脑出血术后病人肺部感染危险因素分析及风险预测模型构建  

Risk factor analysis and establishment of predictive model for pulmonary infection in postoperative patients with intracerebral hemorrhage

作　　者：李景余 臧丽丽 吕慧静 孟庆慧[1] LI Jingyu;ZANG Lili;LYU Huijing;MENG Qinghui(Nursing College of Weifang Medical University,Shandong 261053 China)

机构地区：[1]潍坊医学院护理学院,山东261053

出　　处：《护理研究》2022年第17期3055-3060,共6页Chinese Nursing Research

基　　金：山东省自然科学基金面上项目,编号:ZR2020MH128;潍坊医学院教育教学改革研究项目,编号:2019YB010;潍坊医学院博士启动基金项目,编号:2017BSQD58。

摘　　要：目的:分析脑出血术后病人肺部感染的危险因素并构建风险预测模型,为临床早期筛选高危人群和干预提供参考。方法:选择2019年1月—2020年12月潍坊市某三级甲等医院的298例脑出血术后病人作为研究对象,采用自行编制的脑出血术后病人肺部感染危险因素调查表进行调查,按7∶3的比例随机分为模型组(n=209)和验证组(n=89)。对模型组病人人口统计学特征、既往史及相关并发症发生情况、入院时实验室资料、围术期资料进行统计分析,依据术后是否出现肺部感染分为感染组与非感染组,应用单因素和多因素Logistic回归分析构建风险预测模型,采用Hosmer-Lemeshow检验及受试者工作特征曲线(ROC)评估模型拟合优度和预测效能。结果:脑出血术后肺部感染发生率为17.22%,Logistc回归分析显示,术前格拉斯哥昏迷评分(GCS)(OR=2.394)、美国麻醉医师协会(ASA)分级(OR=1.962)、留置胃管(OR=2.321)、白细胞计数(OR=1.709)、机械通气时间(OR=4.345)为脑出血术后病人发生肺部感染的独立危险因素,相应风险预测模型为:Logit(P)=-3.591+0.873×GCS评分+0.674×ASA分级+0.842×留置胃管+0.536×白细胞计数+1.469×机械通气时间。模型组H-L检验,χ^(2)=6.889,P=0.549,ROC曲线下面积(AUC)为0.748;验证组H-L检验,χ^(2)=6.551,P=0.553,AUC为0.718,准确率为72.0%。结论:构建的模型预测效果良好,可为医护人员筛查脑出血术后发生肺部感染高危人群提供参考。Objective:To analyze the risk factors of pulmonary infection in postoperative patients with intracerebral hemorrhage and build a risk prediction model to provide references for early clinical screening and intervention of high-risk groups.Methods:Totally 298 postoperative patients with intracerebral hemorrhage in a tertiary grade A hospital in Weifang from January 2019 to December 2020 were selected as the research objects.The self-made questionnaire on risk factors of pulmonary infection in postoperative patients with intracerebral hemorrhage was used for investigation.Patients were randomly divided into the modeling group with 209 cases and the validation group with 89 cases in a 7∶3 ratio.The demographic characteristics,past history and related complications,laboratory data at admission and perioperative data of patients in the modeling group were statistically analyzed,they were divided into infection group and non-infection group according to the occurrence of pulmonary infection after operation.A risk model was established by univariate and logistic regression multivariate analysis.The goodness of the fit and predictive validity of the model were verified by the Hosmer-Lemeshow test and the ROC curve,respectively.Results:The incidence of pulmonary infection after intracerebral hemorrhage was 17.22%.Multivariate logistic regression analysis showed that the GCS score(OR=2.394),ASA level(OR=1.962),indwelling gastric tube(OR=2.321),WBC count(OR=1.709),mechanical ventilation time(OR=4.345)were independent risk factors for pulmonary infection in postoperative patients with intracerebralhemorrhage.The prediction model was:Logit(P)=-3.591+0.873×GCS+0.674×ASA+0.842×indwelling gastric tube+0.536×WBC count+1.469×mechanical ventilation time.H-L test in modeling group showed that,χ^(2)=6.889,P=0.549,AUC under ROC curve was 0.748.H-L test in validation group showed that,χ^(2)=6.551,P=0.553,AUC under ROC was 0.718,and the accuracy was 72.0%.Conclusions:The constructed model had good predictive effect and it could

关 键 词：脑出血 肺部感染 危险因素 预测模型 调查研究 

分 类 号：R473.6[医药卫生—护理学]