OBJECTIVES: This study employed a retrospective cohort design to examine patterns of duloxetine utilization and health care costs among elderly fibromyalgia patients. METHODS: Pharmacy and medical claims were analyzed for fibromyalgia patients aged 65+ with Medicare supplemental insurance who initiated duloxetine in 2006. The index date was defined as the dispense date of the first duloxetine prescription filled, with no duloxetine coverage in the prior 90 days. Patients were required to have at least 30 supply days of duloxetine in the 12 months post-index period. Individuals with any diagnosis of diabetic peripheral neuropathic pain or depression during the 12 months pre-index period were excluded. Five study cohorts were constructed based on the index dosage: <30mg, 30mg, 31-59mg, 60mg, and >60mg. Patterns of duloxetine use including changes in dosage, average daily dose (ADD), adherence to duloxetine (medication possession ratio≥0.8 as high adherence) were examined across study cohorts. Regression models were performed to estimate the differences in healthcare costs.RESULTS: A total of 566 fibromyalgia patients were included, with 41, 163, 47, 294 and 21 in the <30mg, 30mg, 31-59mg, 60mg, and >60mg cohorts, respectively. A total of 31.4% of patients experienced any dosage changes (increased dosage: 25.8%; decreased dosage: 15.7%). Among those who changed dosage, patients in the 31-59mg cohort had the shortest time to change (81 days), and patients in the <30mg cohort had the longest (149 days) time. ADD trended upward as index dose increased. Compared with patients in the 60mg cohort, those in the <30mg and >60mg cohorts were less likely to be adherent (odds ratios 0.40 and 0.30, respectively, both p<0.05). Post-index total health care costs were similar across cohorts. CONCLUSIONS: Dosage changes occurred most quickly in fibromyalgia patients with an index dose of 31-59mg of duloxetine. Duloxetine ADD and adherence differ by index dosage, while health care costs remain similar.