Table 3.
Evaluation of predictive model for NIV outcome.
| Dataset | Predictive results | NIV outcome | Total | Sensitivity (%) | Specificity (%) | Youden index (%) | |
|---|---|---|---|---|---|---|---|
| Success | Failure | ||||||
| Training dataset | Success | 25 | 8 | 33 | 71.43 | 77.78 | 49.21 |
| Failure | 10 | 28 | 38 | ||||
| Total | 35 | 36 | 71 | ||||
|
| |||||||
| Testing dataset | Success | 11 | 4 | 15 | 68.75 | 71.43 | 40.18 |
| Failure | 5 | 10 | 15 | ||||
| Total | 16 | 14 | 30 | ||||