摘要
目的探讨观测结果为重复测量证候数据时潜在转移分析在样品聚类中的应用,及聚类后不同类别缺血性中风病患者各证候随时间的变化规律。方法采用潜在转移分析将993例缺血性中风病住院患者按内风、内火、痰湿、血瘀、气虚和阴虚6个证候分别在入院当天、7、14、28天4个时间点上进行聚类,计算所占百分比最大的两组患者在各时点上各证候出现的频率,并绘制线图。结果根据潜在转移分析模型的拟合统计量得出7个潜在状态的模型是首选模型,缺血性中风病住院患者可聚为7个亚组,"内风+内火+痰湿+血瘀"组498人,"血瘀+内风"组251人,"阴虚"组87人,"血瘀+内火+痰湿"组63人,"气虚"组52人,"内火+痰湿"组26人,"血瘀"组16人。经潜在转移分析聚成的7组患者,占比例最大的两组患者各证候的发生率及随时间的变化趋势不尽相同。结论潜在转移分析可用于观测结果变量为多个二值变量且为重复测量证候数据的样品聚类,聚类后患者各证候随时间的变化规律不同,为临床医生确定中药干预证候的最佳时点提供了医学证据,对于缺血性中风患者发病不同阶段各证候组合形式的针对性治疗具有指导意义。
Objective To explore the application of latent transition analysis in a sample cluster of repeatedly-measured syndrome data response variables, and syndrome variation regularity of ischemia stroke pa- tients with different class. Methods Using Latent transition analysis, we gave 993 patients a classification according to 6 syndromes at 4 days, and calculated the syndrome frequency of different days of the two classes that take the largest proportions,then drawed line graph. Results Seven-status model is preferred according to fit statistics, and the seven classes number of patients was 498,251,87,63,52,26 and 16 respectively. The syndrome variation regularity of the two classes that take the largest proportions is dif- ferent. Conclusion Latent transition analysis can resolve the sample classification of binary repeatedly-measured syndrome data response varia- bles. The difference in syndrome variation regularity of classes is more helpful to find medication time, and guide clinical intervention at different stage aimed at patients suffering from ischemia stroke in different c/asses.
出处
《中国卫生统计》
CSCD
北大核心
2013年第6期782-785,共4页
Chinese Journal of Health Statistics
基金
国家重点基础研究计划(973计划)课题(课题编号2003CB517102)
国家重点科技专项“重大新药创制”课题(课题编号2009ZX09502-028)
国家科技重大专项课题(课题编号2011ZX09302-006-01)
关键词
中医
重复测量
潜在转移分析
转移概率
Traditional Chinese medicine
Repeatedmeasurement
Latent transition analysis
Transition probabilities
