Antidepressant drug trials have been criticised because they study atypical populations of depressive patients. The present cross-sectional database analysis was designed to define what constitutes a typical population of patients receiving antidepressants. From a database covering a population of 1,057,053 residents in Piedmont, Italy, and including all community (i.e. outside hospitals) prescriptions reimbursed by the National Health System, all prescriptions of antidepressant drugs dispensed during the first six months of 2000 were extracted. Using the general practice patient code all records were attributed to a sample of patients receiving antidepressants. During the study period 22,135 patients were dispensed one or more prescriptions, yielding a prevalence of use of 27.6 (CI 27.1, 28.0) per 1,000 females and 13.7 (CI 13.4, 14.0) per 1,000 males (female/male ratio 2.01). The prevalence of use progressively increased with age, with the highest rates in subjects over 75 years. The distribution of patients by number of antidepressant prescriptions showed that nearly 50% received only one or two prescriptions over the six months surveyed. Moreover, 18,676 subjects (84%) were prescribed antidepressants together with other medications. These data suggest shifting the focus of antidepressant drug trials from selected to non-selected populations of patients, including the elderly and patients with medical comorbidity, enrolled using entry criteria as close as possible to those adopted in everyday clinical practice. The high proportion of occasional antidepressant users suggests that clinical trials should follow all patients, without excluding those who fail to continue the study medication.