For a study, researchers sought to understand that the decision tree model was used to create a new method for anticipating intravenous immunoglobulin (IVIG) resistance and coronary artery involvement in Kawasaki disease (KD). Retrospective review of the children’s medical records who were admitted to the hospital for KD. Investigators examined the clinical traits and laboratory results in the groups of KD patients who had CADs and IVIG resistance. Predicting IVIG resistance and CADs was the major goal of the decision tree models. About 896 patients in total, 385 women and 511 males, ranged in age from 1 month to 12 years. Around 111 (12.3%) patients exhibited IVIG resistance, while 156 (17.4%) patients had CADs. When compared to the IVIG responsive group, the levels of total bilirubin and NT-proBNP in the IVIG resistant group were significantly higher (0.62±0.8 mg/dL vs. 1.38±1.4 mg/dL and 1231±2136 pg/mL vs. 2425±4459 mL, respectively, P<0.01). The resistant group’s CADs were also more advanced (39/111; 14.9% vs. 117/785; 35.1%, P<0.01) than those of the control group. Total bilirubin (0.7 mg/mL, 1.46 mg/dL) and NT-proBNP (1,561 pg/mL) were used to classify the decision tree for predicting IVIG resistance. The decision tree has 2 layers and 4 nodes, and its evaluation accuracy is 90.5%. The decision tree’s predictive power was assessed using the Receiver Operating Characteristic (ROC), and the area under the curve (AUC) (0.834; 95% CI, 0.675-0.973; P<0.05) revealed relatively higher accuracy. Total bilirubin and NT-proBNP levels in the CAD group were substantially higher than those in the control group (P<0.01) (0.64±0.82 mg/dL vs. 1.04±1.14 mg/dL and 1192±2049 pg/mL vs. 2268±4136 pg/mL, respectively). The decision trees for predicting CADs were segmented into 2 nodes with 83.5% training accuracy and 90.3% evaluation accuracy based solely on NT-proBNP (789 pg/mL). A new algorithm decision tree model was provided that validates the utility of NT-proBNP as a KD predictor by predicting IVIG resistance and CADs in KD.
- Business of Medicine
- Doctor’s Voice