Natural history of suicidal behaviors in a population-based sample of young adults

Background. Suicidal behaviors in young individuals represent an important public health problem. Understanding their natural history and relationships would therefore be of clinical and research value. In this study, we examined the natural histories of several suicidal behaviors and investigated two conceptual models of suicidality (dimensional and categorical) in the context of adolescent and adult-onset suicide attempts. Method. Participants were members of a prospectively studied, representative, population-based school cohort followed since age 6 (n = 3017) through mid-adolescence (17 = 1715) to their early twenties (it = 1684). Outcome measures included suicidal ideation, attempts and completions. Results. Approximately one in 500 individuals died by suicide. About 33% had suicidal ideas and 9.3% made at least one suicide attempt. Over half (4.9%) of the self-reported attempters made their first attempt before age 18. With the exception of current suicidal ideas, non-fatal suicidal behaviors were more prevalent in females. In general, parental and cross-sectional self-reports underestimated suicidality rates. Alkaike (AIC) and Bayesian (BIC) information criteria suggested the ordinal model, and dimensional conceptualization of suicide attempts of different onset age, to be more optimal than its multinomial/categorical counterpart (ordinal: AIC 567.55, BIC 635.67; multinomial: AIC 616,59, BIC 723.83). Both models, nevertheless, identified five common factors of relevance to suicidal diathesis: gender, disruptive disorders, childhood anxiousness and abuse, and suicidal thoughts. Conclusions. Non-fatal suicidal behaviors in adolescents and young adults are more common than suggested by cross-sectional studies and parental reports. The dimensional model may be more useful in explaining the relationship of suicide attempts of different age of onset.