After a Child Shows Up in the ER for Self-Harm, Death Risks Soar for Years

When a child or teenager arrives at an emergency department after deliberately hurting themselves, clinicians face one of the most consequential decisions in pediatric medicine: what happens next? A new population-based study from Wales now offers the clearest picture yet of just how high the stakes are. Researchers tracking more than 611,000 young people aged 5 to 18 over fifteen years found that those who presented to an emergency department with self-harm faced more than a threefold increase in all-cause mortality compared with peers who visited emergency departments for other reasons. The findings, published in eClinicalMedicine, reveal a window of extreme vulnerability in the first year after presentation, when the risk of suicide was more than thirteen times higher and the risk of accidental death nearly fourteen times higher than in the comparison group.

The research team, led by Gabrielle M. Beaudry of McGill University together with colleagues including Ian Kelleher and Seena Fazel, drew on the Secure Anonymised Information Linkage Databank, a national repository that connects anonymised health, social care, and administrative records for the population of Wales. Between April 2008 and November 2023, the team identified 14,007 children and adolescents with at least one emergency department presentation for self-harm and compared them with 597,329 young people who presented for other reasons. Self-harm was defined broadly, encompassing intentional self-injury or self-poisoning regardless of suicidal intent, and was identified through a standardised national coding framework applied across Welsh emergency departments. Follow-up extended for a median of 9.2 years, allowing the researchers to trace outcomes well into early adulthood.

The scale of the dataset is what gives the study its power. Previous investigations, including cohort studies from Canada, England, and Northern Ireland, had already signalled elevated risks of suicide and accidental death after self-harm in young people, but many were confined to selected centres, focused on specific self-harm methods, or excluded younger children. By contrast, the Welsh analysis captured the full developmental range from age 5 through 18 within a single nationally representative cohort, and simultaneously examined suicide, accidental death, other causes of death, and overall mortality. Deaths were ascertained through linkage to national death registration records and classified according to the International Classification of Diseases, tenth revision, with suicide defined to include both intentional self-harm and events of undetermined intent, consistent with current United Kingdom policy practice.

The headline numbers are stark. During follow-up, 114 of the 14,007 young people in the self-harm group died, a rate of 117.6 deaths per 100,000 person-years, compared with 1,337 deaths among 597,329 controls, a rate of 26.5 per 100,000 person-years. Suicide accounted for 37.7 percent of deaths in the self-harm group and accidental causes for 43.9 percent, whereas in the control group more than half of all deaths stemmed from other causes. After adjusting for age and sex, the hazard of suicide in the first year after a self-harm presentation was 13.42 times that of controls, and the hazard of accidental death was 13.89 times higher. Although these risks attenuated after the first year, they remained dramatically elevated, at 4.85-fold for suicide and 5.46-fold for accidental death, indicating a vulnerability that persists for years rather than fading with the acute crisis.

Equally revealing is what the study did not find. Mortality from other causes, including natural causes, showed no meaningful association with self-harm presentation, with an adjusted hazard ratio of 1.09. This specificity matters scientifically: it suggests the excess deaths are not a diffuse marker of general ill health but are concentrated in external causes, particularly suicide and accidents. The researchers caution that the link between self-harm and accidental death may reflect shared underlying vulnerabilities, such as impulsivity, substance use, risk-taking behaviour, and psychosocial adversity, rather than a direct causal effect of self-harm itself. Some deaths recorded as accidents might also represent unrecognised suicides, a known challenge in mortality classification, though the magnitude and persistence of the association suggest accidental mortality is a clinically meaningful outcome in its own right.

Timing emerged as a central theme. Median time to death was shorter in the self-harm group than in controls for every cause examined: 4.9 years versus 7.3 years for suicide, 4.3 versus 6.3 years for accidental death, and 4.5 versus 6.0 years for all-cause mortality. Deaths clustered among young people whose index presentation occurred in later adolescence, between ages 15 and 18, and no suicide deaths were observed following a first self-harm presentation among children aged 5 to 9, though event counts in that youngest group were small. The first-year spike in risk carries an unmistakable clinical message: the twelve months after an emergency department presentation represent a critical window for intervention, when structured follow-up, safety planning, and rapid linkage to mental health services could plausibly alter trajectories.

