OBJECTIVES: The present study investigates the treatment persistence of osteoporosis and assesses its impacts on the risk of fractures among women in Taiwan. METHODS: This is a retrospective cohort study design. Patients who filled prescriptions of alendronate, ibandrobate, risedronate, zoledronate, raloxifene or teriparatide during 2005-2009 were retrieved from the National Health Insurance Research database. Non-persistence was defined when the patients did not refill prescription within 30 days after the previous osteoporosis prescriptions. Relative risks of fracture incidence within 3 years after the initial prescription date in four persistence groups (<30 days, 1 month-1 year, 1-2 years, 2-3 years) were evaluated based upon survival analysis. RESULTS: A total of 1,287,253 patients were included and followed during the study period of 2005-2009. The overall persistence rates among patients with osteoporosis drug treatment were 37%, 29%, 21%, and 9% after 1, 2, 3, and 4 years, respectively. Risks of fractures were higher in the lower persistence group (HR 0.87, p=0.092 in 1 month-1 year persistence; HR 0.61, p<0.001 in 2-3 years persistence). CONCLUSIONS: The relatively high non-persistence rates among patients with osteoporosis treatment and the significant negative association between treatment persistence and fracture incidence indicates potential cost-savings and health gains are substantial if improvement in the treatment persistence can be achieved.