Psychological distress is a driver both of direct and indirect health care costs. Depression compromises functional well-being, such as work productivity. Comorbid anxious features often complicate the recognition of depression and may herald a poor prognosis. We report the results of a cross-sectional naturalistic study to determine the impact of three interventions (no antidepressant, fluoxetine, or tricyclic antidepressant therapy) on relative risk of work days lost in 454 French outpatients with either major or minor depression. Most depressed patients also manifested anxious features (76% with a Hamilton Rating Scale for Anxiety score ≥ 12). The presence of anxiety was related to the severity of depression, work absenteeism, and current social instability. Depression severity (Hamilton Rating Scale for Depression score ≤ 26, including the contributions of anxious symptoms), psychiatric comorbidity, and psychomotor retardation best predicted continued work absenteeism. Patients with major depression were more likely to receive an antidepressant if they had a past history of depressive episodes and/or previous work disability. Patients with minor depression were less likely to receive drug therapy than patients with major depression, despite their current work disability. Among patients who received fluoxetine or a tricyclic antidepressant for at least 8 weeks, fluoxetine was associated with statistically significantly lower mean anxiety and depression scores and fewer work days missed.