FHIR-Starter

LLM supported automatic structuring of medical data

Project description:

Medical findings and doctor's letters are often transmitted in full text between doctors, whereby the formats vary greatly between hospitals, care facilities and doctors in private practice. This lack of standardisation makes it difficult for practice management systems (PVS) and hospital information systems (HIS) to integrate findings into patient files in a structured manner. They are often archived as PDFs or in paper form. This takes a lot of time to read and leads to the loss of important information that is not easily accessible. In addition, full-text information is difficult to use for medical and pharmaceutical research in Germany, which contributes to clinical research increasingly being outsourced abroad.

The project will develop a software service that uses Large Language Models (LLM) and Natural Language Processing (NLP) methods to process and analyse full-text documents and convert them into standardised data formats such as FHIR or OMOP. The service will offer open interfaces that enable service providers, healthcare software providers and secondary users of healthcare data to incorporate the data into their systems in a structured way, making it interoperable and accessible. This will have direct benefits for medical care, as staff will be relieved and detailed information will be more easily available to healthcare professionals. At the same time, it will make it possible to collect medical data in a structured way and make it available to researchers, which will increase Germany's attractiveness for the medical research landscape.

Consortium:

Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e. V., Fraunhofer IESE,
INSIDERS Technologies GmbH,
Charité - University Medicine Berlin

Duration:

February 2025 - January 2028

Budget:

Total funding: € 2.1 million
Funding amount: € 1.6 million