BACKGROUND: Self-harm is common in youth and an important risk factor for suicide. Certain self-harm methods might indicate a higher risk of suicide. The main aim of this study was to determine whether some methods of self-harm in adolescents (10-17 years) and young adults (18-24 years) are associated with a particularly high risk of suicide. A secondary aim was to ascertain how different self-harm methods might affect the probability of psychiatric follow-up.

METHOD: Five Swedish registers were linked in a national population-based cohort study. All nonfatal self-harm events recorded in specialist health care, excluding psychiatry and primary care services, among 10-24 year olds between 2000 and 2009 were included. Methods were classified as poisoning, cutting/piercing, violent method (gassing, hanging, strangulation/suffocation, drowning, jumping and firearms), other and multiple methods. Hazard Ratios (HR) for suicide were calculated in Cox regression models for each method with poisoning as the reference. Odds Ratios (OR) for psychiatric inpatient care were determined in logistic regression models. Analyses were adjusted for important covariates and stratified by age group and treatment setting (inpatient/outpatient).

RESULTS: Among adolescents with initial medical hospitalisation, use of a violent method was associated with a near eightfold increase in HR for suicide compared to self-poisoning in the adjusted analysis [HR 7.8; 95% confidence interval (CI) 3.2-19.0]. Among hospitalised young adult women, adjusted HRs were elevated fourfold for both cutting [4.0 (1.9-8.8)] and violent methods [3.9 (1.5-10.6)]. Method of self-harm did not affect suicide risk in young adult men. Adolescents using violent methods had an increased probability of psychiatric inpatient care following initial treatment for self-harm.

CONCLUSIONS: Violent self-harm requiring medical hospitalisation may signal particularly high risk of future suicide in adolescents (both sexes) and in young adult women. For the latter group this is the case for cutting requiring hospitalisation as well.