急性脑梗死患者泌尿系统感染相关因素分析及预测模型建立
Analysis of the relevant factors and establishment of a predictive model for urinary system infection in pa-tients with acute cerebral infarction
刘丽 1常利 1李合华1
作者信息
- 1. 新乡医学院第一附属医院神经内科 河南新乡 453000
- 折叠
摘要
目的 探讨急性脑梗死患者泌尿系统感染的危险因素,建立感染预测模型,为制定防控措施提供参考.方法 回顾性分析2021年1月～2022年12月新乡医学院第一附属医院223例急性脑梗死患者的临床资料,其中泌尿系统感染者48例纳入感染组,未发生泌尿系统感染者175例纳入未感染组,采用Logistic回归分析筛选脑梗死患者发生泌尿系统感染的危险因素并构建预测模型,采用受试者工作曲线(ROC)评估模型判别效度和截点值.结果 Logisitic回归分析显示女性、年龄≥60岁、糖尿病史、泌尿系统结石、留置导尿管是脑梗死患者发生泌尿系统感染的独立危险因素;根据多因素分析结果构建脑梗死患者泌尿系统感染概率值回归方程为:P=1/[1+e-(-3.883+1.001*性别+0.880*年龄+1.136*糖尿病+1.018*泌尿系统结石+1.957*留置导尿管)].建模组对模型进行内部验证ROC曲线下面积(AUC)为0.823,95％CI(0.762,0.884),灵敏度为87.50％,特异度为64.00％,区分度良好,根据约登指数最大原则选取cut-off点为0.120.结论 建立的预测模型判别效度较佳,可用于识别脑梗死泌尿系统感染高危患者.
Abstract
Objective To explore the risk factors for urinary system infections in patients with a-cute cerebral infarction,establish an infection prediction model,and provide reference for formulating prevention and control measures.Methods A retrospective analysis was conducted on the clinical data of 223 patients with acute cerebral infarction in our hospital from January 2021 to December 2022.A-mong them,48 patients with urinary system infections were included in the infection group,and 175 patients without urinary system infections were included in the non infection group.Logistic regres-sion analysis was used to screen for risk factors for urinary system infections in patients with cerebral infarction and a predictive model was constructed.The discriminant validity and cutoff value of the model were evaluated using receiver operating curve(ROC).Results Logistic regression analysis showed that female,age≥60 years old,history of diabetes,urinary calculi and indwelling catheter were independent risk factors for urinary system infection in patients with cerebral infarction;Accord-ing to the results of multi factor analysis,the regression equation of the probability value of urinary system infection in patients with cerebral infarction was constructed as:P=1/[1+e-(-3.883+1.001*gender+0.880*age+1.136*diabetes+1.018*urinary calculus+1.957*indwelling catheter)].The modeling team conducted internal validation on the model,with ROC area under the curve(AUC)of 0.823,95％CI(0.762,0.884),sensitivity of 87.50％,specificity of 64.00％,and good discrimination.According to the principle of maximum Jordan index,a cut-off point of 0.120 was selected.Conclusion The estab-lished predictive model has good discriminant validity and can be used to identify high-risk patients with cerebral infarction and urinary system infections.
关键词
脑梗死/泌尿系统感染/危险因素/预测模型Key words
Cerebral infarction/Urinary system infection/Risk factors/Prediction model引用本文复制引用
出版年
2024
NETL
NSTL
维普
万方数据