The main objective of this study is to develop a machine learning model that predicts moderate-severe prediction model of pulmonary complications in liver transplantation patients within 14 postoperative day using a real-world, local preoperative and intraoperative electronic health records, not administrative codes.
Postoperative pulmonary complications can increase the length of hospital stay and medical costs. In particular, moderate to severe pulmonary complications, which often require clinical intervention, once occur, will lead to significantly prolonged postoperative hospitalization or even cause permanent damage or death in severe cases. A number of risk-stratified cation models have been developed to identify patients at increased risk of postoperative pulmonary complications. However, these models were built by using the traditional regression analysis. However, the traditional prediction methods have the disadvantages of limited processing power of nonlinear models and outlier, and relatively single selection variables. The obtained models have poor accuracy, and the quantification degree is not enough, so it is difficult to popularize clinical application. Artificial machine learning can use it by analyzing a large number of specific features in the rich data set to identify and learn to accurately predict the diagnosis and prognosis of diseases, and surpass traditional prediction models in dealing with classification problems. The algorithms are flexible, and it is more and more widely used in clinical practice research. However, there are few reports on machine learning models predicting prognostic models related to postoperative pulmonary complications in liver transplantation patients. Therefore, we aimed to build predictive models using artificial machine learning methods to screen for their risk factors in order to provide early intervention and individualized treatment for high-risk patients.
Study Type
OBSERVATIONAL
Enrollment
400
West China Hospital, Sichuan University
Chengdu, Sichuan, China
RECRUITINGPrediction of postoperative moderate-to-severe pulmonary complications
IA total of 72 variables are expected to be included, using 6 types of machine learning methods, including decision tree (DT), logistic regression (LR), random forest (, RF), support vector machine (SVM), extreme gradient lift (XGBoost), and gradient lift decision tree (GBDT) to build a moderate postoperative prediction model
Time frame: August 2024-December 2024
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.