Table 4 Predictive accuracy of the SMO models (best performing cut-off 0.6)
Model | Sensitivity (%) (95%, CI) | Specificity (%) (95%, CI) | PPV (%) (95%, CI) | NPV (%) (95%, CI) | AUROC (95%, CI) | Negelkerke R2 |
|---|---|---|---|---|---|---|
Developmental model (n = 600) | 86 (81–90) | 92 (89–94) | 84 (79–89) | 93 (90–95) | 0·92 (0·90–0·95) | 0·69 |
Validation model (n = 900) | 81 (76–86) | 90 (87–93) | 81 (75–85) | 90 (87–93) | 0·92 (0·88–0·94) | 0·69 |
Sensitivity model with whole data (n = 5243) | 91 (90–92) | 71 (68–74) | 94 (93–95) | 61 (58–64) | 0·92 (0·90–0·95) | 0·58 |