Implementation of Large Language Models in Electronic Health Records

Implementation of Large Language Models in Electronic Health Records | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Implementation of Large Language Models in Electronic Health Records Maxime Griot, Jean Vanderdonckt, Demet Yuksel This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7029913/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Electronic Health Records (EHRs) have greatly improved access to clinical documents which also resulted in new challenges for clinicians, with some spending over 50% of their time on non-clinical tasks. While Large Language Models (LLMs) offer promise for reducing this burden, most implementations focus on synthetic benchmarks rather than clinical deployment. This paper presents a secure, fully on-premises, GDPR-compliant LLM chatbot integrated into the Epic EHR system at a European university hospital. The system utilizes Qwen3-235B with Retrieval Augmented Generation (RAG) for contextual awareness across internal records, regional eHealth documents, and medical literature. During a one-month pilot with 28 physicians from nine specialties, over 400 multi-turn conversations were initiated, with 64% of participants using the tool daily. Primary use cases included chart summarization, targeted retrieval, and support for clinical reasoning. Clinicians emphasized the model’s utility in rapidly synthesizing dispersed information, while diagnostic use remained limited. Our results demonstrate the technical feasibility and adoption patterns of integrating LLMs into production EHR systems, providing a replicable framework for secure clinical AI deployment. Health sciences/Health care/Quality of life Health sciences/Medical research/Translational research artificial intelligence clinical applications clinical investigations electronic health records fhir generative artificial intelligence large language models machine learning transformers user-machine interactions Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7029913","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":479658536,"identity":"f0f687a3-d925-4b74-9899-02f68854d46e","order_by":0,"name":"Maxime Griot","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBElEQVRIiWNgGAWjYJCCA2CSGYg/MDDIgDkJYERYC2PjDAYGHqK0wABjMw9MCwMeLbrtZwwP3ahgsNveznz8sW2bHQ+/RI7hh4c7GPL4cWgxO5OWcDjnDEPynMNsic25bck8kjNyjCUSzzAUSzbg0HIg+cDh3DaGZAlmHkOglgM8BjfSEiQS2xgSNxzAoeX8w4bDuf+gWiwhWpJ/gLTsx6XlBsiWBgY7sBZGsJbkYxBbcPnlxjOgX45JJEgwsyXO7DkH9EvP42MWiW0SxRI4HZZj/DmnxsZegv/wgQ8/yuzk+NkTm2/+bLPJ48fhfSiQSESXl8CrHgTsCaoYBaNgFIyCkQsADAxa/ZqbFNUAAAAASUVORK5CYII=","orcid":"","institution":"Université Catholique de Louvain","correspondingAuthor":true,"prefix":"","firstName":"Maxime","middleName":"","lastName":"Griot","suffix":""},{"id":479658537,"identity":"8be58b2e-108b-4459-9bab-6330c4cd76c7","order_by":1,"name":"Jean Vanderdonckt","email":"","orcid":"","institution":"Université Catholique de Louvain","correspondingAuthor":false,"prefix":"","firstName":"Jean","middleName":"","lastName":"Vanderdonckt","suffix":""},{"id":479658538,"identity":"ec526af7-0f2e-4847-bd6f-fead743bca89","order_by":2,"name":"Demet Yuksel","email":"","orcid":"","institution":"Cliniques universitaires Saint-Luc","correspondingAuthor":false,"prefix":"","firstName":"Demet","middleName":"","lastName":"Yuksel","suffix":""}],"badges":[],"createdAt":"2025-07-02 13:38:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7029913/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7029913/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":86603624,"identity":"ae762f14-0f4b-4b79-962b-d0cd44dde062","added_by":"auto","created_at":"2025-07-13 14:31:55","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1179490,"visible":true,"origin":"","legend":"","description":"","filename":"npjdmimplementationchatbot.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7029913/v1_covered_39892e66-c922-4b96-9879-ee1325504032.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Implementation of Large Language Models in Electronic Health Records","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"artificial intelligence, clinical applications, clinical investigations, electronic health records, fhir, generative artificial intelligence, large language models, machine learning, transformers, user-machine interactions","lastPublishedDoi":"10.21203/rs.3.rs-7029913/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7029913/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Electronic Health Records (EHRs) have greatly improved access to clinical documents which also resulted in new challenges for clinicians, with some spending over 50% of their time on non-clinical tasks. While Large Language Models (LLMs) offer promise for reducing this burden, most implementations focus on synthetic benchmarks rather than clinical deployment. This paper presents a secure, fully on-premises, GDPR-compliant LLM chatbot integrated into the Epic EHR system at a European university hospital. The system utilizes Qwen3-235B with Retrieval Augmented Generation (RAG) for contextual awareness across internal records, regional eHealth documents, and medical literature. During a one-month pilot with 28 physicians from nine specialties, over 400 multi-turn conversations were initiated, with 64% of participants using the tool daily. Primary use cases included chart summarization, targeted retrieval, and support for clinical reasoning. Clinicians emphasized the model’s utility in rapidly synthesizing dispersed information, while diagnostic use remained limited. Our results demonstrate the technical feasibility and adoption patterns of integrating LLMs into production EHR systems, providing a replicable framework for secure clinical AI deployment.","manuscriptTitle":"Implementation of Large Language Models in Electronic Health Records","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-04 03:09:35","doi":"10.21203/rs.3.rs-7029913/v1","editorialEvents":[{"type":"communityComments","content":1}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c09e68de-0be1-4b46-abaa-5c95175c56fc","owner":[],"postedDate":"July 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":50934286,"name":"Health sciences/Health care/Quality of life"},{"id":50934287,"name":"Health sciences/Medical research/Translational research"}],"tags":[],"updatedAt":"2025-07-13T14:23:34+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-04 03:09:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7029913","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7029913","identity":"rs-7029913","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}