Integration of a Large Language Model into the Electronic Health Record

Abstract

Electronic Health Records (EHRs) are an electronic version of patients medical history, in order to ensure that all clinical information is gathered in one place and to ensure a proper follow-up. Due to the generation of vast amounts of data, in some situations it can be complex to obtain a clear and accurate view of the patient's current condition, specially when time is a strong constraint. In this context, the integration of Large Language Models (LLMs) into EHRs can help to summarize clinical content by taking into account the automated summarization of complex patient records, extraction of key clinical concepts, and contextual interpretation of longitudinal health data. So, the implementation of LLMs in EHR environments could have significant potential to enhance workflow efficiency, reduce documentation burden, and improve the accuracy and clarity of patient summaries. In this paper, a distinguishing feature of the development of a specialized LLM framework is proposed, which adapts the summaries according to the profile of the clinical professional. In more detail, the proposal concerns the integration of information from various medical specialty services with an LLM to improve patient diagnosis and follow-up across clinical specialties.