杭州市上城区成年人高血压影响因素的分类树研究

A taxonomic tree study on the influencing factors of hypertension among adults in Shangcheng district of Hangzhou city

目的应用分类树模型分析杭州市上城区成年人高血压相关影响因素,为社区高血压精准防控提供依据。方法对2 762名≥18岁上城区社区居民进行问卷调查、体格检查和实验室检测,应用卡方自动交互检测(CHAID)法建立高血压分类树模型筛选相关因素,采取模型错分概率Risk统计量及受试者工作特征曲线(ROC)下面积对模型进行评估。结果 2 762名≥18岁社区居民中,高血压患者956例,高血压患病率为34.61%,标化患病率为15.83%。分类树模型共4层37个节点,从全部19个变量中筛选出年龄、BMI、中心型肥胖、血脂异常、高血压家族史、糖尿病、吸烟、被动吸烟8个解释变量,其中年龄是最重要的因素,高血压患病率随着年龄的增长而升高;模型显示,在不同年龄层下,高血压影响因素不尽相同。模型错分概率Risk统计量为0.224,利用预测概率绘制的ROC曲线下面积为0.713(95%CI:0.694~0.731),模型拟合效果较好。结论杭州市上城区居民高血压患病率较高,分类树模型不仅可以有效地挖掘高血压相关因素,且能定义不同的亚人群,对精准防控有较高的应用价值。

Objective To analyze the related influential factors of hypertension among residents of Shangcheng district in Hangzhou city by taxonomic tree model;to provide basis for making targeted measurements for community hypertension patients.Methods A total of 2 762 residents aged ≥18 years were selected with multi-stage random sampling method from Shangcheng district,who were subjected to a questionnaire interview,physical examination and laboratory test.The taxonomic tree model was constructed using Chi square automatic interactive detection(CHAID)method,which was evaluated by Riskstatistics and the area under the ROC curve.Results Among the 2 762 participants≥18 years old,956 were hypertension patients,resulting the prevalence rate of 34.61%and age standardized prevalence rate of 15.83%.The taxonomic model was constructed in 4 layers,including 37 nodes,8 explanatory variables including age,BMI,central obesity,dyslipidemia,family history of hypertension,diabetes,smoking and passive smoking,were selected from all 19 variables,among which age was the most important factor,the prevalence rate of hypertension increased with age.The model showed that the influential factors of hypertension were different in different age groups.The Riskstatistic of misclassification probability of the model was 0.224,and the area under the ROC curve drawn by prediction probability was 0.713(95% CI:0.694-0.731),suggesting that the taxonomic tree model fitted well.Conclusion The hypertension prevalence rate is relatively high among community residents of Shangcheng district.The taxonomic tree model could not only identify the major influential factors effectively,but also define the different subgroups,which has high application value in the context of precise prevention and control.