Table 6.
Summary findings of models in gingival bleeding and calculus
| Performance Measures | Predictive model with total samples (%) | Predictive model with training data (%) | Predictive model with validation data (%) |
|---|---|---|---|
| Gingival bleeding | |||
| Sensitivity | 69.05 (67.08, 71.02) | 68.77 (66.48, 71.07) | 69.85 (66.00, 73.71) |
| Specificity | 66.75 (64.87, 68.63) | 66.19 (64.02, 68.36) | 68.48 (64.73, 72.24) |
| PPVa | 64.4 7 (62.50, 66.44) | 63.54 (61.25, 65.83) | 67.26 (63.39, 71.13) |
| NPVb | 71.17 (69.30, 73.04) | 71.22 (69.07, 73.37) | 71.02 (67.29, 74.76) |
| Accuracy | 67.82 (66.46, 69.18) | 67.38 (65.81, 68.96) | 69.14 (66.45, 71.83) |
| AUCc | 73.81 (72.38,75.25) | 73.29 (71.63,74.96) | 75.34 (72.53, 78.15) |
| Predicted prevalence | 49.94 (48.49, 51.40) | 53.86 (52.18, 55.34) | 49.96 (47.04, 52.87) |
| Youden index | 35.93 (33.20, 38.65) | 34.97 (31.81, 38.12) | 38.34 (32.95, 43.72) |
| Dental calculus | |||
| Sensitivity | 67.12 (996, 68.79) | 67.83 (65.91, 69.78) | 65.02 (61.65, 68.39) |
| Specificity | 59.43 (56.94, 61.91) | 59.96 (57.11, 62.82) | 57.73 (52.65, 62.82) |
| PPVa | 77.01 (75.40, 78.61) | 77.15 (75.31, 79.00) | 76.57 (73.32, 79.82) |
| NPVb | 47.16 (44.91,49.42) | 48.3 3 (45.76, 50.90) | 43.72 (39.28, 48.17) |
| Accuracy | 64.57 (63.18, 65.97) | 65.20 (63.60, 66.81) | 62.69 (59.87, 65.51) |
| AUCc | 67.83 (66.19, 69.46) | 68.27 (66.39, 70.16) | 66.55 (63.29, 69.28) |
| Predicted prevalence | 58.34 (56.91 59.78) | 58.54 (56.89, 60.20) | 57.73 (54.86, 60.62) |
| Youden index | 26.54 (23.54, 29.54) | 27.80 (24.36, 31.24) | 22.75 (16.65, 28.86) |
Abbreviations: aPPV positive predictive value, bNPV negative predictive value, cAUC area under the receiver operating characteristic (ROC) curve