Construction and validation of a prolonged postoperative hospital stay prediction model for oral cancer patients
Gao Ying
Shi Yanan
Shi Haoning
Wang Xiaoyan
Cai Zhen
Li Ling
Wang Fengling
Abstract：Objective To construct a prediction model for prolonged postoperative hospital stay in oral cancer patients and validate its predictive value,aiming to provide a reference for clinical use.Methods A total of 448 oral cancer patients hospitalized at Linyi Peo-ple s Hospital from August 1,2016,to November 31,2023,were selected as the study participants,who were randomly as-signed to a model-building group(314 cases)and a validation group(134 cases).Binary logistic regression analysis was con-ducted on the model-building group to identify the factors affecting prolonged postoperative hospital stay and establish a risk prediction model.The model was validated using the receiver operating characteristic curve(ROC),decision curve analysis(DCA),and calibration curves.Results The number of comorbidities(OR=1.378,95％CI:1.103-1.723),number of teeth extracted during surgery(OR=3.008,95％CI:1.986-4.556),neck lymph node dissection(OR=2.578,95％CI:1.356-4.900),tracheostomy(OR=2.857,95％CI:1.341-6.089),and perioperative blood transfusion(OR=3.072,95％CI:1.551-6.083)were identified as influence factors for prolonged postoperative hospital stay in oral cancer patients(P＜0.05).A nomogram prediction model was established and validated,with the model-building group's ROC area under the curve(AUC)being 0.854(95％CI:0.808-0.900),sensitivity of 0.859,and specificity of 0.677;the validation group's AUC was 0.874(95％CI:0.807-0.941),indicating good discrimination.The calibration curves for the model-building and validation groups had P values of 0.297 and 0.829,respectively,indicating good calibration.Moreover,DCA showed that the model had high clinical practical value.Conclusion A higher number of comorbidities,a greater number of teeth extracted during surgery,neck lymph node dissection,tracheostomy,and perioperative blood transfusion are risk vari-ables affecting the prolonged postoperative hospital stay in oral cancer patients.The risk prediction model constructed in this study can accurately identify the risk of prolonged postoperative hospital stay in these patients and has clinical application value.
Keywords：Oral cancerRisk factorsHospital stayPrediction model
Publication Date：2024-07-28
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：10( 69-78 )