美国医疗数据一本通

《美国医疗数据一本通》是一本面向医疗数据从业者的实战手册，带你走通医疗数据的两个世界：一边是临床试验数据（CDISC 的 SDTM 与 ADaM、define.xml 与提交包）那套森严的秩序，另一边是真实世界数据（理赔 claims 与电子病历 EHR、编码体系与 OMOP 通用模型）的庞杂与鲜活。全书 42 章加五个附录，每一章都落到一个能跑的交付物；所有代码只用 SAS 与 Oracle SQL 写成，且全部实测可跑。它不堆砌算法，而是反复回答那个真正的难题--这堆数据为什么长这样、哪里会坑你。从数据摸底、口径定义、指标计算，到 SDTM/ADaM 建域、TFL 输出、提交包组装，再到真实世界研究设计、偏倚规避、统计基本功、向非技术者讲清结论，直到职业地图与新人生存。写给刚入行、想把每一份数据背后的人当人对待的分析师。

U.S. Healthcare Data All-in-One is a hands-on handbook for anyone working with healthcare data, bridging its two worlds: the strict order of clinical-trial data (CDISC's SDTM and ADaM, define.xml, and submission packages) and the sprawl of real-world data (insurance claims and electronic health records, medical coding systems, and the OMOP common data model). Across 42 chapters and five appendices, every chapter is built around a runnable deliverable, and all code-written exclusively in SAS and Oracle SQL-has been actually tested. Rather than piling up algorithms, the book keeps answering the question that really matters: why does this data look the way it does, and where will it trip you up? It moves from data profiling, specification, and metric calculation, through building SDTM/ADaM domains, producing TFLs, and assembling a submission package, to real-world study design, avoiding bias, statistical fundamentals, and explaining results to non-technical audiences-closing with a career map and a new-hire survival guide. Written for analysts entering the field who want to treat the person behind every record as a person.