Objective: To explore the construction of a column line chart-based predictive model for postoperative pulmonary infection severity in tracheostomized patients with cranial brain injuries.
Methods: The study included 187 patients with cranial brain injuries who underwent tracheostomy between December 2021 and June 2023. These patients were categorized into moderate-to-severe and mild groups based on the severity of postoperative pulmonary infections. Logistic regression analysis was employed to pinpoint the autonomous risk elements for the severity of postoperative pulmonary infection in tracheostomized patients with cranial brain injuries, and a column line chart predictive model was established using these identified independent risk factors. Receiver Operating Characteristic (ROC) curves and calibration curves were used to assess the predictive performance and clinical application potential of the column line chart model for postoperative pulmonary infection risk in tracheostomized patients with cranial brain injuries.
Results: Among the 187 patients, 83 (44.39%) experienced moderate-to-severe pulmonary infection. Factors such as age ≥60 years, GCS score <8, a history of long-term smoking, ASA >II, non-washable tracheal tubes, malnutrition, using a ventilator, and longer operative time were more prevalent in the moderate-to-severe group compared to the mild group (P < .05). Multivariate logistic regression analysis revealed that age ≥60 years, GCS score <8, a history of long-term smoking, ASA >II, non-washable tracheal tubes, malnutrition, using a ventilator, and longer operative time were independent risk factors for moderate-to-severe pulmonary infection in tracheostomized patients with cranial brain injuries (P < .05). Build a predictive model based on the above six independent risk factors and plot the ROC curve. ROC curve analysis demonstrated that the AUC values for age ≥60 years, GCS score <8, a history of long-term smoking, ASA >II, non-washable tracheal tubes, malnutrition, using a ventilator, and longer operative time in the column line chart model were 0.578, 0.654, 0.711, 0.652, 0.892, 0.598, 0.712, and 0.752, respectively, indicating good predictive performance of the model.
Conclusion: The column line chart-based predictive model for postoperative pulmonary infection severity in tracheostomized patients with cranial brain injuries has a high discriminative power and predictive accuracy. It provides a reliable and intuitive means of predicting the severity of postoperative pulmonary infections in these individuals, enabling healthcare personnel to implement timely intervention measures, thus reducing the occurrence of pulmonary infections.