OBJECTIVES: While the substantial economic burden of depression is well documented, less is known about the cost of depression by clinical severity levels.  This study’s objective is to document the healthcare resource utilization and costs by severity for patients with major depressive disorder (MDD).  METHODS: Using data from the National Comorbidity Survey-Replication (NCS-R; n=5692), respondents were categorized into clinical severity categories (not clinically depressed, mild, moderate, severe) using standard scales (CIDI/QIDS-SR).  Outcomes were measured over 12 months and included the prevalence of medical services and antidepressant use, the average number of visits and days on antidepressants, the prevalence of treatment adequacy, and medical and drug costs.  Costs were estimated by weighting utilization measures by unit costs obtained for similar services used by depressed patients in a U.S. privately insured claims database for the corresponding time period (2000-2001). Outcomes were compared across depression severity groups using chi-square tests and regression models that generated F statistics adjusted for demographics. RESULTS: Among the 658 depressed respondents, 14.2% were mildly, 38.1% moderately and 47.7% severely depressed.  Respondents in the 3 groups shared similar demographic characteristics.  There was a significant association between severity and the prevalence of mental health services usage (20.8%, 27.9%, and 39.5% respectively, p<0.01) and the average number of mental health practitioner visits.  The use of antidepressants also increased with depression severity (19.8%, 29.8%, and 40.2% respectively, p<0.05).  Similarly, the adequacy of mental health services increased with depression severity (8.9%, 12.1%, and 21.3% respectively, p<0.05).  The average 12-month costs per MDD patient were substantially higher for severe vs. mild (mental health services: $718 vs. $416; general medical services: $133 vs. $57; anti-depressant usage $275 vs. $89).    CONCLUSIONS: There was a significant association between depression severity and treatment usage and costs, as well as between treatment adequacy and severity.