Within the self-harm cohort itself, the researchers identified demographic and clinical factors that shaped risk. Each additional year of age at presentation raised the risk of suicide by 27 percent and all-cause mortality by 22 percent. Female sex was associated with substantially lower risks of suicide, accidental death, and overall mortality, with incidence rate ratios around 0.31 to 0.39, echoing the well-documented pattern in which females are more likely to self-harm but males face higher subsequent mortality. Most strikingly, recurrent self-harm, defined as at least one additional emergency department presentation after the index event, was associated with a more than threefold increase in suicide risk and a 77 percent increase in all-cause mortality. Sensitivity analyses treating recurrence as a time-varying exposure confirmed that risk climbed further after repeat presentations, reinforcing the idea that each return visit is not merely a repeat consultation but a signal of escalating danger.

Perhaps the most sobering finding, however, is one of absence: most suicide deaths in the cohort occurred among young people who had never presented to an emergency department for self-harm. Emergency department-based interventions, however well designed, can only reach the subset of distressed young people who come into contact with acute care. The authors therefore frame their results as an argument for multilevel prevention, combining targeted clinical follow-up for the highest-risk group with broader population and community strategies. Notably, factors often assumed to predict poor outcomes, including area-level deprivation, urban residence, recent contact with Child and Adolescent Mental Health Services, and Looked-After Child status, showed no significant independent associations with mortality among those already presenting with self-harm, suggesting limited prognostic value once a young person has reached emergency care.

The study is not without limitations, and the authors are candid about them. Emergency department records in Wales lack the diagnostic granularity of full ICD-10 coding, so some self-harm presentations, particularly those without overt injury or poisoning, may have been missed, and stratification by self-harm method was not possible. Low event counts in younger children and for specific causes of death limited finer analyses, including sex-stratified models. The study also captured only episodes that reached emergency departments, excluding self-harm managed in primary care or community settings, and residual confounding by unmeasured psychosocial factors such as family adversity or exposure to violence cannot be excluded. Generalisability beyond the Welsh healthcare system, with its publicly funded and free-at-point-of-access emergency care, remains an open question, and the authors call for comparable population-based studies in low- and middle-income countries where surveillance infrastructure is often lacking.

Even with these caveats, the implications for clinical practice are difficult to ignore. Evidence-based therapeutic interventions after adolescent self-harm are known to reduce repeat self-harm and suicide attempts, though a reduction in actual mortality has not yet been demonstrated in trials, likely because death remains a rare outcome in study populations. Current guidance for pediatric mental health emergencies emphasises coordinated emergency care, structured discharge planning, and timely linkage to outpatient services, and this study supplies the mortality data that make such pathways urgent rather than aspirational. The researchers also point toward the development and prospective validation of clinical risk stratification tools that could help emergency clinicians decide who needs the most intensive follow-up. What the Welsh data make clear is that a self-harm presentation in childhood or adolescence is not a single crisis to be managed and forgotten, but a durable marker of elevated mortality risk, one that demands sustained, developmentally informed attention long after the young person leaves the department.

Subject of Research: Long-term mortality risk following emergency department presentation for self-harm among children and adolescents

Article Title: Risk of mortality after emergency department presentation for self-harm among children and adolescents in Wales: a population-based cohort study

Article References: Beaudry, G. M., O'Hare, K., Yu, R., Fazel, S., & Kelleher, I. (2026). Risk of mortality after emergency department presentation for self-harm among children and adolescents in Wales: a population-based cohort study. eClinicalMedicine, 100, Article 104230. https://doi.org/10.1016/j.eclinm.2026.104230

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104230

Keywords: self-harm, children, adolescents, emergency department, suicide, accidental death, mortality, cohort study, Wales, mental health, pediatrics, prevention