Aviation has the ATSB. Every significant air accident is investigated, publicly reported, and disseminated in a way that allows the entire industry to learn from it.
Medicine does not have a single, consistently accessible equivalent at scale. Morbidity and mortality meetings remain internal to institutions, and coronial findings—while public—are dispersed across jurisdictions, inconsistently structured, and difficult to analyse systematically. There is no unified, cross-jurisdictional search system that allows clinicians or researchers to easily explore patterns across thousands of cases.
I have long felt that this gap limits collective learning from preventable harm. When large language models made it feasible to extract structured information from large volumes of unstructured legal text, I built this project.
Coronial is a searchable database of publicly available Australian coroners' findings. Each case is tagged using information explicitly present in the original documents, including clinical specialty, setting of care, contributing factors, medication involvement, system issues, and coronial recommendations where stated. This allows patterns to be explored across cases that would otherwise be navigable only as individual PDFs.
All tagging and case summaries are AI-generated. This enables the scale of the project, but introduces inevitable limitations. Classifications may be incorrect, summaries may omit nuance, and complex cases may be misinterpreted. This database should be used as a discovery tool only, and the original finding should always be consulted before drawing conclusions.
The database is updated weekly. An automated pipeline scrapes newly published findings from all eight state and territory coronial websites and adds them to the database, so the collection grows as new findings are released.
Coronial was created by Dr CY Yew, an Australian anaesthetist with an interest in clinical safety, perioperative medicine, and systems learning from adverse outcomes.