Every child born in Scotland between 1991 and 1998, followed to 2023, shows child mental health services identifying a group that keeps needing care into their thirties.

Child and adolescent mental health services are usually described as treatment for a period of life. A study in European Child & Adolescent Psychiatry suggests they also function as something else: an unusually accurate early identification of the people who will use adult psychiatric services most heavily, decades later.

The study covers every child born in Scotland between 1991 and 1998 — roughly half a million people — followed in linked records until 2023. It is the longest follow-up of young people attending mental health services conducted in the UK to date, and it was led by Professor Ian Kelleher, Chair of Child and Adolescent Psychiatry at the University of Edinburgh.

What a whole birth cohort shows

About 8 per cent of the cohort attended child and adolescent mental health services, or CAMHS, before the age of 18. Four in ten of those children went on to use specialist adult psychiatric services, and among those who attended as teenagers the figure rises to half by their early thirties — well above earlier estimates.

The concentration is the striking part. By age 32, that 8 per cent accounted for 46 per cent of adult psychiatric inpatient bed days and 41 per cent of NHS adult mental health outpatient appointments.

Those shares are not evenly spread across diagnoses. Former CAMHS patients made up 50 per cent of adults diagnosed with eating disorders, 49 per cent of those with personality disorders, 39 per cent of those with schizophrenia and 39 per cent of those with opioid use disorders.

Kelleher drew the conclusion directly, describing CAMHS as “not just a childhood treatment service” but something that “identifies a population that will go on to be high-intensity users” of adult care.

Why the whole-cohort design matters

Most research on what happens to former child mental health patients follows a clinic sample, which cannot say what share of adult service use that group represents, because the denominator is missing. A total birth cohort supplies it: everyone born in those years is in the data whether or not they ever attended a service.

That design is also what makes the concentration figures interpretable. The claim is not that former CAMHS patients have poor outcomes — a clinic study could suggest that — but that a group identifiable in childhood, amounting to one person in twelve, later accounts for nearly half of a national system’s inpatient capacity.

It is worth being clear about what this does not mean. The eight per cent were referred to CAMHS because they were already unwell, and continuity of need in a group selected for need is not a surprising finding on its own. What is new is the magnitude and the persistence: that the continuity holds across fifteen years and more, and that it is this concentrated.

No long-term view of mental health

The study’s second argument is about data infrastructure, and Kelleher made it with a comparison. Scotland tracks long-term outcomes for 27 types of cancer and has no equivalent for mental health.

The researchers call for a Mental Health Observatory that would link childhood and adult mental health records as a matter of routine, so that outcomes could be monitored continuously rather than reconstructed by individual research projects working with historical cohorts.

The practical case is that the present study describes people born between 1991 and 1998, which is the most recent cohort with enough follow-up to be studied this way. Any service change made in the past decade will not appear in data of this kind for another twenty years unless the linkage is permanent.

Dr Jane Morris, Chair of the Royal College of Psychiatrists in Scotland, framed the clinical implication as a case for acting early — stressing the importance of “identifying mental health conditions as early as possible” in order “to prevent irreversible damage to people’s healthy development”.

What the design cannot establish

This is an observational study built from administrative records, and the usual limits apply with some force.

Service use is not the same as illness. The records show who attended services, and attendance depends on referral, availability, waiting lists and family circumstances as well as on need. A child who needed CAMHS and did not reach it is counted among the 92 per cent, which means the cohort’s untreated cases sit in the comparison group and the contrast between the groups is a contrast in service contact first of all.

Nor can the study say whether CAMHS treatment changed any of these trajectories. A design that follows those who attended and those who did not, without randomisation, cannot separate the effect of treatment from the severity that prompted the referral. The finding is predictive, not evaluative, and reading it as evidence that childhood treatment failed would be a straightforward error — the counterfactual is simply absent.

The diagnostic shares need similar care. That former CAMHS patients make up half of adults diagnosed with eating disorders reflects both continuity of illness and the fact that an eating disorder frequently begins in adolescence, which is precisely when CAMHS would have been involved. The statistic describes an overlap in time as well as a prediction.

Scotland’s results also need not transfer. A single national system with one referral pathway and one set of thresholds produces figures shaped by that system, and countries with different arrangements for the transition from child to adult services would not necessarily see the same continuity.

The transition problem

The finding lands on a known weak point. The handover from child to adult mental health services at 18 is widely documented as a point where young people disengage, and a study showing that this is the group that will later consume nearly half of adult inpatient care sharpens the stakes of getting that handover right.

It also suggests a different unit of planning. If one cohort identified in childhood drives this much adult service use, then child and adult mental health provision are not two budgets but one long-run problem — which is close to the argument the authors are making when they ask for records that follow a person rather than a service.

The study was funded by the Academy of Medical Sciences, the UK Department for Science, Innovation and Technology, and Research Data Scotland. The release does not list the full author team.

---

Notes

1. “Adult mental health service outcomes for patients of child and adolescent mental health services: a total Scottish birth cohort study.” European Child & Adolescent Psychiatry, 2026. DOI: 10.1007/s00787-026-03135-y.
2. University of Edinburgh. “Almost half of child mental health patients need adult care too.” EurekAlert!, 4 August 2026. https://www.eurekalert.org/news-releases/1138574.

Photo by cottonbro studio on Pexels.com