TABLE 3.
Diagnostic indices after multiple logistic regression model based on derivation data set.
| Model | SN (95% CI) | SP (95% CI) | LR + (95% CI) | LR- (95% CI) | PPV (95% CI) | NPV (95% CI) | Youden Index | Accuracy |
| LR on 10 formulae, cut point = 0.5 | 97 (93–99) | 82 (67–93) | 5.4 (2.8–10.6) | 0.04 (0.02–0.09) | 96 (92–98) | 87 (71–96) | 0.79 | 94.2 |
| LR on 10 formulae, Optimal cut point = 0.7 | 95 (91–98) | 95 (83–99) | 18.6 (4.8–71.7) | 0.05 (0.03–0.1) | 99 (96–100) | 82 (68–92) | 0.90 | 95.2 |
| LR on 1–5 formulae, cut point = 0.5 | 96 (92–98) | 31 (17–48) | 1.4 (1.1–1.7) | 0.14 (0.06–0.30) | 86 (80–90) | 63 (38–84) | 0.27 | 83.7 |
| LR on 1–5 formulae, Optimal cut point = 0.8 | 78 (71–84) | 67 (50–81) | 2.3 (1.5–3.7) | 0.34 (0.24–0.48) | 91 (85–95) | 41 (29–54) | 0.45 | 75.5 |
| LR on 6–10 formulae, cut point = 0.5 | 94 (89–97) | 69 (52–83) | 3.1 (1.9–4.9) | 0.09 (0.05–0.16) | 93 (88–96) | 73 (56–86) | 0.63 | 89.4 |
| LR on 6–10 formulae, Optimal cut point = 0.7 | 89 (83–93) | 90 (76–97) | 8.7 (3.4–21.9) | 0.13 (0.08–0.19) | 97 (94–99) | 65 (51–77) | 0.79 | 88.9 |
LR10, Logistic regression on 10 formulae; LR 1_5, Logistic regression on first 5 formulae; LR 6_10, Logistic regression on second 5 formulae; Cut, cut point; Optimal cut point, optimal cut point based on sensitivity and specificity in the ROC curve after logistic; CI, Confidence interval; SN, Sensitivity; SP, Specificity; LR +, Positive Likelihood Ratio; LR-, Negative Likelihood Ratio; PPV, Positive Predictive value; NPV, Negative Predictive value.