OBJECTIVES: To assess and compare various approaches for selecting women for a bone mineral density (BMD) examination in three different age groups: (1) the most commonly recommended strategy, i.e., selection based on the presence of at least one major risk factor for osteoporosis (personal history of fracture, maternal history of hip fracture, low weight, early menopause), (2) by weight, and (3) calculation of an individual risk score based on the predictive equation (logistic regression model) including the factors most predictive of osteoporosis in each age group. METHODS: Data from three population-based samples of postmenopausal women of different ages 60-70 years (n=399), 70-80 years (n=392), and 80 years or older (n=3628). Within each age group, the value of these different approaches was primarily assessed in terms of their discriminant value (sensitivity and positive predictive value (PPV)) for osteoporosis. Other women besides osteoporotic women might also be considered at high risk of fracture (and hence treated), in particular osteopenic women (T-score<-1.5) with multiple risk factors for fracture. Hence, we also estimated and compared the overall number of selected women who might be considered at high risk of fracture after BMD testing, according to selection criteria. RESULTS: In each age group and classifying a similar percentage of women at high risk, use of weight as a tool for identifying osteoporotic women has a higher sensitivity and PPV than the currently recommended approach. Increasing the cutoff for weight increases sensitivity. However, identifying the majority (around 80%) of all osteoporotic women requires testing more than half the population. Combining weight with other factors into more complex risk scores usually does not significantly improve the discriminant value of the assessment. When similar percentages of women are selected, more women with osteopenia and multiple risk factors are identified when selection is based on usual referral criteria. However, since more osteoporotic women are identified after selection by weight (higher PPV for osteoporosis), the overall number of women who might be considered at high risk of fracture after BMD testing is higher after selection by weight. CONCLUSIONS: In each age group considered, selecting women for BMD testing based on weight is the simplest and most effective screening method for identifying osteoporotic women as well as other subgroups of women who might also be considered to be at high risk of fracture.
OBJECTIVES: To assess and compare various approaches for selecting women for a bone mineral density (BMD) examination in three different age groups: (1) the most commonly recommended strategy, i.e., selection based on the presence of at least one major risk factor for osteoporosis (personal history of fracture, maternal history of hip fracture, low weight, early menopause), (2) by weight, and (3) calculation of an individual risk score based on the predictive equation (logistic regression model) including the factors most predictive of osteoporosis in each age group. METHODS: Data from three population-based samples of postmenopausal women of different ages 60-70 years (n=399), 70-80 years (n=392), and 80 years or older (n=3628). Within each age group, the value of these different approaches was primarily assessed in terms of their discriminant value (sensitivity and positive predictive value (PPV)) for osteoporosis. Other women besides osteoporoticwomen might also be considered at high risk of fracture (and hence treated), in particular osteopenicwomen (T-score<-1.5) with multiple risk factors for fracture. Hence, we also estimated and compared the overall number of selected women who might be considered at high risk of fracture after BMD testing, according to selection criteria. RESULTS: In each age group and classifying a similar percentage of women at high risk, use of weight as a tool for identifying osteoporoticwomen has a higher sensitivity and PPV than the currently recommended approach. Increasing the cutoff for weight increases sensitivity. However, identifying the majority (around 80%) of all osteoporoticwomen requires testing more than half the population. Combining weight with other factors into more complex risk scores usually does not significantly improve the discriminant value of the assessment. When similar percentages of women are selected, more women with osteopenia and multiple risk factors are identified when selection is based on usual referral criteria. However, since more osteoporoticwomen are identified after selection by weight (higher PPV for osteoporosis), the overall number of women who might be considered at high risk of fracture after BMD testing is higher after selection by weight. CONCLUSIONS: In each age group considered, selecting women for BMD testing based on weight is the simplest and most effective screening method for identifying osteoporoticwomen as well as other subgroups of women who might also be considered to be at high risk of fracture.