危重患者医院获得性肺炎的危险因素分析及预测模型的建立

发布时间：2018-03-03 00:26

  本文选题：医院获得性肺炎　切入点：logistic回归分析　出处：《中华医院感染学杂志》2015年11期 　论文类型：期刊论文

【摘要】：目的通过单因素和多因素分析筛选出与医院获得性肺炎(hospital-acquired pneumonia,HAP)发生相关的危险因素,构建风险预测模型。方法回顾性收集2009年1月-2013年12月医院ICU 539例住院患者临床资料(A组),采用logistic回归分析筛选危险因素、建立风险预测模型,另外前瞻性收集2014年1-6月ICU 88例住院患者资料(B组),以验证该模型的临床预测价值。结果 A组539例患者中发生HAP215例,单因素分析结果显示,患者年龄、APACHEⅡ评分、患心血管系统疾病及多器官功能障碍综合征(MODS)、心肺复苏术、感染性休克、恶性肿瘤、合并肺外感染、低蛋白血症、C-反应蛋白(CRP)值及降钙素原(PCT)值高、机械通气、无创通气、输血、预防性使用抑酸药物等在HAP发生与否的差异有统计学意义;多因素logistic回归分析结果显示,APACHEⅡ评分、MODS、低蛋白血症、PCT值高、无创通气、输血是HAP发生的独立危险因素;利用B组数据验证该公式预测HAP灵敏性为85.2%、特异性为75.6%,阳性预测值为80.38%,阴性预测值为83.75%。结论多因素logistic回归模型能够较好地预测危重患者HAP发生的风险。
[Abstract]:Objective to screen the risk factors associated with hospital acquired pneumonia (HAP) by univariate and multivariate analysis. Methods the clinical data of 539 inpatients with ICU from January 2009 to December 2013 were collected retrospectively. The risk factors were screened by logistic regression analysis, and the risk prediction model was established. In addition, 88 inpatients with ICU from 2014 to June were prospectively collected to verify the clinical predictive value of the model. Results in group A, HAP215 occurred in 539 patients. Univariate analysis showed that the age of the patients was evaluated with Apache 鈪，

本文编号：1558764