Impact of FRAX-based osteoporosis intervention using real world data

Objectives To estimate the implications and accuracy of the most common fracture prevention strategies: (1) fixed threshold by the National Osteoporosis Foundation (NOF), (2) age-dependent threshold by the National Osteoporosis Guideline Group (NOGG) and (3) osteoporotic bone mineral density (BMD). Methods A retrospective cohort of all 50–90 years old female members in a nationally representative payer provider healthcare organization in Israel, with 10 years of follow-up on incident events of major osteoporotic fractures. Since events are less frequent than non-events, balanced accuracy (the average between the accuracy obtained for patients with and without events) was used to measure performance. Results Overall among 141,320 women NOF and NOGG would recommend therapy for 17.3% and 2.8% respectively, with NOF exhibiting higher balanced accuracy: 74.1% vs. 54.2% for incident hip fractures detection and 60.0% vs. 51.6% for a composite outcome of major osteoporotic fractures. In patients with available BMD (n = 16,578) the treatment intervention criteria of NOF, NOGG, osteoporotic femur neck or vertebral BMD were met by 30.5%, 9.3% and 24.6% of the population, with balanced accuracy of 70.1%, 56.5% and 62.3% respectively for hip fractures and 61.4%, 52.8%, 58.1% for major osteoporotic fractures. At the age of 75 years or older the NOF hip fracture risk threshold (3%) was exceeded in most women regardless of risk factors other than age. Conclusions In this large population-based study, detection of patients at high risk of sustaining a major osteoporotic fracture within 10 years was more accurate with the NOF fixed threshold criteria as compared with the age-varying NOGG or BMD-only. However, special consideration and further studies are warranted in patients aged 75 years or older with preserved bone density, which may benefit from non-medicinal interventions.