RESUMO
Background: Over 90% of youth suicide deaths occur in low- and middle-income countries. Despite this relatively little is known about risk factors in this context. Aims: Investigate risk factors for deliberate self-harm (non-fatal) in young people in rural Sri Lanka. Methods: A prospective cohort study of 22,401 individuals aged 12-18 years with complete data on sex, student status, household asset score, household access to pesticides and household problematic alcohol use. Deliberate self-harm was measured prospectively by reviewing hospital records. Poisson regression estimated incidence rate ratios (IRRs) for the association of risk factors with deliberate self-harm. Results: Females were at higher risk of deliberate self-harm compared to males (IRR 2.05; 95%CI 1.75 - 2.40). Lower asset scores (low compared to high: IRR 1.46, 95%CI 1.12 - 2.00) and having left education (IRR 1.61 95%CI 1.31 - 1.98) were associated with higher risks of deliberate self-harm, with evidence that the effect of not being in school was more pronounced in males (IRR 1.94; 95%CI 1.40 - 2.70) than females. There was no evidence of an association between household pesticide access and deliberate self-harm risk, but problematic household alcohol use was associated with increased risk (IRR 1.23; 95%CI 1.04 - 1.45), with evidence that this was more pronounced in females than males (IRR for females 1.42; 95%CI 1.17 - 1.72). There was no evidence of deliberate self-harm risk being higher at times of school exam stress. Conclusion: Indicators of lower socioeconomic status, not being in school, and problematic alcohol use in households, were associated with increased deliberate self-harm risk in young people.