PMS8 Estimating Incidence and Prevalence of Osteoarthritis (OA) in Alberta Using Administrative Claims Data

Abstract

ence-in-difference method was used to estimate changes in fracture rates two years before and after ZOL or OBP initiation. Generalized estimating equation models were used to test the hypothesis of differential changes in fracture rates between ZOL and OBP users, controlling for age, gender, treatment (ZOL vs. OBP), and time (pre-period vs. follow-up period). RESULTS: A total of 3,102 ZOL and 36,961 OBP users met the study criteria. Over the two-year follow-up, ZOL users experienced a significant reduction in fracture rates compared to the two-year prior to ZOL (13.4% vs. 11.2%; pϭ0.008) while fracture rates significantly increased for OBP users (9.0% vs. 9.5%; pϭ0.019). Multivariate regression estimated that the probability of experiencing any fracture decreased by 1.97% between pre-and follow up period for ZOL users (pϭ0.004), increased by 0.46% for OBP users (pϭ0.001), and the difference-in-difference effect was 2.43%, suggesting that ZOL users experienced a significant decrease in fracture rates relative to OBP users (pϽ0.001). CONCLUSIONS: This is the first comparative analysis evaluating fracture rates two years before and two years after the initiation of ZOL and OBP. Use of ZOL significantly reduced fracture rates, compared with the use of OBP.