Summary features of the three most studied tools, as deemed relevant for the selection of the instrument in clinical and research settings
| FRAX | QFracture | GARVAN | |
|---|---|---|---|
| Feasibility | |||
| Number of clinical risk factors | 11 | 19 | 5 |
| Requirement for BMD | Optional | No | Optional |
| Algorithm accessible for individual use | Yes | No* | Yes |
| Applicability | |||
| Male and female | Yes | Yes | Yes |
| Age range | 40–90 | 35–100 | 50–96 |
| Prediction intervals | 10 | 1, 2, …, 10 | 5, 10 |
| Type of fracture—hip | Yes | Yes | Yes |
| Type of fracture—MOP | Yes | Yes | Yes |
| Countries | 57 | UK only | 3 |
| Inclusion in national guidelines | Yes | Yes | Yes |
| Validity | |||
| Validated in a separate cohort | Yes | Yes | Y (BMD only) |
| Independent validation† | Yes | Yes | Y (BMD only) |
| Number of validation studies | 26 | 3 | 6‡ |
| Population basis for validation, N | 4 624 438 | 3 485 952§ | 229 162 |
| Population basis for validation, countries | 9¶ | UK only | 3¶ |
| Average quality of studies (QUADAS-2) | Globally similar (see online supplementary figure S2) | ||
| Duration of follow-up equal to tool estimation interval | Yes | Yes (10 year only) | Yes (5 and 10 year) |
| Consideration of national fracture epidemiology | Yes | No | No |
| Consideration of background mortality | Yes | No | No |
| AUC (95% CI)—hip, females, without BMD | 0.74 (0.68 to 0.80) | 0.89 (0.88 to 0.89) | NA |
| AUC (95% CI)—hip, females, with BMD | 0.79 (0.73 to 0.85) | NA | 0.74 (0.61 to 0.87) |
| AUC (95% CI)—hip, males, without BMD | 0.71 (0.65 to 0.77) | 0.87 (0.86 to 0.88) | NA |
| AUC (95% CI)—hip, males, with BMD | 0.77 (NA)** | NA | 0.85 (NA)** |
| AUC (95% CI)—MOP | †† | †† | †† |
*QFracture was removed from the website in 2012. Only the updated version is now available, but is not suitable for meta-analysis as it has only been the subject of one validation study.
†That is, by independent research groups.
‡Only with BMD.
§Does not include the updated QFracture (2012) study.
¶We did not consider the study that included data from 10 countries.
**One study only.
††Comparison is inadequate because of different definitions of MOP for each tool.
AUC, area under the curve; BMD, bone mineral density; MOP, major osteoporotic fractures; NA, not applicable/not available; QUADAS, Quality Assessment Tool for Diagnostic Accuracy Studies.