中医证候要素研究中潜在变量模型的应用

Application of latent variable models in research of TCM syndrome elements

通过应用几种潜在变量模型统计方法,对1 280例高血压病现场调查资料进行盲法的分析和数据处理,探索了疾病中医证候分类和证候要素提取的研究方法。提出病证结合中医证候要素研究的思路,即在临床流行病学全信息调查的基础上,采用潜在变量模型中因子分析模型、结构方程模型、项目反应理论中的等级反应模型、潜在类别模型等研究方法。研究结果认为疾病的中医证候分类和证候要素的提取研究采取潜在变量模型等系列分析方法是适宜的。

The research methods of disease classification of TCM syndromes and syndrome elements extraction were explored through applying the statistical methods of several latent variable models.The blind analysis and data treatment were conducted on the field survey material of 1 280 cases of hypertension.The author presented the research idea of TCM syndrome elements of disease-syndrome combination,i.e.on the base of clinical epidemiological overall information survey,several research methods were applied,including factor analysis model and structural equation model in latent variable models,and graded response model in item response theory and latent class model.The result indicated that the analysis methods of variable models were suitable for the research on disease classification according to TCM syndromes and syndrome elements extraction.