基于中医真实疗效病证结合多个主要终点协同评价方法的构建

中医药临床研究疗效需要科学的评价方法彰显，而疗效结局指标评价至关重要。前期研究提出了病证结合多个主要终点评价法的概念并进行了应用探索。然而，如何确定多个主要终点最优样本量、实现终点间动态转换以及建立多个终点类型重复测量模型却是亟需进一步深入研究的关键环节。本研究围绕中医药病证结合多个主要终点临床试验开展过程中的三个阶段进行：设计阶段，引入经典诊断试验的思路，采用ROC曲线建立最优化样本量调整策略；实施阶段，开展两阶段适应性研究，盲态下进行期中分析，通过Bonferroni校正因子法和总检验效能的循环迭代，适应性调整样本量和单阶段检验水准，并进行主次要终点的动态转换；分析阶段，建立混合终点重复测量多水平模型，通过与传统单终点模型比较，对模型进行验证和优化。本研究将以两个同期开展的RCT为实例，构建病证结合多个主要终点协同评价方法，为提高中医临床真实疗效评价的客观化和标准化提供新思路和新方法。

The development of a scientific evaluation tool is a prerequisite to evaluate the effect of Traditional Chinese Medicine (TCM), and the outcomes evaluation is of vital importance. In the previous studies, we proposed the concept of evaluation mode with multiple primary endpoints based on combination of diseases and symptoms, and the method has been applied. However, how to determine the optimal sample size of multiple primary endpoints, realize the dynamic transformation between primary and secondary endpoints, and establish the hierarchical model with multiple endpoint types and repeated measurements, all of these are the key points that need to be further studied. This study will focus on the development of clinical trials of TCM with multiple primary endpoints based on combination of diseases and symptoms, and include three stages: In the design stage of clinical trials: we will introduce the idea of classic diagnostic tests, and use the ROC curve to establish the optimal sample size adjustment strategy of multiple primary endpoints based on combination of diseases and symptoms. In the implementation stage of clinical trials: we will use the two-stage adaptive adjustment strategy, according to the results of the interim analysis in blindness, use Bonferroni-based correction factor to calculate single stage alpha, through the cyclic iteration of single stage alpha and total power, the sample size and single stage alpha level will be adjusted adaptatively. In addition, the dynamic transformation of primary and secondary endpoints will be carried out according to the results of the interim analysis. In the analysis stage of clinical trials: we will establish a hierarchical model with multiple endpoint types and repeated measurements, then verify and optimize the model by comparing it with the traditional single endpoint models. This study will take two clinical trials conducted in the same period as examples to construct an evaluation mode with multiple primary endpoints based on combination of diseases and symptoms, which will provide new ideas and methods for improving the objectification and standardization of the real effect evaluation of TCM.