Little information is available about identification of mental distress by general hospital physicians (GHPs). We compared, among patients admitted in a General Hospital, depressed patients with patients identified as depressed by the GHPs. A random sample of 1,039 patients were screened with the 12-item General Health Questionnaire. Afterwards, all high scorers and a probability sample of low scorers were interviewed with a variety of measures, including the Composite International Diagnostic Interview-Primary Care version (CIDI-PHC). GHPs recorded the presence of depression on a Physician Encounter Form. Patients were more likely to have depression detected on medical than surgical wards. Of the 195 patients who had a depression, the GHPs assessed 32.5% as depressed. A number of factors associated with CIDI diagnoses were not significantly associated with being identified by the GHPs - female gender, two or more life events in the previous year, and a previous history of depression. The identification by the GHPs was associated with a higher probability of contacts with medical professionals and of antidepressant drug prescriptions during the year which followed the first interview.