This article reports on the efficacy and safety of the selective serotonin reuptake inhibitor, fluoxetine, in 213 patients (ages 11-23 years) treated by psychiatrists/neurologists (PN) or general practitioners/internists (GPI). Data were derived from naturalistic drug utilization observation (DUO) studies with fluoxetine (n = 18,759 patients). Data collection--at the start and the end of the observation period (< or =6 weeks)--included patient characteristics, diagnoses, medication, co-medication, efficacy, and adverse events (AEs). Nonparametric statistics and descriptive p values (two-tailed) were used. Analyses revealed various differences between PN (n = 56) and GPI (n = 157) samples as to patient and treatment characteristics (p < 0.001-0.08). Based on both Clinical Global Impression (CGI; all p < 0.001) and self-assessment (total n = 47; Zung SDS, all p < or = 0.003), both PN and GPI patients showed improvements in their symptomatology over time, including suicidality (all p < 0.001; there were no group differences). Overall AE rates were higher in PN patients (p < 0.01; 17.9% vs. 4.5%); the frequency and type of AEs in both subgroups were typical for fluoxetine and the total DUO samples. In fact, AE rates were lower compared to controlled trials. Findings suggest that PN patients were more severely ill at observation start and suffered a more complicated treatment course. However, clinical efficacy showed highly significant improvements in both subgroups; AE rates were low in both--although higher in PN patients. Thus, results support a positive benefit/risk ratio of fluoxetine use for this young patient population.