Abstract: Objective To analyze the risk factors of unplanned readmission in postoperative patients with colorectal cancer by logistic regression model and decision tree. Methods Clinical data of 1,383 patients with colorectal cancer who underwent surgery in department of gastrointestinal surgery from March 2018 to December 2019 were analyzed retrospectively. Logistic regression model and decision tree model were established respectively, and the results of the two models were compared. Results Seventy-one patients were readmitted unplannedly, with an unplanned readmission rate of 5.13%. Logistic regression model showed that postoperative complications, TNM stage≥III and more than 2 preoperative complications were independent risk factors. Decision tree model showed that postoperative complication was the main risk factor, followed by more than 2 preoperative complications, TNM stage≥III, enterostomy and operation methods. The area under the ROC curve of the logistic regression model was 0.773 and that of the decision tree model 0.790. There was no significant difference in the area under the ROC curve between the two models（Z=0.414,P>0.05）. Conclusion Postoperative complications,TNM stage≥III and more than 2 preoperative complications are important influencing factors of unplanned readmission in postoperative patients with colorectal cancer. Logistic regression model and decision tree model have high consistency in the analysis of risk factors. Clinical nurses should take targeted prevention and nursing interventions to reduce the rate of unplanned readmission.

Key words: logistic regression, decision tree, colorectal cancer, unplanned readmission