BACKGROUND: Depression, a common disorder often treated by family physicians, may be both underdiagnosed and undertreated. The objective of this study was to determine whether the diagnosis and treatment of depression by family physicians could be improved through an educational strategy. METHODS: In this study, conducted between July and December 1997, 42 family physicians in Newfoundland were randomly assigned to an intervention group (3-hour case-based educational session on clinical practice guidelines [CPGs] for depression and access to a psychiatrist for consultation) or to a control group (receipt of CPGs without educational session or access to the psychiatrist). Physicians were asked to keep a log of patients with newly diagnosed depression and to record information on severity of depression, medications and referrals to mental health professionals. Patients were asked to complete the Centre for Epidemiologic Studies Depression (CES-D) scale before treatment and after 6 months of follow-up. The primary outcome measure was the "gain" score (difference between first and last CES-D scores). RESULTS: During the study period physicians in the intervention group diagnosed 91 new cases of depression (mean 4.1 per physician) and those in the control group diagnosed 56 (mean 2.8 per physician); the difference was not significant. Most patients (91.2% in the intervention group and 89.3% in the control group received a prescription for an antidepressant on their first visit. Similar proportions (46.2% in the intervention group and 37.5% in the control group) took their medication for the full 6 months; however, significantly more patients in the intervention group were taking an antidepressant at the 6-month follow-up (56% v. 39.3%, p = 0.02). The mean number of visits per patient was similar in the 2 groups (7.7 in the intervention group and 7.6 in the control group). Physicians in the intervention group consulted the psychiatrist 9 times. The overall rate of referrals to psychiatrists and other mental health professionals was 10.9%; however, referrals were significantly higher in the intervention group (15.4% v. 3.5%, p = 0.05). After 6 months of follow-up, a significant difference in gain scores was detected between the intervention and control groups for both the patient's self-rated CES-D scores (mean gain score 19.3 v. 15.5 respectively, p = 0.04) and the physicians' ratings of depression severity before treatment and at 6 months (mean gain 1.1 v. 0.7 respectively, p = 0.02). INTERPRETATION: The educational strategy had a modest beneficial effect on the outcomes of patients with depression, but there are still concerns regarding the low rates of drug treatment and referral to mental health professionals by family physicians.