Original Research Prediction of Suicidality and Violence in Hospitalized Adolescents: Comparisons by Sex

Objective: To examine psychological correlates of suicidality and violent behaviour in hospitalized adolescents and the extent to which these associations may be affected by their sex. Method: A sample of 487 psychiatric inpatients (207 male, 280 female), aged 12 to 19 years, completed a battery of psychometrically sound self-report measures of psychological functioning, substance abuse, suicidality, and violent behaviour. We conducted multiple regression analyses to determine the joint and independent predictors of suicide risk and violence risk. In subsequent analyses, we examined these associations separately by sex. Results: Multiple regression analysis revealed that 9 variables (sex, age, hopelessness, self-esteem, depression, impulsivity, alcohol abuse, drug abuse, and violence risk) jointly predicted suicide risk and that an analogous model predicted violence risk. However, we found several differences with respect to which variables made significant independent contributions to these 2 predictive models. Female sex, low self-esteem, depression, drug abuse, and violence risk made independent contributions to suicide risk. Male sex, younger age, hopelessness, impulsivity, drug abuse, and suicide risk made independent contributions to violence risk. We observed a few additional differences when we considered male and female subjects separately. Conclusions: We found overlapping but distinctive patterns of prediction for suicide risk and violence risk, as well as some differences between male and female subjects. These results may reflect distinct psychological and behavioural pathways for suicidality and violence in adolescent psychiatric patients and differing risk factors for each sex. Such differences have potential implications for prevention and treatment programs. (Can J Psychiatry 2007;52:572–580) Information on funding and support and author affiliations appears at the end of the article