Table 4 Prediction model accuracy (validation set)
High-risk prediction | Low-risk prediction | |||
|---|---|---|---|---|
Prediction model | Swedish index for medical priority dispatch | Prediction model | Swedish index for medical priority dispatch | |
Number of variables in model | 11 | – | 10 | – |
Cut-off, % | 10 | – | 90 | – |
Sensitivity, % | 89 | 77 | 10 | 1 |
Specificity, % | 51 | 38 | 97 | 99 |
Positive predictive value, % | 27 | 20 | 89 | 64 |
Negative predictive value, % | 96 | 89 | 33 | 32 |
AUROC | 0.789 | – | 0.737 | – |
Highest priority dispatch, n (%) | 488 (56) | 563 (65) | – | – |
Lowest priority dispatch, n (%) | – | – | 64 (7) | 11 (1) |
Patients with high-risk condition given lowest priority, n (%) | 3 (5) | 3 (27) | ||