Abstract
Bone mineral density (BMD) measurements are widely used to estimate the risk of osteoporotic fractures. In addition, many other risk factors have been identified, sonic of which are known to add to the risk independently of BMD measurements. The combination of BMD with such risk factors increases the gradient of risk/standard deviation (SD) than that achieved by BMD alone. In this paper, we report the fracture probabilities according to age, gender, and relative risk, and have investigated the effects of changes in the gradient of risk for osteoporotic fractures on the sensitivity and specificity of assessments, modeled on the population of Sweden. Ten-year risks of hip, clinical vertebral, forearm, or proximal humeral fracture were computed with increments in gradient of risk that varied from 1.5 to 6.0 per SD change in skeletal risk. The identification of high-risk groups had little effect on the specificity of assessments, but increased the sensitivity over a wide range of assumptions. The inclusion of all four fracture types had little effect on sensitivity, but increased the positive predictive value of the test. Positive predictive value also increased with age, so that values greater than 50% were obtained testing women at the age of 65 years with modest gradient of risk of 2.0-2.5/SD when small segments of the population were targeted (0.5-5%). Screening of women to direct intervention at the age of 65 years and targeting 25% of the population could save up to 23% of all fractures in women over the next 10 years by the use of multiple tests with a moderate gradient of risk (RR = 2.0/SD). Such gradients might be achieved with the use of multiple risk factors to identify patients at risk. (C) 2002 by Elsevier Science Inc. All rights reserved.