← All posts

Universal school-based mental health prevention is having a reckoning. Here's what the 2025 evidence actually says.

April 25, 2026 · Research · by Heardify

Something is shifting in the school-based mental health prevention literature, and practitioners should know about it. For about a decade, the consensus position in the field has been that universal, school-wide prevention programs (CBT-based, mindfulness-based, or hybrid) produce small but meaningful improvements in adolescent mental health at population scale. That consensus is now openly contested, including by researchers who have spent careers building it.

Two publications from the last 12 months capture the state of the debate. The first is a 2025 cluster-randomised controlled trial of the OurFutures Mental Health program, published in eClinicalMedicine in late 2025. The second is a formal debate series in Child and Adolescent Mental Health, also 2025, where four different research groups argue different positions on where universal school-based mental health prevention should go next.

Read together, they suggest the field is recalibrating, and that some of the things schools have been told for years about what universal prevention can and cannot do are being revisited with more candour than before. Worth walking through, if you're a program administrator making decisions this year.

The OurFutures trial

Grummitt, O'Dean, Birrell, Newton, Barrett and colleagues published Efficacy of a school-based, universal prevention programme for depression and anxiety in adolescents (OurFutures Mental Health): a two-arm cluster-randomised controlled trial in eClinicalMedicine in November 2025. The trial ran from 2023 to 2024 with Year 8 and 9 students across 10 Australian secondary schools, randomised at the school level to either the OurFutures Mental Health intervention or an active control arm of usual health education.

OurFutures Mental Health itself is a six-lesson, web-based program using a cognitive-behavioural approach, redesigned in a 2024 JMIR Pediatrics mixed-methods participatory process to be trauma-informed and gender-and-sexuality-diversity affirmative. This sits firmly in the mainstream of what digital, universal school-based CBT prevention looks like in 2026. Well-resourced, carefully iterated, and on the larger end of what schools realistically deploy.

Headline findings:

  • Anxiety symptoms: modest short-term prevention signal in the intervention arm.
  • Mental health knowledge: clearer positive signal in the intervention arm.
  • Depression symptoms: no reliable effect. The authors were explicit that further work is needed to improve depression-prevention efficacy.
  • Methodological caveats: high participant attrition, several schools withdrawing post-randomisation, and the trial being underpowered for smaller effects. The authors flagged all three.

The authors' own framing was measured. The program "may be efficacious" for knowledge and short-term anxiety prevention, not for depression. That's a careful read of their own results, and it matters because the commercial and policy framing of similar programs usually isn't that careful.

What the CAMH debate adds

The Child and Adolescent Mental Health debate series from 2025 put the broader question directly. Andrews and Foulkes argued that universal school-based prevention, as currently practiced, has accumulated enough null and mixed results that resources should be redirected toward targeted and indirect approaches with a stronger evidence base. Their position, summarised: after years of scaling up CBT-based and mindfulness-based universal programs, the population-level effects on depression and anxiety symptoms are not what was hoped for, and the opportunity cost of continued investment in universal delivery is real.

A counter-position came from Birrell, Grummitt and colleagues (the same group behind the OurFutures trial), Hetrick and colleagues on youth-led research, Coetzee and colleagues on low- and middle-income contexts, and Greenberg in a commentary warning against throwing out the broader prevention literature. These authors don't argue that current programs are working as well as earlier claims suggested. They argue that the response should be redesigning universal programs (with more personalisation, more attention to school climate, more peer-mediated elements, more student co-design) rather than abandoning them.

A few specific points worth pulling out of the debate:

  • The effect-size ceiling. Multiple commentators acknowledged that typical meta-analytic effect sizes for universal CBT-based school programs cluster around d = 0.1 to 0.3, which at population scale is small per student but potentially meaningful in aggregate. Whether that's a worthwhile return depends on cost and opportunity-cost framing.
  • The "universal versus targeted" framing may be wrong. A common thread in the pro-universal responses was that the binary choice between universal and targeted approaches is itself misleading. Proportionate universalism, which means providing a baseline for everyone while graduating support upward for those with greater need, is closer to what several authors proposed.
  • Youth involvement in program design. Both Hetrick and the OurFutures group emphasised that programs designed with students and iterated based on their feedback perform differently than programs designed for students and dropped in.
  • The depression-specific problem. Several debate contributors pointed out, as the OurFutures trial did, that adolescent depression is harder to prevent with universal programs than adolescent anxiety. Whether this is a signal issue (depression outcome measures are less sensitive in universal populations), a mechanism issue (depression has more heterogeneous causes), or a dose issue (universal programs don't deliver enough intensity to shift depression trajectories) is genuinely unresolved.

