OBJECTIVE: To profile treatment-resistant depression (TRD) patients healthcare costs and medical care patterns as their illness progresses. METHODS: The MEDSTAT MarketScan(r) Database for 1995-2000 was used. Patients with a depression diagnosis, suicide attempt, or those treated with electro-convulsive therapy were considered. TRD patients were those who either switched or augmented their initial four-week (minimum) antidepressant prescription with at least one more antidepressant prescribed for at least four weeks. Demographic, treatment, and cost profiles were constructed for periods covered by each subsequent antidepressant medication switch or augmentation. Total medical expenditures per day (year-2000 dollars) were calculated and compared for periods between the index date (entry into study) and each subsequent medication switch or augmentation. Negative binomial count regression models were used to assess the impact of factors on number of medication switches or augmentations occurring during the study period. RESULTS: The sample included 7,737 TRD patients; 72% female, 60% employees and 39% in managed care plans. The mean number of medication switches or augmentations following the index prescription was 2.4. Average total healthcare expenditures increased 102% from $563 per month at the initial antidepressant switch date to $1140 per month after six additional medication switches or augmentations. The number of medication switches or augmentations was significantly influenced by the following factors: existence of comorbid mental health problems, type of antidepressant prescribed at the index date, type of depression diagnosis, treatment under a managed care plan, single or family insurance coverage, and length of time patients were followed. CONCLUSIONS: Most treatment-resistant patients had multiple medication switches or augmentations. Average monthly expenditures more than doubled as the TRD illness progressed through six additional medication switches or augmentations. A better understanding of factors associated with the number of medication switches may lead to promising interventions improving care for patients with treatment-resistant depression.