There is a substantial gap in this field between what the evidence supports and what is being delivered, and the gap runs in a consistent direction: the better-evidenced interventions are more expensive, and the cheaper ones are being scaled regardless.
Interventions with strong evidence
Cognitive-behavioural therapy for anxiety and depression. Among the best-evidenced psychological treatments in adolescence, with substantial trial support. Effects are meaningful and are largest for anxiety.
The constraint is access. Waiting times for adolescent mental health services are long in most systems, and the workforce required to deliver it is not being trained at anything like the rate needed.
Sleep. The causal evidence is unusually strong for this field, because sleep can be manipulated experimentally. Restriction studies produce measurable mood effects; extension improves them. Adolescent circadian phase shifts later at puberty, and school start times in most systems are misaligned with it.
Studies of delayed school start times have found improvements in sleep duration and in some mental health and attendance measures. This is one of the few structural interventions with reasonable evidence and it is implemented rarely, because it is logistically disruptive rather than because it does not work.
Physical activity. Meta-analyses find modest but consistent effects on depressive symptoms in adolescents. Effect sizes are smaller than for therapy and the intervention is much more available.
Family-based intervention for specific conditions. For adolescent eating disorders in particular, family-based treatment has substantially better evidence than individual approaches.
Interventions with moderate or mixed evidence
School-based universal programmes. Meta-analyses generally find small effects. Some programmes perform considerably better than others, and the average conceals wide variation.
Universal delivery also has an efficiency problem: most recipients do not have the condition being prevented, so the average effect is diluted by design.
Mindfulness in schools. A large trial of a school-based mindfulness programme found no benefit on the primary mental health outcomes, which was a significant result given how widely such programmes had been adopted. Smaller studies have found positive effects. The honest position is that the evidence does not support the confidence with which these programmes have been rolled out.
Digital and app-based interventions. Evidence is growing and is mixed. Effects in trials are frequently modest, and real-world engagement is dramatically lower than trial engagement — dropout rates in unsupported app use are very high.
Interventions with weak evidence or documented risk
One-off awareness talks and assemblies. No good evidence of behavioural effect. Same finding as in every other domain covered on this site.
Detailed accounts of self-harm or disordered eating delivered to general audiences. Guidance in this area consistently cautions against method description, and there is a body of research on contagion effects following detailed media coverage of suicide that informs those cautions.
Peer-led programmes without professional oversight. Some have been found effective; others raise concerns about burden placed on the peer supporters and about escalation of risk not being recognised.
Screening without a treatment pathway. Identifying adolescents with symptoms and then placing them on a two-year waiting list is not a null intervention. It creates expectation and does not meet it, and there is reasonable argument that it can be harmful.
The access problem is the main problem
It is worth stating plainly that the largest single constraint is not knowledge of what works but availability of it.
In most systems with published data, a majority of adolescents with diagnosable conditions do not receive treatment, and among those referred, waiting times are frequently measured in many months.
Under those conditions, debates about which programme is marginally better are somewhat beside the point. The binding constraint is workforce.
What a family can actually do
Prioritise sleep, seriously and structurally. Devices out of the bedroom, consistent timing, and recognition that a chronically sleep-deprived adolescent will not respond to anything else while she remains sleep-deprived.
Maintain physical activity, framed around enjoyment rather than health, since the adherence evidence favours the former.
Seek help early rather than at crisis, because the waiting times mean that a referral made at the point of crisis arrives too late by definition.
Ask what specific therapy is being offered and what its evidence base is. "Support" covers an enormous range, and the difference between structured evidence-based therapy and general supportive contact is substantial.
And be sceptical of anything being sold as transformative. The honest summary of this field is that the effective interventions produce moderate improvements, take time, and are hard to access — which is less appealing than the alternative on offer, and considerably more likely to be true.