The aim of this study was to determine the threshold of fracture probability at which interventions become cost-effective. We modeled the effects of a treatment costing $500/year, given for 5 years, that decreased the risk of all osteoporotic fractures by 35%, followed by a waning of effect for 5 years. Sensitivity analyses included a range of effectiveness (10%-50%) and a range of intervention costs (200-500 dollars/year). Data on costs and risks were from Sweden. Costs included direct costs and costs in added years of life, but excluded indirect costs due to morbidity. A threshold for cost-effectiveness of 60,000 dollars per quality-adjusted life-year (QALY) gained was used. Costs of added years were excluded in a sensitivity analysis for which a threshold value of 30,000 dollars per QALY was used. In the base case, intervention was cost-effective when treatment was targeted to women at average risk at age of >or=65 years. Irrespective of the efficacy modeled (10%-50%) or of cost of intervention (200-500 dollars/year) segments of the population at average risk could be targeted cost-effectively: The lower the intervention cost and the higher the effectiveness, the lower the age at which intervention was cost-effective. With the base case (500 dollars/year; 35% efficacy) treatment in women was cost-effective with a 10 year hip fracture probability that ranged from 1.4% at the age of 50 years to 4.4% at the age of 65 years. The exclusion of osteoporotic fractures other than hip fracture would increase the threshold to a 9%-11% 10 year probability because of the substantial morbidity from fractures other than hip fracture, particularly at younger ages. We conclude that the inclusion of all osteoporotic fractures has a marked effect on intervention thresholds, that these vary with age, and that available treatments can be cost-effectively targeted to individuals at moderately increased risk.