To determine the feasibility and effectiveness of treatments for depressed elderly medical inpatients, MEDLINE was searched for potentially relevant articles published from January 1987 to August 1997, using the keywords "depression or depressive disorder" (exploded) and "aged." The bibliographies of relevant articles were searched for additional references. Fifteen reports met the following inclusion criteria: (a) published in English or French; (b) minimum age criterion of 55 and over or mean age 65 and over; (c) subjects admitted to the medical service of an acute care hospital; (d) used accepted criteria for depression; (e) examined treatment(s) for depression; and (f) reported outcomes as a depression diagnosis and/or symptom level. Information was abstracted independently from each article by two reviewers, tabulated, and compared. The limited evidence suggests contraindications to treatment in 38% to 87% of subjects who received a heterocyclic antidepressant compared to 4% of subjects who received the selective serotonin reuptake inhibitor (SSRI) fluoxetine; rates of discontinuation and study completion were similar for heterocyclics, the SSRIs, and psychostimulants. All of the treatments (including social support/psychotherapy) appeared to be at least modestly effective in the short term.