糖尿病国际功能、残疾和健康分类综合核心组套的图建模  

Graphical modeling of International Classification of Functioning,Disability and Health comprehensive core sets for diabetes mellitus

作　　者：林枫[1] 江钟立[1] 吴亚文[1] 李红卫[1] 王萍[1] 田智慧[1] 程少强[1] 

机构地区：[1]南京医科大学第一附属医院,南京210029

出　　处：《中国康复医学杂志》2012年第4期308-314,共7页Chinese Journal of Rehabilitation Medicine

基　　金：江苏省社会发展支撑计划(BE2009613)

摘　　要：目的:绘制基于国际功能、残疾和健康分类(ICF)的糖尿病功能地形图,为糖尿病的功能研究提供新的依据。方法:收集200例糖尿病患者的便利样本,以糖尿病ICF综合核心组套中的99个分类项目为变量进行图建模。缺失值以多重填补法进行填补,采用"最小的绝对缩减和变量选择算子"发掘变量之间的条件依存关系,采用自举法重采样技术和置信区间检验法强化模型的信度和效度,以R软件和Pajek2.04进行建模和分析。结果:在99个分类项目中,有61个相互联系构成了功能地形图中的最大的独立结构,并且可以解析出一个由44个分类项目组成的2-核。"s220"(眼球的结构)、"s6100"(肾)、"d760"(家庭人际关系)、"d455"(到处移动)和"d450"(步行)因连接范围较大而居于地形图的中心地位。结论:图建模所绘制的糖尿病功能地形图可以揭示功能类别之间的复杂的关系结构。这些关系结构既可以从临床知识中寻找到依据,也可以为应用ICF指导糖尿病康复的临床实践和科学研究提供线索。Objective： To provide new supports for studies of functionings in diabetes mellitus by functioning mapping which based on the International Classification of Functioning, Disability and Health （ICF）. Method： Graphical modeling was based on a convenience sample of 200 diabetic persons. The 99 categories of comprehensive ICF core sets for diabetes mellitus were defined as variables. Missing values were imputed by multiple imputation method. The ＂least absolute shrinkage and selection operator（LASSO）＂ was used for mining condi- tional dependencies between the variables. Bootstrap resampling method and confidence interval approach were used to enhance the reliability and validity of model selection. R software and Pajek 2.04 were used for graphical modeling and analysis. Result： In the 99 ICF categories, there are 61 interconnected categories which organized into the maximal independent component in the functioning mapping. A 2-core composed of 44 categories can be decomposed from the maximal component. ＂s220＂（structure of eyeball）, ＂s6100＂（kidneys）, ＂d760＂（family relationships）, ＂d455＂（moving around） and ＂d450＂（walking） are centrally positioned categories because of their high connections. Conclusion： Functioning mapping by graphical modeling can reveal complex relational structures embedded in functioning categories. These relational structures have evidences from clinical knowledge. They also provide clues for using ICF to guide clinical practices and scientific studies in diabetic persons.

关 键 词：糖尿病 国际功能 残疾和健康分类 图建模 康复 网络分析 

分 类 号：R587.1[医药卫生—内分泌] R492[医药卫生—内科学]