The common data model (CDM) is an important tool to facilitate the standardized integration of multi-source heterogeneous healthcare big data, enhance the consistency of data semantic understanding, and promote multi-party collaborative analysis. The data collections standardized by CDM can provide powerful support for observational studies, such as large-scale population cohort study. This paper provides an in-depth comparative analysis of the data storage structure, term mapping pattern, and auxiliary tools development of the three international typical CDMs, then analyzes the advantages and limitations of each CDM and summarizes the challenges and opportunities faced in the CDM application in China. It is expected that exploring the advanced technical concepts and practical patterns of foreign countries in data management and sharing will provide references for promoting FAIR (findable, accessible, interoperable, reusable) construction of healthcare big data in China and solving the current practical problems, such as the poor quality of data resources, the low degree of semantization, and the inabilities of data sharing and reuse.
通用数据模型（CDM）是促进多源异构健康医疗大数据标准化整合、增强数据语义理解一致性、推动多方协同分析的重要工具，经CDM标准化后的数据集合可为开展大规模人群队列等观察性研究提供有力支撑。本文深入比较分析了三项国际典型医学CDM的数据存储结构、术语映射模式和辅助工具研发情况，系统梳理各模型的优势、局限，总结了我国在CDM应用过程中所面临的挑战与机遇。期望通过探索国外在健康医疗大数据开放共享过程中的先进技术理念与实践模式，为推动我国健康医疗数据资源FAIR化建设，即数据可发现（findable）、可访问（accessible）、可互操作（interoperable）和可重用（reusable），解决当前数据资源质量不佳、语义化程度低、无法实现打通共享和重复利用等实际问题提供借鉴。.