摘要
目的:确立科学的分类节点变量,为建立病例分组做准备。方法:在运用正态性检验、核密度分布图以及Box-Cox转换等方法的基础上,用非参数检验Kruskal Wallis方法和多因素回归方法分析选取分组的分类节点变量。结果:分别对住院天数和住院费用这两种统计分类指标的35大类统计,筛选的主要分类节点变量为ICD-10国际疾病编码前一位码、是否手术、疾病危重度、伴随症和转归情况,也有少数大类涉及性别、年龄、入院情况、是否随诊、是否抢救以及是否感染等变量。结论:选取的分类节点变量科学。
Objective:To decide scientific classification node variables so as to provide the basis for the case grouping.Methods:To select the classification node variables by applying non parametric test(Kruskal Wallis method) and multivariate regression analy-sis based on the basic methods which test the normality, nuclear density map and Box-Cox transformation.Results:Statistical 35 cate-gories of the statistical classification of length of hospital stay and hospitalization expenses respectively, ICD-10 International Classifi-cation of Diseases coding, whether take surgery, critical illness degree, along with the disorder were selected as primary classificationnode variables. There were the outcomes of the gender, age, admission, whether follow-up, whether rescue and whether infect were re-ferred.Conclusion:The grouping results were scientific.
出处
《中国卫生经济》
北大核心
2014年第9期20-23,共4页
Chinese Health Economics
基金
美国中华医学基金会(GMB)项目资助(C09-986)
关键词
分类节点变量
影响因素
住院天数
住院费用
绩效评价
classification node variables
influencing factor
length of hospital stay
hospitalization expense
performance evaluation
