目的运用倾向评分法分析真实世界中参芪扶正注射液对门冬氨酸氨基转移酶(AST)水平的影响。方法基于全国20家医院的医院信息系统(HIS)数据,选取年龄在18~80岁之间的患者,将使用参芪扶正注射患者作为观察组;将使用1种功效类似的中药注射剂患者作为对照组。运用广义增强模型(GBM)倾向评分加权法,平衡大量混杂因素的偏倚,利用多元Logistic回归估计,分析参芪扶正注射液对AST水平的影响。结果平衡57个混杂因素后,数据显示使用参芪扶正注射液比使用1种功效类似的中药注射剂导致AST发生异常变化的可能性要小,但差异无统计学意义。敏感性分析对潜在偏倚不敏感,认为本研究可能不存在没有观测到的变量。结论基于现有的HIS数据和研究方法认为,使用参芪扶正注射液对AST指标异常变化的影响不显著。 Objective To analyze the effect of Shenqi Fuzheng injection on aspartate aminotransferase (AST) in the real world by propensity score method. Methods Based on the hospital information system (HIS) data of 20 hospitals in China, patients aged 18 to 80 years were enrolled in this study. Shenqi Fuzheng injection patients were used as observation group. One kind of traditional Chinese medicine injection with similar efficacy was used as control group. The GBM propensity score weighted method was used to balance the bias of a large number of confounding factors. Multivariate logistic regression was used to evaluate the effect of Shenqi Fuzheng Injection on AST. Results After 57 confounding factors were balanced, the data showed that the use of Shenqi Fuzheng Injection was less likely to cause abnormal changes of AST than the use of a traditional Chinese medicine injection with similar efficacy, but the difference was not statistically significant. Sensitivity analysis is not sensitive to potential bias, and we think there may be no unobserved variables in this study. Conclusion Based on the existing HIS data and research methods, it is concluded that the use of Shenqi Fuzheng injection has no significant effect on the abnormal changes of AST.