Exhaustive CHAID分类树与logistic回归在脑卒中危险因素中的应用

Exhaustive CHAID decision tree and logistic regression analyses of risk factors for stroke

目的应用exhaustive CHAID分类树模型与logistic回归分析来分析北京社区居民脑卒中危险因素以及不同特征人群的重点干预因素,为加强北京市居民脑卒中的干预提供科学依据。方法于2007年6月至8月,采用整群抽样方法,对北京10 108名社区居民进行问卷调查、体格检查及检测空腹血糖、血脂。采用logistic回归与exhaustive CHAID分类树分析相结合来探讨影响北京市居民脑卒中的因素。结果 logistic回归分析和exhaustive CHAID分类树分析显示年龄、性别、踝臂指数(ABI)、高血压、腹型肥胖、高密度脂蛋白胆固醇、吸烟状况、工作强度为脑卒中的危险因素;Exhaustive CHAID分类树分析揭示老年者ABI贡献大,不容忽视中年者糖尿病。Logistic回归分析和exhaustive CHAID分类树分析的ROC曲线下面积分别为0.803和0.778,模型可靠。结论对脑卒中的防治,要在总体把握的情况下,对不同的高危人群应采取不同的防制措施。

Objective To analyze risk factors for stroke by logistic regression model and exhaustive CHAID decision tree methods.Methods A total of 10 108 community residents were recruited for study by a cluster sampling method.Data were collected by questionnaire,physical examination and blood testing for fasting plasma lipids and glucose concentrations.The risk factors were determined by logistic regression model and exhaustive CHAID decision tree methods.Results Stroke was significantly associated with age,sex,abdominal obesity,HDL,hypertension,smoking,ankle-braehial index（ABI）,intensity of work,and so on by logistic regression and exhaustive CHAID decision tree methods（P〈0.05）.The exhaustive CHAID decision tree analysis suggested that ABI was important for predicting the risk of stroke in elderly people,so as diabetes in middle-aged population.The curved areas were 0.803 and 0.778 respectively for the logistic regression model and the exhaustive CHAID analysis.Conclusion Different measures should be taken to prevent residents from the occurrences of stroke based on their specific risk factors.