Prediction of hospital-acquired pneumonia after traumatic brain injury based on the PCA-Logistic regression analysis model
FENG Jin-zhou
LIU Fa-jian
KUANG Yong-qin
JIANG Hua
Abstract：Objective To explore the application of principal components analysis (PCA)-Logistic regression analysis to prediction of the hospital-acquired pneumonia (HAP) in the patients with traumatic brain injury (TBI), and to find the pathophysiological patterns and important risk factors related to their clinical prognoses. Methods Original dataset was constituted by the data of 108 patients with TBI derived from the Database of the Trauma Center of Sichuan Provincial People's Hospital from 2011 to 2017. The primary outcome was HAP. After the dataset evaluation and cleaning, the PCA-Logistic regression model was built to identify the risk factors related to HAP. Receiver Operating Characteristic (ROC) curve was used to evaluate the PCA-Logistic regression model. Results The PCA-Logistic regression model analysis found the important clinical indicators affecting the patients HAP, and the PCA-Logistic regression model was evaluated by ROC curve. The HAP outcome model had good predictive power (sensitivity, 83.9% ; specificity, 94.8%; AUC, 0.949). Conclusions PCA-Logistic regression analysis can effectively mine the clinical variables of the patients with TBI and establish a clinical prognosis prediction model. The abnormal parenteral nutrition support after severe TBI may be an important clinical factor affecting the occurrence of HAP in the patients with TBI.
Keywords：Traumatic brain injuryHospital-acquired pneumoniaPrincipal component analysisLogistic regression analysis
Publication Date：2019-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：4( 35-38 )