What this means for a counsellor or district lead

If you're reading the field from a practitioner seat, a few things are reasonable to take away.

Temper claims about depression prevention specifically

A universal program claiming it "prevents depression and anxiety in adolescents" is doing more work than the evidence supports. The anxiety side of that claim is stronger. The depression side is considerably weaker, in the OurFutures trial, in the broader meta-analytic literature, and in the acknowledgement of researchers who are pro-universal. If a vendor makes a prevention claim about depression specifically, ask which studies back it and what the depression-specific effect size was. A lot of programs cite composite depression-and-anxiety effects without noting that the anxiety component is doing most of the work.

Universal programs are not a substitute for a measurement-and-pathway system

The debate series repeatedly returned to the point that universal programs delivered in isolation don't move population outcomes much. What they can plausibly do is raise mental health knowledge, improve help-seeking attitudes, and reduce stigma, which then interact with a school's measurement and pathway infrastructure (see the companion piece, what happens after a positive screen) to produce downstream effects. A universal program without a system to catch students who surface as needing more is half a program.

Attrition and implementation fidelity aren't footnotes

The OurFutures trial reported high attrition and school withdrawal as main limitations. Most universal school programs face the same constraints, and most published trials under-report how much signal is being eaten by implementation issues versus program-design issues. For practical decisions, this means that the effect size you'll see in your school isn't the effect size reported in the paper. It's the paper's effect size multiplied by whatever fraction of your students actually complete the program, which is often substantially less than 100%. Budget for this.

Program selection should be boring

The debate implicitly endorses a few practical principles for choosing a universal program:

  1. Prefer programs that have been trialled with populations similar to yours, at scale, with published effect sizes.
  2. Prefer programs with a redesign and iteration history (OurFutures redesigned theirs based on a participatory process, which is the pattern).
  3. Be sceptical of one-off workshop formats. Dose matters. Programs of six or more sessions consistently outperform single-session or two-session formats across the broader universal school-based mental health review literature.
  4. Treat any program with marketing stronger than its citations as a warning sign.

Personalisation is the next design frontier

Birrell and colleagues specifically flagged personalised prevention (programs that adapt content based on a student's baseline measures, demographics, or emerging profile) as one of the directions the field is moving. This is a technology-enabled direction and one where platform products will likely differentiate. It's also not a solved problem yet, and early claims in this space should be read carefully. The evidence base for personalised, AI-adapted, or tailored-content universal programs is still thin.

Where this leaves the measurement conversation

A theme running through both the OurFutures paper and the CAMH debate is that measurement is inseparable from the prevention program. The trial used the PHQ-A and GAD-7 adapted for adolescents as primary outcomes, the appropriate choice for an under-15 Australian secondary-school population. Had the trial used adult PHQ-9 scoring without the adolescent-specific adaptation, the results would likely look different, not because the program effect is different, but because the instrument would be less sensitive to the age group.

For schools standing up measurement programs alongside prevention programs, the instrument-age question is the single most common place to get it wrong. The AACAP screening tools resource lists adolescent-validated alternatives for nearly every commonly used adult instrument. The Strengths and Difficulties Questionnaire and the RCADS remain two of the most useful starting points for practitioners.

Measurement mistakes show up in the trial literature too. Studies that use adult instruments on adolescent populations consistently report smaller or null effects, which then get cited back as evidence that prevention programs don't work. Sometimes the program doesn't work. Sometimes the instrument wasn't sensitive to the population the program was designed for. Distinguishing the two requires reading past the headline effect size.

What this blog is not saying

For the avoidance of doubt: this piece is not arguing that schools should stop running universal mental health prevention programs. The field hasn't concluded that. What the 2025 evidence is saying, carefully, is that the claims made on behalf of universal prevention over the past decade need to be recalibrated, that depression prevention in particular is harder than earlier enthusiasm suggested, and that the next generation of programs will probably look different from the current generation. Programs that will hold up are likely to be more personalised, more participatory in design, more integrated with measurement and pathway infrastructure, and more modest in their claims.

Schools currently running a universal program should keep running it if it's well-implemented, and should strengthen the measurement-and-pathway layer around it. Schools evaluating a new program should read the underlying trials, not the vendor deck.

Further reading

Heardify publishes twice-weekly on student wellbeing, measurement, and school mental health programs. Back to heardify.me or read more from the blog.