儿科多烯磷脂酰胆碱输液渗漏Logistic回归预测模型的建立及其应用
发布时间：2018-10-26 09:12
【摘要】：目的筛选儿科多烯磷脂酰胆碱输液渗漏有统计学意义的危险因素,初步建立logistic回归模型,评价该模型的预测价值。方法选择儿科脉输注多烯磷脂酰胆碱患儿153例,其中103例作为模型训练组用于模型的创建,另外50例作为模型验证组用于模型的评判。采用单因素和多因素非条件Logistic回归分析模型训练组中患儿发生输液渗漏的危险因素。通过优化组合方式建立Logistic回归模型,对其稳定性进行验证,并将得到的预测模型代入验证数据集进行评价。结果年龄、同一血管进针次数、护士工作年限及用药时间4个危险因素在经过Logistic回归分析之后进入模型;利用该回归模型对模型验证组中50例多烯磷脂酰胆碱输注病人进行预报,其曲线下面积、灵敏性、特异性分别为0.983,92.3%,91.6%。结论 Logistic回归分析能够筛选出对儿科多烯磷脂酰胆碱输液渗漏有意义的危险因素;该Logistic回归模型对儿科多烯磷脂酰胆碱输液渗漏风险有初步预判的作用。
[Abstract]:Objective to screen the risk factors of polyene phosphatidylcholine infusion leakage in pediatrics and to establish a logistic regression model and evaluate the predictive value of the model. Methods 153 children with Paediatric vein infusion of polyenylphosphatidylcholine were selected, of which 103 were used as model training group and 50 as model validation group. Univariate and multivariate non-conditional Logistic regression analysis was used to analyze the risk factors of transfusion leakage in the training group. The Logistic regression model is established by optimizing the combination method, and its stability is verified, and the prediction model is added to the validation data set to evaluate the stability of the model. Results four risk factors, age, times of injection of the same blood vessel, working life of nurses and duration of medication, entered the model after Logistic regression analysis. The regression model was used to predict 50 patients with polyene phosphatidylcholine infusion in the model verification group. The area, sensitivity and specificity under the curve were 0.983C 92.3% and 91.6%, respectively. Conclusion Logistic regression analysis can screen out the risk factors of percolation of pediatric polyenylphosphatidylcholine infusion, and the Logistic regression model has a preliminary predictive effect on the risk of percolation of pediatric polyene phosphatidylcholine infusion.
【作者单位】： 四川大学华西第二医院儿科;出生缺陷与相关妇儿疾病教育部重点实验室;
【基金】：国家自然科学基金(81501301) 国家临床重点专科建议项目(1311200003303)资助
【分类号】：R473.72
，
本文编号：2295250
[Abstract]:Objective to screen the risk factors of polyene phosphatidylcholine infusion leakage in pediatrics and to establish a logistic regression model and evaluate the predictive value of the model. Methods 153 children with Paediatric vein infusion of polyenylphosphatidylcholine were selected, of which 103 were used as model training group and 50 as model validation group. Univariate and multivariate non-conditional Logistic regression analysis was used to analyze the risk factors of transfusion leakage in the training group. The Logistic regression model is established by optimizing the combination method, and its stability is verified, and the prediction model is added to the validation data set to evaluate the stability of the model. Results four risk factors, age, times of injection of the same blood vessel, working life of nurses and duration of medication, entered the model after Logistic regression analysis. The regression model was used to predict 50 patients with polyene phosphatidylcholine infusion in the model verification group. The area, sensitivity and specificity under the curve were 0.983C 92.3% and 91.6%, respectively. Conclusion Logistic regression analysis can screen out the risk factors of percolation of pediatric polyenylphosphatidylcholine infusion, and the Logistic regression model has a preliminary predictive effect on the risk of percolation of pediatric polyene phosphatidylcholine infusion.
【作者单位】： 四川大学华西第二医院儿科;出生缺陷与相关妇儿疾病教育部重点实验室;
【基金】：国家自然科学基金(81501301) 国家临床重点专科建议项目(1311200003303)资助
【分类号】：R473.72
，
本文编号：2295250
本文链接：https://www.wllwen.com/linchuangyixuelunwen/2295250.html
最近更新
教材专著
