Understanding the circumstances of death and how they can be prevented is critical for improving how we live our lives. At the national level, over- or under-reporting of deaths can profoundly affect policy decisions, altering global economies and the day-to-day lives of citizens. At the individual level, understanding how and why deaths occurred may prevent similar deaths or serious harm from occurring in the future.
National statistics and electronic health records are often used to examine mortality rates. However, these datasets do not provide insight into the circumstances that led to death, which is provided in coroners’ reports following inquests. Collecting and analysing coroners’ reports in England and Wales has been extremely difficult until the development of the Preventable Deaths Tracker database (https://preventabledeathstracker.net/) in 2020. Previous research has used this database to explore deaths across a range of health and care settings, including deaths involving medicines, sepsis, falls, and maternal care. However, data from coroners and health care systems have not been linked before.
This project will establish methods for linking electronic health records with coroners’ reports to identify new insights and methods from such linkages and codesign tools and dashboards that could support health care professionals to learn following death to deliver safer care. It is an opportunity to harness indexed and searchable health records via collaboration across the Preventable Deaths Tracker and Cogstack (https://www.kcl.ac.uk/research/cogstack); to apply or customise specialised natural language processing models for longitudinal analysis or predictive modelling. The project will include the application of artificial intelligence, codesign, user testing, and epidemiology to identify clinical insights and unlock new pathways for death prevention.