摘要
目的:建立脑卒中发病概率预测模型并对模型预测效果进行评价。方法:回顾性分析152例脑卒中病例资料,采用Logistic回归分析研究影响脑卒中发病的危险因素,建立发病概率预测模型并利用ROC曲线进行评价。结果:影响脑卒中发病的危险因素包括年龄、高血压、糖尿病、高密度脂蛋白、低密度脂蛋白等,ROC曲线下面积为0.884,模型具有较好的预测价值。结论:Lo-gistic回归预测概率模型评价效果较好,能较为准确地预测脑卒中发病概率。
Objective. To establish and evaluate a predictive model of incidence probability of stroke. Methods. The clinical data of 152 cases of stroke were retrospectivly analyzed. The risk factors of stroke were studied by logistic Regression method and the efficiency of this model was evaluated with ROC method. Results. The main risk factors of stroke patients include age, hypertension, diabetes mellitus, HDL and LDL. The area of ROC curve was 0. 884, suggesting that the predictive value was satisfactory. Conclusion.. The useful predictive model by Logistic Regression which can be used to predict the incidence of stroke have been established.
出处
《神经损伤与功能重建》
2011年第3期196-198,共3页
Neural Injury and Functional Reconstruction
关键词
脑卒中
危险因素
脑血管疾病
预测模型
stroke
risk factors
cerebrovascular disease
predictive model
