应用决策树和logistic回归模型分析胃肠道穿孔修补术后手术部位感染的影响因素  

Influencing factors for surgical site infection after gastrointestinal perforation repair surgery:analysis based on decision tree and logistic regression model

作　　者：胡登敏 胥润[1] 李建[1] 艾余 HU Deng-min;XU Run;LI Jian;AI Yu(Department of General Surgery,The Third Hospital of Mianyang,Mianyang 621000,China;Department of Nephrology,The Third Hospital of Mianyang,Mianyang 621000,China)

机构地区：[1]绵阳市第三人民医院(四川省精神卫生中心)普通外科,四川绵阳621000 [2]绵阳市第三人民医院(四川省精神卫生中心)肾病内科,四川绵阳621000

出　　处：《中国感染控制杂志》2024年第7期826-832,共7页Chinese Journal of Infection Control

摘　　要：目的应用决策树和logistic回归模型分析胃肠道穿孔修补术后并发手术部位感染(SSI)的影响因素。方法选取2018年1月—2023年1月于绵阳市某医院进行胃肠道穿孔修补术的患者为研究对象,收集患者临床资料,根据术后是否发生SSI将患者分为SSI(+)组(n=41)和SSI(-)组(n=322)。使用单因素和多因素logistic回归分析胃肠道穿孔修补术后并发SSI的影响因素,并建立相关决策树预测模型。结果363例胃肠道穿孔修补术患者术后发生SSI 41例,发病率为11.29%。单因素分析结果显示,两组患者身体质量指数(BMI)、清蛋白水平、术前应用抗菌药物、术前腹痛时间和手术时间比较,差异均有统计学意义(均P<0.05)。多因素logistic回归分析显示,BMI较高(OR=2.059,95%CI:1.103~3.842)、清蛋白水平<35 g/L(OR=2.761,95%CI:1.312~5.811)、术前腹痛时间≥24 h(OR=3.589,95%CI:1.659~7.763)和手术时间≥2 h(OR=3.314,95%CI:1.477~7.435)是胃肠道穿孔修补术后患者并发SSI的独立危险因素(P<0.05),术前应用抗菌药物是其保护因素(OR=0.338,95%CI:0.166~0.690,P<0.05)。基于上述因素建立预测胃肠道穿孔修补术后患者并发SSI风险的决策树模型。模型验证结果显示,受试者工作特征曲线下面积(AUC)为0.811(95%CI:0.794~0.825)。结论胃肠道穿孔修补术后患者并发SSI的危险因素包括BMI较高、清蛋白水平<35 g/L、术前腹痛时间≥24 h和手术时间≥2 h,保护因素为术前应用抗菌药物,基于影响因素构建的决策树模型对胃肠道穿孔修补术后患者并发SSI风险具有良好的预测能力。Objective To analyze the influencing factors for surgical site infection(SSI)after gastrointestinal perforation repair surgery by decision tree and logistic regression model.Methods Patients who underwent gastrointestinal perforation repair surgery at a hospital of Mianyang City from January 2018 to January 2023 were selected as the research subjects.Clinical data of the patients were collected.Patients were divided into the SSI(+)group(n=41)and the SSI(-)group(n=322)based on whether SSI occurred after surgery.Influencing factors for SSI after gastrointestinal perforation repair surgery were analyzed by univariate and multivariate logistic regression.Re-levant decision tree prediction model was constructed.Results Among the 363 patients who underwent gastrointestinal perforation repair surgery,41 developed postoperative SSI,with an incidence of 11.29%.Univariate analysis results showed that there were statistically significant differences between two groups of patients in body mass index(BMI),albumin level,preoperative antimicrobial use,duration of preoperative abdominal pain,and duration of surgery(all P<0.05).Multivariate logistic regression analysis showed that higher BMI(OR=2.059,95%CI:1.103-3.842),albumin levels<35 g/L(OR=2.761,95%CI:1.312-5.811),duration of preoperative abdominal pain≥24 hours(OR=3.589,95%CI:1.659-7.763),and duration of surgery≥2 hours(OR=3.314,95%CI:1.477-7.435)were independent risk factors for postoperative SSI in patients after gastrointestinal perforation repair surgery(P<0.05),while preoperative antimicrobial use was a protective factor(OR=0.338,95%CI:0.166-0.690,P<0.05).The decision tree model based on the above factors was constructed to predict the risk of SSI in patients after gastrointestinal perforation repair surgery.Validation of the model showed that the area under the receiver operating characteristic(ROC)curve(AUC)was 0.811(95%CI:0.794-0.825).Conclusion The risk factors for postoperative SSI in patients after gastrointestinal perforation repair surgery include high BMI,a

关 键 词：胃肠道穿孔修补术 手术部位感染 危险因素 LOGISTIC回归 决策树 

分 类 号：R181.32[医药卫生—流行病学]