Skip to main content

Table 3 Performance of the 8-gene XGBoost model and clinical warning signs in the independent cohort. 95% confidence intervals (CIs) from bootstrapping are shown in parentheses for each metric. For the NNP of warning signs, the lower CI bound is omitted as the 95% CI contained negative values due to the sum of PPV and NPV being less than 1 (indicating no gain in certainty according to the Predictive Summary Index). LR+, positive likelihood ratio; LR−, negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value; NNP, number needed to predict

From: An 8-gene machine learning model improves clinical prediction of severe dengue progression

  Sensitivity % Specificity % LR+ LR− PPV % NPV % NNP
8-gene XGBoost model 86.4 (68.2–100.0) 79.7 (75.5–83.9) 4.3 (3.2–5.5) 0.2 (0.01–0.4) 20.9 (16.7–25.6) 99.0 (97.7–100.0) 5.0 (4.0–6.8)
Warning signs 77.3 (58.3–94.1) 39.7 (34.7–44.9) 1.3 (0.9–1.6) 0.6 (0.2–1.1) 7.4 (4.3–10.9) 96.6 (93.3–99.3) 25.4 (NA–185.6)