The authors examined the ability of nonpsychiatric house staff to accurately diagnose delirium at the time of consultation. Of 221 consultations over a 5-year period, 46% were misdiagnosed by the house staff. House staff on the general medicine wards and the nonintensive care unit environment did significantly better than those on the surgical wards and intensive care units. Age, gender, and race of the patient did not overall influence incorrect diagnoses; however, when a misdiagnosis occurred, women were more often given a diagnosis of a depressive disorder, whereas men were more often given a "no diagnosis" label. Finally, the consultees improved over an academic year in accurately identifying women as delirious, whereas no such learning curve existed for men.