A long-term follow-up study of adolescent psychiatric in-patients. Part IV. Predictors of a non-negative outcome

Fifteen years after first hospitalization 932 patients were followed up by record linkage to national registers of criminality, disability and causes of death. On the basis of hospital records, all patients were rediagnosed a ccording to DSM-IV and scored on factors postulated to have predictive powe r with regard to a non-negative outcome, defined as having avoided entry in to the above-mentioned registers at follow-up. Logistic regression showed t hat in males psychiatric diagnosis (odds ratio (OR) = 2.4, 95% confidence i nterval (CI)= 1.3-4.3), no psychoactive substance use (OR = 3.2, CI = 1.2-8 .0), no disciplinary problems in school (OR = 4.1, CI = 2.4-6.9) and an IQ of greater than or equal to 90 (OR = 1.9, CI = 1.03-3.4) were strong and in dependent predictors of a non-negative outcome. In females, psychiatric dia gnosis (OR = 2.2, CI = 1.3-3.6), no psychoactive substance use (OR = 2.4, C I = 1.5-3.8), DSM-IV GAF greater than or equal to 40 at admission (OR = 2.9 , CI = 1.8-4.6) and an IQ of greater than or equal to 90 (OR = 1.7, CI = 1. 01-2.8) were strong and independent predictors of a non-negative outcome. O nly 5.1% of male patients with psychoactive substance use disorder and disc iplinary problems in school had a nonnegative outcome, while as many as 76. 6% of female patients without psychoactive substance use and a DSM-TV GAF o f greater than or equal to 40 at hospitalization had a non-negative outcome at 15 years follow-up.