Depression, family environment, and adolescent suicidal behavior

Objective: To assess the specific influence of family relationship difficulties, over and above the effect of depression, on the risk of adolescent suicidal behavior. Method: The study was based on the clinical data summaries, ''item sheets,'' of children and adolescents who attended the Maudsley Hospital during the 1970s and 1980s. Two hundred eighty-four cases of suicidal behavior, defined as suicidal ideas, attempts, or threats (mean age 13.9 years, SD 2.6), were compared with 3,054 nonsuicidal controls, using stepwise logistic regression controlling for age and sex. Results: The following variables were each independently associated with suicidal behavior: an operationally defined depressive syndrome, odds ratio (OR) = 4.4 (95% CI 3.1 to 6.3); family discord, OR = 1.5 (95% CI 1.1 to 2.0); disturbed mother-child relationship, OR = 1.5 (95% CI 1.1 to 2.0); and familial lack of warmth, OR = 1.6 (95% CI 1.1 to 2.3). Twenty-seven percent of the suicidal cases met operational criteria for depression. In a separate analysis of nondepressed cases (n = 198), female gender, OR = 2.4 (95% CI 1.7 to 3.2), and conduct symptoms, OR = 1.4 (95% CI 1.02 to 1.95), were independent risk factors for suicidal behavior. Among the depressed cases (n = 73), gender and conduct symptoms did not affect the risk of suicidal behavior. Conclusions: Although depression is the largest single risk factor for teenage suicidal behavior, family relationship difficulties make a significant independent contribution to this risk. Depression also interacts with gender, so that the excess risk of suicidal behavior in females is confined to nondepressed cases.