决策树联合Logistic回归模型探讨骨科清洁手术术后切口感染病原菌特点及危险因素  

Analysis of pathogen distribution characteristics and risk factors of incision infection after orthopedic clean surgery with decision tree model joined Logistic regression model

作　　者：杨蓓蓓 张晓钰[1] 姜凤丽[1] 张晋[1] 乌伊萍 李兴欢 刘冬[1] Yang Beibei;Zhang Xiaoyu;Jiang Fengli;Zhang Jin;Wu Yiping;Li Xinghuan;Liu Dong(Department of Clinical Pharmacy,Baoji Central Hospital,Baoji 721006,China)

机构地区：[1]宝鸡市中心医院临床药学室,陕西宝鸡721006

出　　处：《实用药物与临床》2023年第4期323-328,共6页Practical Pharmacy and Clinical Remedies

基　　金：陕西省科学技术厅科技计划项目(2022SF-569)。

摘　　要：目的探讨我院骨科清洁手术切口感染的病原菌特点及危险因素。方法收集2018年1月-2020年12月在我院骨科行清洁手术的2264例患者资料,分析切口感染的病原菌特点,并分别建立决策树CHAID模型和Logistic回归模型,分析切口感染的影响因素,更全面地探讨导致切口感染的危险因素。结果手术后有31例(1.37%)患者发生切口感染,感染患者切口分泌物中共培养出致病菌15株,其中革兰阳性球菌10株,革兰阴性杆菌5株。决策树CHAID模型分析结果显示,手术时间>2 h、术前抗菌药物使用时间≥10 d以及既往有手术经历均是切口感染的危险因素,并且术前抗菌药物使用时间≥10 d的患者术后感染的风险最高;Logistic回归结果显示,合并糖尿病、手术时间>2 h、术前抗菌药物使用时间≥10 d以及既往有手术经历是感染发生的独立危险因素。结论通过分析切口感染的病原菌,并且利用决策树CHAID模型与Logistic回归相结合,从不同方面描述影响术后切口感染发生的因素及其作用,为制定骨科手术切口感染的预防和治疗策略提供充分的实践依据,降低了院内的感染率。Objective To analyze the risk factors and the distribution characteristics of pathogenic bacteria of incision infection after orthopedic clean surgery in our hospital.Methods The clinical data of 2264 patients undergoing clean surgery from January 2018 to December 2020 were collected,the characteristics of pathogens of incision infection were analyzed,CHAID decision tree model and Logistic regression model were respectively established to analyze the influencing factors of infection,and the risk factors of incision infection were comprehensively analyzed.Results There were 31 cases(1.37%)with incision infection after operation,and 15 strains of pathogenic bacteria were isolated and cultured from the incision secretion,of which 10 strains were Gram-positive bacteria,and 5 strains were Gram-negative bacteria.CHAID decision tree model analysis showed that the potential dangerous factors of infection included operation time>2 h,preoperative antibiotic use time≥10 d and the previous surgical records,and the highest risk factor was preoperative antibiotic use time≥10 d.Logistic regression model showed that combination of diabetes,surgery time>2 h,preoperative antibiotic use time≥10 d and the previous surgical records were the independent risk factors for infection.Conclusion By analyzing the pathogenic bacteria of incision infection,CHAID decision tree model and Logistic regression model are combined to describe the influencing factors of surgical site infection and their roles from different aspects.The effective combination of the two models can help to provide evidence and reference for the prevention and treatment of surgical site infection in orthopedic surgery,and thus reducing the nosocomial infection rate.

关 键 词：病原菌 决策树CHAID模型 LOGISTIC回归模型 危险因素 切口感染 

分 类 号：R619.3[医药卫生—外科学] R446.5[医药卫生—临床医学]