Evaluation of decision rules to identify postmenopausal women for intervention related to osteoporosis.

Abstract

Decision rules for intervention that utilize screening tools for bone mineral density (BMD) testing and incorporating the BMD findings and other risk factors to identify high-risk women to prevent fracture have not been evaluated. We examine the sensitivity and specificity of decision rules for intervention based on two pre-BMD screening tools: Simple Calculated Osteoporosis Risk Estimation (SCORE) and a Study of Osteoporotic Fractures (SOF)-based tool. Women 60 years of age and older without previous osteoporosis diagnosis were randomly selected from a managed care population and invited to receive a BMD test. Four hundred sixteen women had complete information and were included in the study. Women were classified as high risk requiring intervention using three different criteria: World Health Organization (t-score -2.5 or less), National Osteoporosis Foundation (t-score -2.0 or less, or -1.5 or less with one or more risk factors), and the SOF-based criteria (prior fracture; or age 60-64 with t-score less than -2.5 or age 65 or older with z-score less than -0.43 and five or more risk factors). SCORE identified 82% of the women as appropriate for BMD testing, whereas the SOF-based tool identified 26%. Sensitivity and specificity were 89.8%-96.5% and 23.8%-34.8%, respectively, for the decision rule using SCORE as the screening tool and 30.5%-84.9% and 76.0%-95.8%, respectively, for the decision rule based on SOF screening criteria. SCORE correctly identified more women who were at high risk for intervention, whereas the SOF-based tool correctly identified more women who do not meet intervention criteria. The appropriate selection of a screening tool depends upon the objective for intervention and trade-off between not identifying women for BMD testing who are at high risk and identifying more women for BMD testing who are at low risk.