Study: FHIR-GPT Can Enhance Health Data Interoperability
Now that we are almost at the two-year mark of the launch of ChatGPT, perhaps we are finally entering an era where Artificial Intelligence (AI) can fully come to life in the healthcare arena?
In the wake of the 2022 launch of ChatGPT, many players in the healthcare IT space quickly embraced AI, mostly through their marketing messaging and discussions around ideal use cases. Of course, there have also been plenty of discussions around the benefits and dangers of AI in healthcare.
When it comes to health data interoperability, it also seems that we have been on the cusp of leveraging AI for transforming EHR data – until now.
A recent study published in the journal NEJM AI showcased an AI model developed by Northwestern Medicine investigators that improved the transformation of EHR data into standardized health resources more efficiently than current methods.
Called FHIR-GPT, this new model can improve the conversion of electronic patient record data into the Fast Healthcare Interoperability Resources (FHIR) standard. FHIR-GPT uses large language models to interpret and translate clinical data into the FHIR format, allowing it to be shared across different health systems.
According to the study, the model achieved a 90 percent success rate in converting EHR data into FHIR resources. It also improved the exact match rates of existing NLP pipelines by 3 percent for routes, 12 percent for dose quantities, 35 percent for reasons, 42 percent for forms, and more than 50 percent for timing schedules.
"This is going to greatly accelerate the pace of breaking down the walls between different health systems that hinder the aggregation of health data and the exchange of the data for performing large-scale research, which we are in great need of, especially with the advancement of the generative AI and large language model technology," said Yuan Luo, PhD, associate professor of Preventive Medicine, director of the Center for Collaborative AI in Healthcare, and chief AI officer at the Northwestern University Clinical and Translational Sciences (NUCATS) Institute.
Achieving effective health data exchange capabilities is highly complex, and much has been said about AI serving as a key solution for achieving simplification. Based on this new study, it seems that FHIR-GPT could be a viable solution.
RosettaHealth can assist with any health information challenges you might have, book a free consultation with one of our interoperability experts.