真实世界研究中的因果推断及粗化精准匹配

Causal Inference and Coarsened Exact Matching in Real-world Study

对因果推断及粗化精准匹配进行介绍,促进其在真实世界研究中的应用。基于文献研究,对因果推断与真实世界研究、匹配的原理及作用、粗化精准匹配及其应用进行全面介绍。真实世界研究应使用匹配等方法提升因果推断的证据力度,为药品监管及相关决策提供依据;粗化精准匹配比倾向性评分匹配等方法更加有效、便捷,原理与方法更加科学合理,但也有对原人群的代表性不够强等局限性;粗化精准匹配可独立使用,也可与其他匹配方法共同使用,研究者应当基于数据与研究设计选择合适的匹配方法。

To introduce causal inference and coarsened exact matching, and to promote its application in real-world study. Based on literature review, the study introduced causal inference and real-world study, principle and implication of matching methods, coarsened exact matching and its application. Real-world research should use matching methods etc to enhance the evidence strength of causal inference, so as to provide a basis for drug regulation and related decision-making. Coarsened exact matching is more effective and convenient than the tendency score matching method, and the principle and method are more scientific and reasonable, while some limitation such as the poor representativeness of study population exists. Coarsened exact matching can be used independently or in combination with other matching methods, thus appropriate matching method should be chosen based on the actual situation of data and research design.