Which measure of adolescent psychiatric disorder—diagnosis, number of symptoms, or adaptive functioning—best predicts adverse young adult outcomes?
- 1Division of Child and Adolescent Psychiatry, University of Washington, Seattle, USA
- 2Department of Epidemiology, University of Washington
- 3Department of Biostatistics, University of Washington
- 4Division of Epidemiology, Columbia University, USA
- Correspondence to: Ann Vander Stoep, Division of Child and Adolescent Psychiatry, University of Washington, Box 359300, Seattle, WA 98195-9300, USA; annv{at}u.washington.edu
- Accepted 13 June 2001
Abstract
Objective: To test the ability of psychiatric diagnosis, symptom count, and adaptive functioning in adolescence to predict failure to complete secondary school and criminal involvement in young adulthood.
Design: Community-based cohort study.
Setting: Two counties in upstate New York, USA
Participants: 181 adolescents interviewed in 1983 and 1985–86 who were randomly selected in 1975 from a probability area sampling of representative families with 1–10 year old children
Main results: Compared with adolescents without psychiatric disorders, adolescents with depressive, anxiety, disruptive, and substance abuse disorders were 2.86–9.21 times more likely to fail to complete secondary school. Compared with adolescents without disruptive disorders, adolescents with disruptive disorders were 4.04 (1.96–8.32) times more likely to get in trouble with police during young adulthood. The positive predictive value of each measure of adolescent psychiatric disorder for school non-completion was higher in the lowest SES stratum and for young adult criminal involvement was higher for boys. Combining knowledge of symptom counts, age, gender, and social class in a logistic regression model yielded 89% sensitivity and 87% specificity for predicting future school non-completion at the p ≥ 0.13 cut off. The optimal cut off value in a model incorporating knowledge of disruptive symptoms and demographic characteristics yielded 75% sensitivity and 76% specificity for predicting future criminal involvement.
Conclusions: Screening children and adolescents for psychiatric disorders can identify those at high risk of adverse young adult outcomes. Future school and community adjustment can be predicted as easily and accurately on the basis of a simple count of psychiatric symptoms as by applying more complex diagnostic algorithms. Screening youth for psychiatric symptoms in neighbourhood, school, or primary care settings is a logical first step for early intervention to promote increased school completion and decreased criminal activity in young adulthood.
