代谢综合征组分的正交因子分析

结合代谢综合征(MS)组分的正交因子分析实例,介绍正交因子分析模型的原理及其在医学研究中的应用.使用1998-2001年上海社区人群MS患病率的现况研究资料,选取35～65岁资料完整的1877例女性作为研究对象,使用SPSS统计软件对MS相关组分进行正交因子分析,并计算因子得分.因子分析提取的前6个互不相关的公共因子可以基本反映原始指标86%的信息.通过正交旋转后6个公共因子的实际意义很清晰,依次主要反映肥胖、血压、血糖、胰岛素、TG和HDL-C指标的信息.根据给出的因子得分矩阵计算个体各因子得分及总得分.结果表明,MS的同一组分(除血脂外)内的变量呈高度相关性,但不同组分间虽有统计学关联,但非高度相关.正交因子分析的价值在于探查有高度相关关系的变量群,进而为探究其共同的潜在病理生理机制提供线索.

Orthogonal factor analysis on metabolic syndrome

To elucidate the principal of orthogonal factor analysis, using an example of factor analysis of metabolic syndrome. The basic structures and the fundamental concepts of orthogonal factor analysis were introduced and data involving 1877 women aged of 35-65 years, selected from a cross-sectional study, which was conducted in 1998 - 2001 in Shanghai, were included in this study. Factor analysis was carried out using principle components analysis with Varimax orthogonal rotation of the components of the metabolic syndrome. The different components of the metabolic syndrome were not linked closely with the other components and loaded on the six different factors,which mainly reflected by the variables of obesity, blood pressure, plasma glucose, plasma insulin, triglycerides and HDL-cholesterol respectively. Six major factors of the metabolic syndrome were uncorrelated with each other and explained 86% of the variance in the original data. The factor score and total factor score for the individual could be obtained according to the component score coefficient matrix. Although the components of the metabolic syndrome were related statistically, the finding of six factors suggested that the components of the metabolic syndrome did not show high degrees of intercorrelation. As a linear method of data reduction, the mode reduced a large set of measured intercorrelation variables into a smaller set of uncorrelated factors, which explained the majority of the variance in the original variables. Factor analysis was well suited for revealing underlying patterns or structure among variables showing high degrees of intercorrelation.