目的·建设上海糖尿病临床专病大数据库，挖掘临床数据信息价值，开展真实世界研究工作。方法·糖尿病数据来源于上海申康医院发展中心的医联工程所汇集的临床数据，原始临床数据需经过脱敏加密、清洗、标准化、信息提取以及结构化等数据处理步骤，然后再根据具体研究目的和内容，采取医学统计或机器学习方法开展数据分析工作。结果·糖尿病数据库现已存储2013—2022年212万例糖尿病患者在37家医院1.5亿次的诊疗数据。通过临床分析展现了糖尿病疾病在现实环境中的基本特征和发展趋势；利用构建回顾性队列可以发现糖尿病的潜在风险因素；聚类分析、网络分析等机器学习方法能够揭示糖尿病疾病的内在规律和相互关系。结论·上海糖尿病临床专病大数据库的建立不仅可以总结和展现糖尿病临床现状，还可以利用真实世界临床数据开展研究获得更多具有临床价值的科研成果。

Objective ·To construct a clinical database of diabetes in Shanghai, mine the value of clinical data, and carry out real-world study. Methods ·The data were extracted from Shanghai Link Healthcare Database. All original clinical data have undergone standard processes such as desensitization, encryption, cleaning, standardization, information extraction and structuring, and clinical data were analyzed by the method of medical statistics or machine learning according to different research contents. Results ·The database has imported the clinical data of 150 million visits and treatment records of 2.12 million diabetic patients in 37 municipal hospitals over a ten-year period from 2013 to 2022. The overall analysis showed the basic characteristics and development trends of all aspects of diabetes disease in real-world settings, the potential risks of diabetes are discovered by constructing retrospective cohort, and the inherent patterns of the disease are revealed by using machine learning methods such as cluster analysis and network analysis. Conclusion ·The establishment of Shanghai Diabetes Clinical Database can not only summarize and show the clinical status of diabetes, but also obtain more scientific achievements with realistic clinical value by real-world clinical data study.