Posts in tags: ai

In a previous post, I described how we created a semantic layer over FHIR to support AI-driven capabilities such as semantic search and risk stratification. The response prompted a number of follow-up questions, many of them less about AI itself and more about the architectural decisions behind it: Why use FHIR narratives? Why are embeddings useful? What role does retrieval play? Why does chunking matter? And what exactly changes when agents are introduced into the architecture?
There is no shortage of discussion about the opportunities that AI may create within health and social care. Modern language models appear increasingly capable of interpreting narrative, answering questions and identifying relationships across both structured and unstructured information. In a domain where information is often recorded and understood as a mixture of records, observations and stories, it is tempting to conclude that the way data is represented no longer matters. In other words, perhaps structured data is dead.
The recent Shared Care Record Summit Learnathon explored how International Patient Summaries (IPS) can be generated, shared, and consumed using FHIR standards and secure data sharing protocols. Our team showcased how a FHIR International Patient Summary (IPS) can be assembled on-the-fly from federated data using the NHS National Record Locator (NRL) and demonstrated patient-mediated sharing powered by an AI digital assistant.