INCIDENCE AND PREDICTORS OF SUICIDAL BEHAVIORS IN A LONGITUDINAL SAMPLE OF YOUNG ADOLESCENTS

Objective: This analysis examines 1-year transition probabilities and baseline predictors for suicidal behaviors in young adolescents. Metho d: Adolescents from a two-stage, community-based longitudinal study we re classified into suicidal behavior categories (attempt, plan, ideati on, and none) for baseline and follow-up years. Transition probabiliti es for movement among categories were calculated, and polytomous logis tic regression analysis was used to examine predictors of suicidal beh aviors. Results: Among those with no suicidal behaviors at baseline, 1 -year incidence rates were 1.3% for attempts and 1.7% each for plans a nd ideation. Increasing family cohesion was protective for suicide att empts (odds ratio [OR] = 0.9). Female subjects were more likely than m ales to report plans (OR = 8.9) and ideation (OR = 4.1). Increasing im pulsivity (OR = 2.3), prior suicidal behavior (OR = 10.6), and undesir able life events (OR = 1.1) were significant predictors of plans. Conc lusions: While there are a number of predictors of suicidal behaviors, the false-positive rate is high. Focusing on proximal risk factors, p articularly stressors in adolescent development, may overlook the fund amental role of underlying mental disorder and familial factors-both b iological and environmental. Suicide and suicidal behaviors are the re sult of a constellation of adverse factors requiring a range of interv entions for prevention.