Table 2.
Diagnostic values of the risk profile at different cut-off points in the total risk score
| Cut-off in the total risk score | % at high risk group | Sensitivity (%) | Specificity (%) | Σa (%) | PV+ (%) | PV- | Pfalls |
| 0 vs. ≥ 1 | 782 | 95.1 | 11.7 | 106.8 | 41.1 | 78.8 | 0.10 vs. 0.34 |
| 0–1 vs. ≥ 2 | 69.7 | 86.1 | 22.1 | 108.2 | 41.8 | 71.0 | 0.11 vs. 0.36 |
| 0–2 vs. ≥ 3 | 61.5 | 79.9 | 33.8 | 113.7 | 43.9 | 72.1 | 0.13 vs. 0.39 |
| 0–3 vs. ≥ 4 | 51.4 | 70.8 | 47.3 | 118.1 | 46.6 | 71.4 | 0.14 vs. 0.43 |
| 0–4 vs. ≥ 5 | 46.0 | 66.0 | 54.5 | 120.5 | 48.5 | 71.2 | 0.15 vs. 0.46 |
| 0–5 vs. ≥ 6 | 39.2 | 60.4 | 64.0 | 124.4 | 52.1 | 71.4 | 0.17 vs. 0.50 |
| 0–6 vs. ≥ 7 | 35.4 | 55.6 | 68.0 | 123.6 | 53.0 | 70.2 | 0.17 vs. 0.52 |
| 0–7 vs. ≥ 8 | 32.6 | 52.1 | 71.2 | 123.3 | 54.0 | 69.6 | 0.18 vs. 0.54 |
| 0–8 vs. ≥ 9 | 29.8 | 48.6 | 74.3 | 122.9 | 55.1 | 69.0 | 0.19 vs. 0.56 |
| 0–9 vs. ≥ 10 | 27.9 | 46.5 | 76.6 | 123.1 | 56.3 | 68.8 | 0.20 vs. 0.57 |
| 0–10 vs. ≥ 11 | 21.4 | 41.0 | 85.6 | 126.6* | 64.8 | 69.1 | 0.22 vs. 0.91 |
| 0–15 vs. ≥ 16 | 5.9 | 13.9 | 97.7 | 111.6 | 80.0 | 63.6 | 0.28 vs. 0.77 |
a sum of sensitivity + specificity; * maximum Σ; PV+ positive predictive value; PV- negative predictive value; Pfalls probability of recurrent falls in low risk versus high risk group. The sample used in LASA to develop the fall risk profile included relatively healthy community dwelling older persons of which a large part reported zero previous falls [16]. In contrast, all participants of this study have a history of at least one recent fall and also include people living in a home for the elderly. Thus, the participants in our study are expected to have poorer mobility and more functional limitations and, on average, to score higher on the risk profile. The diagnostic values presented here are the recalculated values using the data of 426 independently living participants of the LASA study who fell at least once [16].