基于列线图的空腹血糖受损发生风险预测模型的构建与验证  

Construction and validation of a risk prediction model for impaired fasting glucose on column charts

作　　者：甄紫伊 刘蕾 孟继娴 傅一婷 马小慧 孙金菊 ZHEN Ziyi;LIU Lei;MENG Jixian;FU Yiting;MA Xiaohui;SUN Jinju(Epidemiology and Health Statistics,School of Public Health,Shenyang Medical College,Shenyang 110034,China;Department of Community Nursing,School of Nursing,Liaoning University of Traditional Chinese Medicine,Shenyang 116600,China)

机构地区：[1]沈阳医学院公共卫生学院流行病与卫生统计学教研室,沈阳110034 [2]辽宁中医药大学护理学院社区护理学教研室,沈阳116600

出　　处：《中国医科大学学报》2025年第1期18-23,共6页Journal of China Medical University

基　　金：辽宁省教育厅科学研究经费项目(LJKR0555);沈阳医学院硕士研究生科技创新基金(Y20220515)。

摘　　要：目的探讨空腹血糖受损(IFG)发生的危险因素,构建基于列线图的IFG发生风险的预测模型并验证。方法收集2022年8月至12月沈阳市某医院体检门诊常规体检者(n=3037)的临床资料,将研究对象按照7∶3随机分为训练组(n=2126)和验证组(n=911)。通过LASSO回归分析筛选预测变量,采用logistic回归分析进一步筛选并构建列线图预测模型,由验证组对模型可行性进行内部验证,应用受试者操作特征(ROC)曲线的曲线下面积(AUC)和Hosmer-Lemeshow拟合优度检验对模型效能进行评价。结果3037例研究对象中2880例未发生IFG,157例发生IFG。结果显示,年龄(OR=1.04,95%CI:1.02~1.05)、体重指数(OR=1.10,95%CI:1.05~1.17)、收缩压(OR=1.01,95%CI:1.00~1.03)、甘油三酯(OR=1.20,95%CI:0.99~1.51)以及高血压(OR=1.28,95%CI:1.04~1.59)是IFG发生的独立危险因素。基于以上变量成功构建了列线图预测模型。预测模型中训练组预测IFG发生的AUC为0.722(95%CI:0.68~0.77),验证组预测IFG发生的AUC为0.907(95%CI:0.87~0.94)。拟合优度检验结果显示训练组与验证组模型差异比较无统计学意义(P>0.05),实际概率与模型预测概率基本一致,模型校准度较好。结论本研究构建了包括年龄、体重指数、收缩压、甘油三酯以及高血压5个变量的IFG发生风险的预测模型,有助于尽早筛选IFG高危人群并及时干预。Objective To discuss the risk factors for impaired fasting glucose(IFG)and construct and validate a predictive model based on column charts of the risk of IFG occurrence.Methods This retrospective study included 3037 individuals who underwent routine physical examinations at a hospital in Shenyang between August and December 2022.The population was randomly divided into a training group(n=2126)and a validation group(n=911)in a 7∶3 ratio,and physical examination data were collected.LASSO regression analy-sis was used to screen predictive variables and logistic regression analysis was used to further screen and construct a column chart predictive model.The validation group was used to conduct an internal validation of the feasibility of the model,and the area under the curve(AUC)of receiver operator characteristic(ROC)and goodness of fit tests were used to evaluate the model effectiveness.Results Among the 3037 included individuals,2880 did not experience IFG and 157 did.The results showed that age(OR=1.04,95%CI:1.02-1.05),body mass index(BMI,OR=1.10,95%CI:1.05-1.17),systolic blood pressure(SBP,OR=1.01,95%CI:1.00-1.03),triglycerides(TG,OR=1.20,95%CI:0.99-1.51),and a history of hypertension(OR=1.28,95%CI:1.04-1.59)were independent risk factors for IFG occurrence in this population.Based on these variables,a column chart prediction model was constructed.In the training group,the model predicted an AUC of 0.722(95%CI:0.68-0.77)for IFG occurrence,while in the validation group,it predicted an AUC of 0.907(95%CI:0.87-0.94)for IFG occurrence.The results of the Hosmer-Lemeshow goodness of fit test showed that the models of the training and validation groups were not significantly different(P>0.05);that is,the actual probability was consistent with the prediction probability of the model,and the models calibration was good.Conclusion A risk prediction model for IFG occurrence that included five variables:age,BMI,SBP,TG,and history of hypertension could be construted.This model might help to identify high-risk groups for I

关 键 词：空腹血糖受损 危险因素 预测模型 列线图 

分 类 号：R587.1[医药卫生—内分泌]