摘要
目的:探究连续血糖监测系统血糖检测误差的因素,构建校准模型。方法:以实际患者为研究对象,使用辅理善瞬感扫描式血糖仪测量间质液血糖浓度和常规指尖血的末梢血糖浓度,并采用Lasso回归模型进行特征筛选,以线性模型为例,构造分类校准模型。结论:性别、饮酒史、年龄、家族史、检测时间、体重、吸烟史均会影响连续血糖监测系统的血糖检测误差,将其纳入校准模型,可提高血糖检测的准确性。
Objective:To explore the factors of blood glucose detection error in continuous blood glucose monitoring system and establish a calibration model.Methods:With actual patients as the rescarch object,the blood gloose concentration of interstitial fluid and the peripheral blood glucose concentration of conventional fingertip blood were measured by the auxiliary quick sensing scanning ghucose meter,and the feature screening was performed by Lasso regnesion model.The lincar model was used as un example to construct a cssification calibration model.Conclusion;Gender,drinking history,ngo,family history,detection time,weight and smoking history all affect the blood glu-cose detection errors of contirnucus blood glucose monitoring system,which can be included in the calibration model to improve the accuracy of blood glucose detection.
作者
陶国强
高楠
杨予潇
汪星星
寇祖亮
TAO Guoqiang;GAO Nan;YANG Yuxaio;WANG Xingxing;KOU Zuliang
出处
《计量与测试技术》
2024年第7期114-118,共5页
Metrology & Measurement Technique
基金
浦东新区科委项目(项目名称:围手术期血糖的智能监测与优化手术管理,项目编号:PKJ2021-Y89)。
