The objectives of this prospective cohort study were to 1) determine the prevalence of depressed mood, 2) identify the characteristics associated with it, and 3) evaluate the recognition rate of depressed mood by clinicians. The study population was a cohort of 401 elderly patients, aged 75 years and older, admitted to the internal medicine service of a tertiary care academic medical center in Western Switzerland over six months. We excluded patients with severe cognitive impairment, terminal disease or those living in a nursing home. Data on demographics, medical, physical, social and mental status were collected upon admission. Presence of depressed mood was defined as a score > or = 6 on the Geriatric Depression Scale (GDS), short form (15-item). An independent reviewer performed a discharge summary abstraction to assess recognition rate. Subjects' mean age was 82.4 years, 60.9% were women. Overall, 90 patients (22.40%) had an abnormal GDS score (> or =6). Compared to those without a depressed mood, these subjects were (all p<0.05) older (83.5 vs 82.0 years), more frequently living alone (66.7 vs 55.0%), dependent in both basic activities of daily living (BADL) and instrumental ADL (48.9 vs 36.0%, and 91.1 vs 84.9%, respectively), and cognitively impaired (47.8 vs 27.7% with MMSE score<24). In addition, they had more comorbidities (Charlson index 1.6 vs 1.2). In multivariate analysis, an independent association remains for subjects living alone (OR 1.8, 95%CI 1.1-3.0), with cognitive impairment (OR 1.9, 95%CI 1.1-3.2), and comorbidities (OR 1.3 per point, 95%CI 1.1-1.5). Detection rate during the index hospitalization was only 16.7% (15/90). In conclusion, depressed mood was frequent but rarely detected in this population. These findings emphasize the need to improve screening efforts, and to develop additional strategies such as using a pre-screening question to enhance clinical recognition.