为探讨中药临床试验重复测量数据的统计分析方法，对中药复方胃康胶囊治疗十二指肠溃疡的多中心随机对照盲法临床试验的疗效进行秩和检验和重复测量分析的比较。结果表明，重复测量因子（时间作用）、医院间与组间的交互作用、重复测量因子与医院间以及重复测量因子与组间的交互作用对总疗效的影响均有显著意义；交互作用的进一步分析提示，每家医院内组间差异对疗效有显著性影响，但其影响随医院的不同而不同；同时，同一组内各医院疗效评价结果的差异也有显著意义。从拟合残差模型的疗效结果及症状总分的重复测量方差分析结果看，重复测量因子对疗效的影响有显著意义，也即不同时点的疗效有所不同；但总的来说，组间疗效差异无显著意义，即表明试验药与对照药雷尼替丁对症状的疗效无显著性差异。上述结果表明，对于多中心临床试验采用多个时点观测的情况，由于常常存在相关的交互作用的影响，应用重复测量分析方法有助于更合理地分析和解释试验的结果。 In order to explore the statistical analysis method of repeated measurement data in clinical trials of traditional Chinese medicine, the curative effect of multi-center randomized controlled blind trial of traditional Chinese medicine Huikang Capsule in the treatment of duodenal ulcer was compared by rank sum test and repeated measure analysis. The results showed that the influence of repeated measures (time effect), hospital-to-group interactions, repeated measures and inter-hospital and repeated measures and interaction between groups on the overall efficacy were significant; interaction The analysis suggests that the differences between groups in each hospital have a significant effect on the efficacy, but the impact varies with the hospital; meanwhile, the differences in the evaluation results of the hospitals in the same group are also significant. According to the result of repeated-measures ANOVA of the results of the fitting residual model and the symptom score, the effect of repeated measures on the efficacy is significant, that is, the curative effect is different at different time points; however, in general, No significant difference between the efficacy of the test, indicating that the test drug and the control drug ranitidine no significant difference in the efficacy of symptoms. The above results indicate that using multiple time-point observations for multi-center clinical trials, the use of repeated measures analysis methods can help to more rationally analyze and interpret the results of the trial because of the often associated effects of interactions.