OBJECTIVES:To describe the prevalence of risk factors for women at high risk of fracture in a population-based sample of postmenopausal women who were not using hormone replacement therapy (HRT), to demonstrate how the estimated prevalence of women at high risk of future fracture is affected by the different criteria used for classification, and to characterize the populations identified and missed by each of the criteria. A key study objective was to compare the proportion of postmenopausal women at high risk of fracture in a managed care population using several different definitions of who is at high risk. DESIGN:The Osteoporosis Population-based Risk Assessment study, a randomized trial of three screening strategies. SETTING:Conducted at Group Health Cooperative in western Washington state. PARTICIPANTS:Women aged 60 to 79 who had not used HRT for at least 12 months were chosen at random. MEASUREMENTS:In one of the trial arms, 428 women had their bone mineral density (BMD) measured at the hip and spine (L1-L4) using dual energy x-ray absorptiometry. Minimum t scores and z scores at all sites were used for classification. Risk factors for fractures were assessed at the time of the BMD scan. RESULTS:Guidelines based on the Study of Osteoporotic Fractures classified 25.1% of the women as being at high risk of fracture, compared with 30.0% and 68.0% using World Health Organization (WHO) recommendations and National Osteoporosis Foundation guidelines, respectively. Classification based on low BMD alone (WHO) failed to include more than 50% of women who had already experienced a clinical fracture. CONCLUSIONS:Prevalence of women at high risk of fracture not using HRT varies notably depending on the criteria used for identification. The criteria used to identify women to target for primary and secondary prevention of osteoporotic fractures has major implications for population-based prevention strategies.