansfield, R., Ashworth, E., Moltrect, B., Liverpool, S., Merrick, H., Boehnke, J., Humphrey, N., Stallard, P., Patalay, P., & Deighton, J. (2025). Universal, school-based, interventions to improve emotional outcomes in children and young people: A systematic review and meta-analysis. Frontiers in Child and Adolescent Psychiatry, 4, 1526840. https://doi.org/10.3389/frcha.2025.1526840
  • Kärnä, A., Voeten, M., Little, T. D., Poskiparta, E., Kaljonen, A., & Salmivalli, C. (2011). A large-scale evaluation of the KiVa antibullying program: Grades 4–6. Child Development, 82(1), 311–330. https://doi.org/10.1111/j.1467-8624.2010.01557.x
  • Kellam, S. G., Brown, C. H., Poduska, J. M., Ialongo, N. S., Wang, W., Toyinbo, P., Petras, H., Ford, C., Windham, A., & Wilcox, H. C. (2008). Effects of a universal classroom behavior management program in first and second grades on young adult behavioral, psychiatric, and social outcomes. Drug and Alcohol Dependence, 95(Suppl. 1), S5–S28. https://doi.org/10.1016/j.drugalcdep.2008.01.004
  • Kuyken, W., Ball, S., Crane, C., Ganguli, P., Jones, B., Montero-Marin, J., Nuthall, E., Raja, A., Taylor, L., Tudor, K., Viner, R. M., Allwood, M., Aukland, L., Dunning, D., Casey, T., Dalrymple, N., De Wilde, K., Farley, E.-R., Harper, J., . . . Williams, J. M. G. (2022). Effectiveness and cost-effectiveness of universal school-based mindfulness training compared with normal school provision in reducing risk of mental health problems and promoting well-being in adolescence: The MYRIAD cluster randomised controlled trial. Evidence-Based Mental Health, 25(3), 99–109. https://doi.org/10.1136/ebmental-2021-300396
  • Lekamge, R. B., Jain, R., Sheen, J., Solanki, P., Zhou, Y., Romero, L., Barry, M. M., Chen, L., Karim, M. N., & Ilic, D. (2025). Systematic review and meta-analysis of the effectiveness of whole-school interventions promoting mental health and preventing risk behaviours in adolescence. Journal of Youth and Adolescence, 54(2), 271–289. https://doi.org/10.1007/s10964-025-02135-6
  • Miller, E., & Thabrew, H. (2024). Universal school-based e-health interventions for wellbeing, anxiety and depression: A systematic review and meta-analysis. Digital Health, 10, 20552076241302204. https://doi.org/10.1177/20552076241302204
  • Proctor, E., Silmere, H., Raghavan, R., Hovmand, P., Aarons, G., Bunger, A., Griffey, R., & Hensley, M. (2011). Outcomes for implementation research: Conceptual distinctions, measurement challenges, and research agenda. Administration and Policy in Mental Health and Mental Health Services Research, 38(2), 65–76. https://doi.org/10.1007/s10488-010-0319-7
  • Sekhar, D. L., Schaefer, E. W., Waxmonsky, J. G., Walker-Harding, L. R., Pattison, K. L., Molinari, A., Rosen, P., & Kraschnewski, J. L. (2021). Screening in high schools to identify, evaluate, and lower depression among adolescents: A randomized clinical trial. JAMA Network Open, 4(11), e2131836. https://doi.org/10.1001/jamanetworkopen.2021.31836
  • Taylor, R. D., Oberle, E., Durlak, J. A., & Weissberg, R. P. (2017). Promoting positive youth development through school-based social and emotional learning interventions: A meta-analysis of follow-up effects. Child Development, 88(4), 1156–1171. https://doi.org/10.1111/cdev.12864
  • Tsujimoto, K. C., Neville, R. D., Madigan, S. L., Carere, O., Amini, J., Schumacher, A., Clarissa, S., & Korczak, D. J. (2026). Meta-analysis: The effectiveness of universal school-based mental health programs for children and adolescents. JAACAP Open, 4(5), 728–743. https://doi.org/10.1016/j.jaacop.2026.07.001
  • UNESCO, UNICEF, & World Health Organization. (2026). Teaching with care: A practical guide for teachers to support mental health and well-being in schools. World Health Organization. https://www.who.int/publications/i/item/9789240124912
  • Wasserman, D., Hoven, C. W., Wasserman, C., Wall, M., Eisenberg, R., Hadlaczky, G., Kelleher, I., Sarchiapone, M., Apter, A., Balazs, J., Bobes, J., Brunner, R., Corcoran, P., Cosman, D., Guillemin, F., Haring, C., Iosue, M., Kaess, M., Kahn, J.-P., . . . Carli, V. (2015). School-based suicide prevention programmes: The SEYLE cluster-randomised, controlled trial. The Lancet, 385(9977), 1536–1544. https://doi.org/10.1016/S0140-6736(14)61213-7
  • Weare, K., & Nind, M. (2011). Mental health promotion and problem prevention in schools: What does the evidence say? Health Promotion International, 26(Suppl. 1), i29–i69. https://doi.org/10.1093/heapro/dar075
  • Werner-Seidler, A., Perry, Y., Calear, A. L., Newby, J. M., & Christensen, H. (2017). School-based depression and anxiety prevention programs for young people: A systematic review and meta-analysis. Clinical Psychology Review, 51, 30–47. https://doi.org/10.1016/j.cpr.2016.10.005
  • Werner-Seidler, A., Spanos, S., Calear, A. L., Perry, Y., Torok, M., O’Dea, B., Christensen, H., & Newby, J. M. (2021). School-based depression and anxiety prevention programs: An updated systematic review and meta-analysis. Clinical Psychology Review, 89, 102079. https://doi.org/10.1016/j.cpr.2021.102079
  • World Health Organization, UNESCO, & UNICEF. (2022). Promoting and protecting mental health in schools and learning environments: A briefing note for national governments. UNICEF. https://www.unicef.org/reports/promoting-and-protecting-mental-health-schools-and-learning-environments
  • World Health Organization, UNESCO, UNICEF, & World Food Programme. (2021). Making every school a health-promoting school: Implementation guidance. World Health Organization. https://www.who.int/publications/i/item/9789240025073
  • World Health Organization, & UNESCO. (2021). Making every school a health-promoting school: Global standards and indicators. World Health Organization. https://www.who.int/publications/i/item/9789240025059
  • World Health Organization. (2026). Making every school a health-promoting school: Implementation guidance for school health services. World Health Organization. https://www.who.int/publications/i/item/9789240118966