Abstract
BACKGROUND: Despite the growing recognition of osteoporosis in men, screening tools validated for this population remain underutilized. This study aimed to evaluate the diagnostic performance of three U.S. Preventive Services Task Force (USPSTF)-recommended risk assessment tools for identifying osteoporosis among older men. METHODS: This retrospective cross-sectional study included 2,824 men aged ≥65 years who underwent dual-energy X-ray absorptiometry (DXA) at two teaching hospitals between 2015 and 2024. Osteoporosis was defined as a T-score ≤ -2.5 at the femoral neck, total hip, or lumbar spine. Three screening tools, the Osteoporosis Self-Assessment Tool (OST), Osteoporosis Index of Risk (OSIRIS), and Simple Calculated Osteoporosis Risk Estimation (SCORE)-were calculated for each participant. Diagnostic accuracy was assessed using receiver operating characteristic (ROC) curve analysis, with the area under the curve (AUC) compared using the DeLong test. Optimal thresholds were identified by maximizing the Youden index. RESULTS: Among the 2,824 men included in the study, 598 (21.2%) were diagnosed with osteoporosis. All three USPSTF-recommended tools demonstrated acceptable discrimination. The AUCs were 0.7339 (95% CI, 0.7121-0.7556) for OST, 0.7168 (95% CI, 0.6939-0.7396) for OSIRIS, and 0.7153 (95% CI, 0.6931-0.7375) for SCORE. Pairwise comparisons showed that OST performed significantly better than OSIRIS (ΔAUC = 0.0171, p = 0.0124) and SCORE (ΔAUC = 0.0186, p = 0.0002), whereas OSIRIS and SCORE demonstrated comparable performance (ΔAUC = 0.0015, p = 0.8523). The optimal thresholds were -0.7 for OST, -1.5 for OSIRIS, and 11.5 for SCORE, producing sensitivities of 0.7743, 0.6739, and 0.6890 and specificities of 0.5714, 0.6698, and 0.6352. CONCLUSIONS: Among USPSTF-endorsed screening tools, OST demonstrated the best overall diagnostic performance for identifying osteoporosis in older men. Its simplicity and favorable sensitivity-specificity balance support its utility as a practical first-line screening approach in clinical settings.