Objectives: To compare two well-validated tools - the FRAX without bone mineral density (BMD) and the Osteoporosis Self-Assessment Screening Tool (OST) - in predicting osteoporosis and to define thresholds above and below which it would be reasonable to recommend omitting BMD testing. Design: Retrospective review. Setting: General practices in Western Australia. Participants: Individuals aged 70 and older responding to a prospective audit of osteoporosis investigation and management for whom dual-energy X-ray absorptiometry scan results and clinical risk factor data were available (N = 626). Measurements: Receiver operating characteristic (ROC) curves were compared, upper and lower thresholds for omission of screening BMD were proposed, and the statistical performance measures for the tests are reported. Results The areas under the ROC curves for the OST (0.76-0.82) were slightly better than for FRAX without BMD 10-year major osteoporotic fracture risk (0.64-0.76) in predicting osteoporosis at the defined sites. At defined lower thresholds, the tests were comparable in identifying a group with low osteoporosis risk (sensitivity 89.6-92.2%, specificity 35.0-39.9%), translating into 33.5% to 36.1% of tests saved at a cost of missing 7.8% to 10.4% of individuals with osteoporosis on BMD criteria. It was not possible to identify a useful upper threshold. Conclusion: At the defined thresholds, the OST is as good as FRAX without BMD in identifying a low-risk population subgroup for whom screening BMD can reasonably be omitted. This could reduce costs and improve access to treatment. © 2014, The American Geriatrics Society.