论文部分内容阅读
目的分析手足口病重症病例的危险因素,为降低手足口病重症病例发生率提供理论依据。方法福建省2008-2013年手足口病重症确诊病例资料来源于疾病监测信息报告管理系统,从同期报告的轻症病例确诊病例中随机选取与重症病例相同数目的病例。运用决策树中的CHAID法分析重症病例发生的危险因素,因变量为重症是否发生,自变量为实验室结果、发病到就诊时间、发病月份、年龄、性别、城乡。使用分割样本验证模型,样本被随机均分为训练样本与测试样本。结果福建省2008-2013年共报告13971例手足口确诊病例,其中重症1503例(10.8%),轻症12468例,共3006例病例进入模型。训练样本(1523例)模型中包括实验室结果、发病到就诊间隔时间、年龄、性别4个预测变量,实验室结果是最佳预测变量,模型对70.1%的训练样本进行了正确分类,测试样本(1483例)模型对66.5%的个案进行了正确分类。结论决策树是一种有效预测手足口病重症病例的方法。
Objective To analyze the risk factors of HFMD cases and provide a theoretical basis for reducing the incidence of HFMD cases. Methods The data of severe confirmed cases of hand, foot and mouth disease in Fujian Province from 2008 to 2013 were collected from the Disease Surveillance Information Report Management System. The same number of cases as the severe cases were randomly selected from the cases of mild cases confirmed in the same period. The CHAID method in decision tree was used to analyze the risk factors of severe cases. Whether the dependent variables were severe cases or not, independent variables were laboratory results, onset to treatment time, onset month, age, sex, urban and rural areas. Using the split sample validation model, samples were randomly divided into training samples and test samples. Results A total of 13971 cases of hand, foot and mouth were reported in Fujian province from 2008 to 2013, of which 1503 were severe cases (10.8%) and 12468 were mild cases. A total of 3006 cases were included in the model. The training sample (1523 cases) included the laboratory results, the interval from onset to treatment, age and gender, and the laboratory results were the best predictors. The model correctly classified 70.1% of the training samples, and the test samples (1483 cases) model correctly classified 66.5% of the cases. Conclusion Decision tree is an effective method to predict HFMD cases.