Food and Drug Administration data show that most anti-depressant studies in youth do not show drug effect. The few positive studies used rigorous diagnostic screening procedures, suggesting major depressive disorder (MDD) may not be a persistent condition in a subgroup of youth. To investigate persistence of MDD, we serially assessed a cohort of inpatients admitted to the hospital with a clinical diagnosis of MDD. Assessments included a structured diagnostic interview, the Diagnostic Interview for Children and Adolescents-Revised (DICA-R), and measures of depressive symptomatology. Of 66 subjects (40 girls; mean age, 14.4 +/- 2.2 years), 34 (51.5%) met DICA-R criteria for MDD at the initial postadmission assessment. Of these, only 8 (23.5%) met DICA-R criteria for MDD at any subsequent assessment. Similar reductions were found on other ratings of depression. In conclusion, MDD did not persist in this sample. The findings suggest a multigated assessment procedure should be employed before randomization in antidepressant clinical trials.