| NHANES III Development Cohort43 (n = 2995) | Clinical Validation Cohort (n = 346) | |||||
|---|---|---|---|---|---|---|
| Hip Fracture | Hip Fracture | Major Osteoporotic Fracture | ||||
| MORES | Universal | MORES | Universal | MORES | Universal | |
| Population (n) | 10,000 | 10,000 | 10,000 | 10,000 | 10,000 | 10,000 |
| Sensitivity (MORES) | 0.930 | — | 0.800 | — | 0.800 | — |
| Specificity (MORES) | 0.590 | — | 0.695 | — | 0.695 | — |
| Fracture risk† | 0.135 | 0.135 | 0.056 | 0.056 | 0.123 | 0.123 |
| Osteoporosis prevalence | 0.048 | 0.048 | 0.043 | 0.043 | 0.043 | 0.043 |
| Relative risk of fracture with treatment‡ | 0.630 | 0.630 | 0.630 | 0.630 | 0.575 | 0.575 |
| Adherence to treatment§ | 0.700 | 0.700 | 0.700 | 0.700 | 0.700 | 0.700 |
| Predicted cases | 480 | 480 | 430 | 430 | 430 | 430 |
| True positive | 446 | 480 | 344 | 430 | 344 | 430 |
| False negative | 34 | — | 86 | — | 86 | — |
| True negative | 5,617 | 9,520 | 6,651 | 9,570 | 6,651 | 9,570 |
| False positive | 3,903 | — | 2,919 | — | 2,919 | — |
| Referred for DXA (true positive + false positive) | 4,350 | 10,000 | 3,263 | 10,000 | 3,263 | 10,000 |
| Predicted hip fractures: MORES/DXA | ||||||
| True positive | 44.66 | 48.02 | 14.27 | 17.84 | 29.75 | 37.19 |
| False negative | 4.54 | 0.00 | 4.82 | 0.00 | 10.59 | 0.00 |
| Total | 49.19 | 48.02 | 19.09 | 17.84 | 40.33 | 37.19 |
| Unscreened (no DXA) | 64.80 | 64.80 | 24.08 | 24.08 | 52.93 | 52.93 |
| Number needed-to-screen (95% CI) | 279 (257–306) | 596 (—) | 654 (485–1132) | 1,604 (—) | 259 (192–449) | 636 (—) |
↵* Formulas used in the calculation were adapted from Nelson et al.52 and are available upon request.
↵† Fracture risks for the NHANES III cohort were obtained from Kanis et al.58 Fracture risks for the current study were obtained from FRAX for both hip and major osteoporotic fractures and reflect the characteristics of the study population.54
↵‡ Risk reduction estimates are supported by clinical trials, and the estimates for hip fractures are those used by Nelson et al.52 The point estimate for the relative risk reduction for major osteoporotic fractures is a weighted average from published reports.19,20
↵§ Compliance with treatment estimates are from Nelson et al.52
DXA, dual-energy X-ray absorptiometry; NHANES, National Health and Nutrition Examination Survey.