Although cognitive variables have been shown to be useful in predicting outcomes in late-life depression, there has not yet been a comprehensive study in younger persons with depression. The clinical symptoms and cognitive performance of participants were evaluated at admission to one of two university teaching hospitals and again at 3 months after remission and discharge. A total of 52 participants with a DSM-IV diagnosis of major depressive disorder, aged between 20 and 60 years and with a Hamilton Depression Rating Scale score 17 > or = entered the study. The sample for this paper comprises the 48 subjects (mean age 37.9 years, s.d.=10.7) who received admission and follow-up assessments; an attrition rate of 7.7%. More perseverative errors on the shortened Wisconsin Card Sorting Test at admission predicted a worse clinical outcome at follow-up. Poor event-based prospective memory and more perseverative errors on the shortened Wisconsin Card Sorting Test at admission predicted worse social and occupational outcome at follow-up. These results suggest that a brief cognitive screen at hospital admission, focusing on executive function, would have a useful prognostic value in depression. Determining early predictors of individuals at risk of poorer outcomes is important for identifying those who may need altered or additional treatment approaches.