摘要
目的建立高血压患者颈动脉硬化发生风险的预测模型并评价。方法回顾性分析我院2016年6月至2018年12月的高血压患者,收集患者的临床资料、主要生化指标、颈动脉超声颈动脉内膜-中层厚度值等,应用SPSS 25.0和R 3.6.1软件进行统计分析,建立预测模型并评价。结果单因素和多因素分析表明,年龄≥60岁、低密度脂蛋白升高、碱性磷酸酶升高为高血压患者颈动脉硬化的独立危险因素。根据多因素logistic回归模型筛选的变量建立风险列线图,C-index为0.770,模型区分能力尚可。ROC曲线计算AUC为0.770,模型准确性尚可。Calibration曲线、Decision曲线评价模型一致性及获益性尚可。结论高血压患者颈动脉硬化发生风险预测模型列线图准确性尚可,对医务人员在临床工作直观的、个体化的风险分析具有一定的参考价值,但由于本研究样本量较少、病例来源局限,仍需要通过大样本、多中心的研究进行验证。
Objective To establish a predictive model for the risk of carotid atherosclerosis in patients with hypertension and evaluate it.Methods Patients with hypertension in our hospital from June 2016 to December 2018 were analyzed retrospectively.The clinical data,main biochemical indexes and intima-medium thickness value of carotid artery were collected.SPSS 25.0 and R 3.6.1 software were used for statistical analysis,and the prediction model was established and evaluated.Results Univariate and multivariate analysis showed that age≥60 years,elevated low density lipoprotein cholesterol and alkaline phosphatase were independent risk factors for carotid atherosclerosis in patients with hypertension.According to the variables selected by Logistic regression model,the risk nomograph was established.The C-index is 0.770,and the distinguishing ability of the model is good.The AUC calculated by ROC curve is 0.770,and the accuracy of the model is good.The evaluation of Calibration curve and Decision curve shows that the consistency and benefit of the model are still available.Conclusion The accuracy of the nomograph for the risk of carotid atherosclerosis in patients with hypertension is reasonable,which has a certain reference value in the clinical work and individual risk analysis.However,due to the small sample size and the limited source of cases,it still needs to be verified by large sample and multicenter research.
作者
沈群弟
Shen Qundi(Department of Clinical Laboratory,Shaoxing Seventh People′s Hospital,Shaoxing 312000,China)
出处
《中国医院统计》
2020年第3期211-214,共4页
Chinese Journal of Hospital Statistics
关键词
高血压
颈动脉硬化
风险预测模型
列线图
hypertension
carotid atherosclerosis
risk prediction model
nomograph
