Application of Logistic Regression and Decision Tree Models in the Analysis of Influencing Factors of Surgical Patients Hospitalized for More Than 30 Days
Objectives This study aims to analyze the main factors affecting surgical patients'hospitalization for more than 30 days to promote continuous improvement of medical quality.Methods The medical records of surgical patients from the gastrointestinal surgery department of a certain hospital from January 1,2019 to December 31,2021 were retrieved.Level Ⅲ and Ⅳ surgery patients aged ≥ 14 years were screened.Systematic sampling plus cluster sampling were used to select a group of cases,and the medical record information of 356 patients was retrospectively investigated.SPSS20.0 software was used for analysis,and single factor analysis,binary logistic regression,and decision tree regression were used to analyze the main factors affecting hospitalization for more than 30 days.Results The results of binary Logistic regression showed that postoperative complications(OR=5.565),critical value(OR=3.285),preoperative hospital stay ≥ 7 days(OR=1.264),Operation Time ≥ 180 minutes(OR=2.678),and blood transfusion(OR=2.903)were the independent influencing factors.The results of decision tree analysis showed that the influencing factors of hospitalization over 30 days were hospitalization ≥ 7 days,blood transfusion,postoperative complications,critical value,and hypoalbuminemia.The AUC(area under the ROC curve)of the two models was 89％and 80.4％,respectively(P＜0.001),and the model predictions had a certain accuracy.Conclusions The results of the two models are complementary to each other and have certain accuracy and practical significance as the prediction of over 30 days of hospitalization.Hospitals should focus on patients with malignant tumors and patients with a history of abdominal surgery.By improving the level of medical technology,optimizing service processes,strengthening perioperative management,and reducing the incidence of complications,patients'hospitalization days and hospitalization expenses can be reduced.
Logistic regressionDecision treeHospitalized for more than 30 daysInfluencing factors
NETL
NSTL
万方数据
