Key Symptoms in Bronchial Asthma Diagnosis: Identification via Decision Tree Modeling.

Abstract

Clinical features are essential for the diagnosis of bronchial asthma. However, the relative importance of different features remains unclear. To evaluate the relative importance of various clinical features in bronchial asthma using a decision tree model. We retrospectively analyzed data from multiple prospective cohorts (2005-2020), including 4,417 patients with asthma-like symptoms. Using the 2025 Global Initiative for Asthma guidelines as the reference standard, we developed a symptom-based decision tree model using the Exhaustive Chi-squared Automatic Interaction Detector algorithm. Diagnostic performance was evaluated using receiver operating characteristic curves, sensitivity, specificity, positive predictive value, and negative predictive values. Using a decision tree model, 12 key characteristic variables were identified from clinical manifestations. The model demonstrated excellent performance in diagnosing asthma, achieving a classification accuracy of 93.9% (95% CI, 93.17-94.58) and an area under the receiver operating characteristic curve of 0.973 (95% CI, 0.968-0.978), significantly improving asthma diagnosis rates. Decision tree models based on key clinical symptoms can effectively support physicians in their clinical reasoning training and optimize diagnostic workflows.