目的 基于"成分反映活性，活性指向功效"的中药质量控制研究思路，建立用于赤芍饮片等级评价的Logistic回归模型。方法 采用超高效液相色谱（UPLC）法测定质量标志物芍药苷的含量，以赤芍抗凝血生物效价，羟自由基抑制率和DPPH抑制率作为生物活性评价指标，运用Logistic回归分析法将各批次芍药苷含量和生物活性指标进行关联分析，最终建立用于赤芍4个等级（优、良、中、差）评价研究的"主成分分析-Logistic回归"模型。结果 等级评价结果显示，各批次赤芍的等级概率达到了95%以上。16个批次的赤芍饮片分布在优级、良级和中级各有5个批次，差级有1个批次。结论 初步建立了不同批次赤芍饮片等级评价的新方法，并用于赤芍饮片质量的评价。

Objective A logistic regression model for grade evaluation of Paeoniae Radix Rubra medicinal slices was constructed based on the quality control idea of traditional Chinese medicines that "ingredients reflect activity and activity expresses effect". Methods Q-marker content of paeoniflorin was tested by ultra-high performance liquid chromatography (UPLC). Anticoagulation valence, inhibition rate of hydroxyl radical and DPPH clearance rate were used as evaluation indexes of biological activity. Correlations between paeoniflorin content in different batches and bioactivity indexes were analyzed by the logistic algorithm. Finally, a "principal component analysis-Logistic regression" model for grade evaluation of Paeoniae Radix Rubra was constructed. Results According to grade evaluation results, the grade probability of different batches of Paeoniae Radix Rubra was higher than 95%. Among 16 batches of Paeoniae Radix Rubra, 15 batches were evaluated excellent, good and moderate (five for each), and one batch was evaluated poor. Conclusion A new grade evaluation method for Paeoniae Radix Rubra medicinal slices is constructed preliminarily. It is applicable to quality evaluation of Paeoniae Radix Rubra medicinal slices.

国家自然科学基金项目（81501229）；国家自然科学基金项目（81773919）；陕西省创新人才推进计划青年科技新星项目（2017KJXX-71）；陕西省"特支计划"青年拔尖人才项目；国家中药标准化项目（ZYBZH-C-QIN-45）