Table 4.
Predictive power for NIV failure of the updated HACOR score
| Cutoff value | SE | SP | PPV | NPV | + LR | −LR |
|---|---|---|---|---|---|---|
| Training cohort | ||||||
| After 1–2 h of NIV, N = 1451 | ||||||
| > 7 | 84.9% | 67.3% | 59.8% | 88.6% | 2.59 | 0.22 |
| > 10.5 | 59.9% | 89.6% | 76.8% | 79.6% | 5.76 | 0.45 |
| > 14 | 29.5% | 97.9% | 89.1% | 70.8% | 14.31 | 0.72 |
| After 12 h of NIV, N = 1133 | ||||||
| > 7 | 84.0% | 71.2% | 58.5% | 90.2% | 2.92 | 0.22 |
| > 10.5 | 55.0% | 91.8% | 76.3% | 80.9% | 6.67 | 0.49 |
| > 14 | 20.9% | 99.5% | 95.1% | 72.2% | 39.86 | 0.80 |
| After 24 h of NIV, N = 942 | ||||||
| > 7 | 77.9% | 73.5% | 56.6% | 88.2% | 2.94 | 0.30 |
| > 10.5 | 51.7% | 93.4% | 77.7% | 81.3% | 7.84 | 0.52 |
| > 14 | 21.0% | 99.2% | 92.4% | 73.9% | 27.4 | 0.80 |
| Validation cohort | ||||||
| After 1–2 h of NIV, N = 728 | ||||||
| > 7 | 89.9% | 45.3% | 57.4% | 84.6% | 1.64 | 0.22 |
| > 10.5 | 67.7% | 76.5% | 70.3% | 74.3% | 2.88 | 0.42 |
| > 14 | 29.0% | 92.5% | 76.0% | 61.4% | 3.86 | 0.77 |
| After 12 h of NIV, N = 633 | ||||||
| > 7 | 90.5% | 51.2% | 56.7% | 88.4% | 1.85 | 0.19 |
| > 10.5 | 60.3% | 79.0% | 66.9% | 73.8% | 2.87 | 0.50 |
| > 14 | 27.5% | 95.2% | 80.0% | 65.0% | 5.66 | 0.76 |
| After 24 h of NIV, N = 552 | ||||||
| > 7 | 90.5% | 53.2% | 56.3% | 89.3% | 1.93 | 0.18 |
| > 10.5 | 66.1% | 78.3% | 67.0% | 77.5% | 3.04 | 0.43 |
| > 14 | 23.5% | 95.2% | 76.5% | 65.1% | 4.87 | 0.80 |
NIV noninvasive ventilation, HACOR heart rate, acidosis, consciousness, oxygenation, and respiratory rate, SE sensitivity, SP specificity, PPV positive predictive value, NPV negative predictive value, + LR positive likelihood ratio, –LR negative likelihood ratio