ClinicalTrials .gov (CT .gov) is the largest publicly accessible registry of clinical studies, yet its registry-oriented architecture and heterogeneous adverse event (AE) terminology limit systematic pharmacovigilance (PV) analytics. AEs are typically recorded as investigator-reported text rather than standardized identifiers, requiring manual reconciliation to identify coherent safety concepts. We present the ClinicalTrials .gov Transformation Database (CTG-DB), an open-source pipeline that ingests the complete CT .gov XML archive and produces a relational database aligned to standardized AE terminology using the Medical Dictionary for Regulatory Activities (MedDRA). CTG-DB preserves arm-level denominators, represents placebo and comparator arms, and normalizes AE terminology using deterministic exact and fuzzy matching to ensure transparent and reproducible mappings. This framework enables concept-level retrieval and cross-trial aggregation for scalable placebo-referenced safety analyses and integration of clinical trial evidence into downstream PV signal detection.
CTG-DB：基于本体的ClinicalTrials.gov转换数据库，用于支持跨试验的药物安全性分析 / CTG-DB: An Ontology-Based Transformation of ClinicalTrials.gov to Enable Cross-Trial Drug Safety Analyses
这篇论文开发了一个名为CTG-DB的开源工具，它能将ClinicalTrials.gov上杂乱无章的临床试验数据，特别是药物不良反应文本，自动转换成标准化的、可关联分析的数据库，从而让研究人员能更方便地进行大规模、跨研究的药物安全性评估。
源自 arXiv: 2603.15936
