Analysis Of Factors Influencing Prolonged Hospital Stay After Enhancde Recovery After Surgery In Gynecologic Cancer Patients And Model Construction

Objective: To explore and analyze the main risk factors for the occurrence of prolonged postoperative hospital stay after enhanced recovery after surgery in gynecologic cancer patients,and to construct and validate a predictive model for the risk of prolonged postoperative hospital stay after enhanced recovery after surgery in gynecologic cancer patients.Methods: Based on literature review,expert consultation and self-designed questionnaire on factors influencing prolonged postoperative hospital stay in enhanced recovery after surgery for gynecologic cancer patients.575 gynecologic cancer patients who underwent enhanced recovery after surgery from January 2020 to December 2021 in a tertiary care hospital in Shenyang,Liaoning Province were retrospectively collected.gynecologic cancer patients from January 2020 to June 2021 were included in the modeling group(133actual occurrence,103 predicted occurrence),and gynecologic cancer patients from July2021 to December 2021 were included in the validation group(40 actual occurrence,55 predicted occurrence).Patient data were collected using a risk factor questionnaire for inclusion in the study.Predictors included in the model were screened using univariate and binary logistic regression analysis,with the occurrence of prolonged postoperative hospital stay as the dependent variable and the included screening factors as the independent variables,to develop a predictive model for prolonged postoperative hospital stay in enhanced recovery after surgery for gynecologic cancer patients,which was presented as a column line graph.The ROC curve was used to assess the discriminatory ability of the model,and the sensitivity,specificity and correctness were used to assess the classification of the model.Results: The risk prediction model for prolonged postoperative hospital stay in gynecologic cancer patients incorporated intraoperative blood transfusion(OR=3.276),operative time(OR=1.011),avoidance of mechanical bowel preparation(OR=0.398),goaldirected fluid therapy(OR=0.297),early postoperative activity(OR=0.204),catheter removal within 24 h after surgery(OR=0.555),D-Dimer(OR=2.340),and the occurrence of postoperative complications(OR=2.126)for a total of 8 predictors.Model equation: z=-0.171+1.187×intraoperative blood transfusion+0.011×operative time-0.933×avoidance of mechanical bowel preparation-1.214×goal-directed fluid therapy-1.590×early postoperative activity-0.588×extraction of catheter within 24 h after surgery+0.001×DDimer+0.754×occurrence of postoperative complications.ROC The area under the curve was 0.8537 with a 95% CI of(0.8164-0.8911),a sensitivity of 87.2%,a specificity of70.4%,a Jorden index of 0.576,and an optimal cut-off point of 0.244.The predictive validity of this model for the occurrence of prolonged postoperative hospital stay in gynecologic cancer patients was moderate.The area under the ROC curve for external validation was 0.8303 with a 95% CI of(0.7674-0.8932),sensitivity was 70.00%,specificity was 78.79%,and the overall correctness of the model was(104+28)/172=76.74%.Conclusion: The risk prediction model constructed in this study for prolonged postoperative hospital stay in gynecologic cancer patients is effective in predicting the occurrence of prolonged postoperative hospital stay,and can provide a reference for nursing staff to take preventive interventions for postoperative patients with prolonged postoperative hospital stay in enhanced recovery after surgery for gynecologic cancer at an early stage.