Table 4 Logistic regression models for predicting severe periodontitis and their performance
From: A rapid, non-invasive tool for periodontitis screening in a medical care setting
Predictor | Model 1: Questionnaire, demographic data and biomarkers | Model 2: Questionnaire and demographic data | Model 3: Questionnaire only | ||||||
|---|---|---|---|---|---|---|---|---|---|
Contributing to model | Reference category | B | Contributing to model | Reference category | B | Contributing to model | Reference category | B | |
Predictor | |||||||||
Q1. Gum disease | + | Yes | 1.573 | ||||||
Q2. Own teeth/gum health | + | Negative | 1.152 | ||||||
Q3. Gum treatment | + | Yes | 2.100 | + | Yes | 2.073 | + | Yes | 2.235 |
Q4. Loose teeth | + | Yes | 1.277 | + | Yes | 1.306 | |||
Q5. Lost bone | |||||||||
Q6. Tooth appearance | + | Yes | 1.590 | + | Yes | 0.973 | |||
Q8. Mouthwash use | + | 1–7 days | 1.745 | + | 1–7 days | 1.440 | + | 1–7 days | 1.181 |
Age (years) | + | > 39 | 1.455 | + | > 39 | 1.615 | |||
Sex | + | Male | 1.272 | + | Male | 1.091 | |||
Smoking | + | Yes | 2.007 | ||||||
Albumin concentration | + | n/a | 0.727 | ||||||
Chitinase activity | |||||||||
Protease activity | |||||||||
Model performance | |||||||||
AUROCC (95% CI) | 0.89 (0.85–0.95) | 0.82 (0.75–0.89) | 0.78 (0.71–0.86) | ||||||
Predicted probability cut-off | 0.250 | 0.214 | 0.273 | ||||||
Hosmer-Lemeshow | 0.827 | 0.963 | 0.717 | ||||||
Sensitivity (95% CI), (%) | 86 (71–95) | 80 (66–90) | 65 (52–79) | ||||||
Specificity (95% CI), (%) | 78 (68–86) | 70 (60–79) | 81 (73–88) | ||||||
PPV (95% CI), (%) | 62 (52–71) | 56 (48–64) | 62 (48–75) | ||||||
NPV (95% CI), (%) | 93 (86–97) | 88 (81–93) | 83 (76–90) | ||||||