血瘀证量化诊断流行病学调查数据的判别分析

目的：运用统计学方法研究对于血瘀证诊断具有较好区分能力的诊断方法和诊断标准。方法：在现代流行病学原则指导下进行血瘀证诊断的多中心、大样本临床资料收集，对其结果进行判别分析处理。结果：对75个变量进行35步的逐步判别，共有33个项目进入方程，由此建立了血瘀型和非血瘀型的判别函数，并对33个自变量对血瘀证判别的贡献度进行区分。并对2004例患者进行了对判别结果的回顾性检验，总预测正确率为91．6％。结论：逐步判别分析可筛选出对鉴别两类具有不同属性的人群有较大贡献的条目，从而使其结果具有较好的区分度。

A Discriminant Analysis of Epidemiological Survey Data on Quantitative Diagnosis of Blood Stasis Syndrome

Here we report an investigation into the diagnostic approaches that are in a better position to differentiate the blood stasis syndrome(BSS).We used epidemiological survey data on BSS collected from three Chinese areas,and made a discriminant analysis of the database.We made 35 stepwise discriminant analysis of 75 items,allowing 33 items into the equation/function.Discriminant functions of BSS and Non-BSS were established.The contributions of 33 items were differentiated.We had a 91.6% prediction accuracy for 2004 cases,through a retrospection comparison.It is concluded that the stepwise discriminant analysis can be employed to screen the items that have a better contribution to differentiating the patients of different diseases.