老年骨折术后手术部位感染Logistic预测模型建立及其价值

目的运用Logistic多元回归分析建立老年骨折术后手术部位感染(SSI)数学预测模型并评估其预测感染的效能。方法选取2015年1月-2019年12月江南大学附属医院就诊的1034例老年骨折手术患者为研究对象,根据术后感染情况将患者分为感染组和对照组,利用Logistic多元回归分析术后SSI的危险因素并建立预测模型,通过受试者工作特征曲线(ROC)评价预测模型的效能。结果1034例患者中有60例患者术后发生SSI,发生率为5.80%;多元Logistic回归分析结果显示:年龄≥65岁、手术时间≥3 h、II/III类手术切口、受伤至手术时间≥8 h、引流管留置时间≥5 d、住院时间≥7 d是术后SSI的独立危险因素(P<0.05);老年骨折患者术后SSI的预测模型为:Logit(P)=2.438+0.641×(年龄)+0.581×(手术时间)+0.504×(II/III类手术切口)+0.482×(受伤至手术时间)+0.653×(引流管留置时间)+0.543×(住院时间),Hosmer-lemeshow拟合优度检验,模型的预测概率和实际发生率比较无统计学差异;验证试验ROC曲线下面积为0.840,95%CI为0.760~0.919,随机抽取100例进行Logistic回归预测模型验证其总准确率为76.00%。结论年龄≥65岁、手术时间≥3 h、II/III类手术切口、受伤至手术时间、引流管留置时间≥5 d、住院时间≥7 d是老年骨折患者术后SSI的危险因素,Logistic回归分析构建的预测模型对术后SSI的预测效能较好。

Establishment of Logistic prediction model for postoperative surgical site infection in elderly fracture patients and its value

OBJECTIVE To establish the mathematical prediction model for postoperative surgical site infection(SSI)in the elderly fracture patients by multivariate logistic regression analysis and analyze its efficiency in prediction of infection.METHODS A total of 1034 elderly fracture patients who were treated in Affiliated Hospital of Jiangnan University from Jan 2015 to Dec 2019 were enrolled in the study and divided into the infection group and the control group according to the status of postoperative infection.Multivariate logistic regression analysis was performed for risk factors for the postoperative SSI,the prediction model was established,and the efficiency of the prediction model was evaluated by means of receiver operating characteristic(ROC)curve.RESULTS Of the 1034 patients,60 had postoperative SSI,with the incidence rate 5.80%.Multivariate logistic regression analysis showed that the no less than 65 years of age,operation duration no less than 3 hours,type II/III surgical incision,time interval between injury and surgery no less than 8 hours,drainage tube indwelling time no less than 5 days and length of hospital stay no less than 7 days were the independent risk factors for the postoperative SSI(P<0.05).The prediction model for the postoperative SSI in the elderly fracture patients was Logit(P)=2.438+0.641×(age)+0.581×(operation duration)+0.504×(type II/III surgical incision)+0.482×(time interval between injury and surgery)+0.653×(drainage tube indwelling time)+0.543×(length of hospital stay),Hosmer-lemeshow goodness of fit test showed that there was no significant difference between the prediction probability and the actual incidence rate.Verification test indicated that the area under ROC curve was 0.840,95%CI was 0.760~0.919,totally 100 cases were randomly selected for the verification of the Logistic regression prediction model,and the total accuracy rate was 76.00%.CONCLUSION The no less than 65 years of age,operation duration no less than 3 hours,type II/III surgical incision,time interval between injury and surgery,drainage tube indwelling time no less than 5 days and length of hospital stay no less than 7 days are the risk factors for the postoperative SSI in the elderly fracture patients.The efficiency of the prediction model established based on Logistic regression analysis is high in prediction of the postoperative SSI.