Enhancing Real-World Data Extraction in Clinical Research: Evaluating the Impact of the Implementation of Large Language Models in Hospital Settings

Enhancing Real-World Data Extraction in Clinical Research: Evaluating the Impact of the Implementation of Large Language Models in Hospital Settings | 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 Research Article Enhancing Real-World Data Extraction in Clinical Research: Evaluating the Impact of the Implementation of Large Language Models in Hospital Settings Bin Wang, Junkai Lai, Han Cao, Feifei Jin, Qiang Li, Mingkun Tang, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3644810/v3 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Sep, 2024 Read the published version in European Heart Journal - Digital Health → Version 3 posted You are reading this latest preprint version Show more versions Abstract Aims This study aims to assess the feasibility and impact of the implementation of the ChatGLM for real-world data (RWD) extraction in hospital settings. The primary focus of this research is on the effectiveness of ChatGLM-driven data extraction compared with that of manual processes associated with the electronic source data repository (ESDR) system. Methods and results The researchers developed the ESDR system, which integrates ChatGLM, electronic case report forms (eCRFs) and electronic health records (EHRs). The LLaMA (Large Language Model Meta AI) model was also deployed to compare the extraction accuracy of ChatGLM in free-text forms. A single-center retrospective cohort study served as a pilot case. Five eCRF forms of 63 subjects, including free-text forms and discharge medication, were evaluated. Results Data collection involved electronic medical and prescription records collected from 13 departments. The ChatGLM-assisted process was associated with an estimated efficiency improvement of 80.7% in the eCRF data transcription time. The initial manual input accuracy for free-text forms was 99.59%, the ChatGLM data extraction accuracy was 77.13%, and the LLaMA data extraction accuracy was 43.86%. The challenges associated with the use of ChatGLM focus on prompt design, prompt output consistency, prompt output verification, and integration with hospital information systems. Conclusion The main contribution of this study is to validate the use of ESDR tools to address the interoperability and transparency challenges of using ChatGLM for RWD extraction in Chinese hospital settings. Medical Informatics Challenge Data extraction Electronic health records Interoperability Large language models Prompt Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Full Text Additional Declarations The authors declare no competing interests. Supplementary Files Supplementaryfile1promptexample.docx Supplementaryfile2PartelectronicCaseReportForm.docx Supplementary file 2 Part electronic Case Report Form GraphicalAbstract.png Graphical Abstract SupplementaryFigure1.png Supplementary Figure 1 The percentage of time saved by ChatGLM-Assisted compared with manual processes from three field types for patients in various departments. Cite Share Download PDF Status: Published Journal Publication published 11 Sep, 2024 Read the published version in European Heart Journal - Digital Health → Version 3 posted You are reading this latest preprint version Show more versions 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-3644810","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":254481996,"identity":"8fa782cf-330b-4046-87d1-299cd59819dc","order_by":0,"name":"Bin Wang","email":"","orcid":"https://orcid.org/0000-0003-0012-9835","institution":"Department of Cardiology, Beijing Tsinghua Changgung Hospital, School of Clinical Medicine, Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Wang","suffix":""},{"id":254481997,"identity":"2d264807-32f5-439c-ae20-82fa8aae7b50","order_by":1,"name":"Junkai Lai","email":"","orcid":"https://orcid.org/0000-0002-2272-3870","institution":"Institute of Automation, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Junkai","middleName":"","lastName":"Lai","suffix":""},{"id":254481998,"identity":"5c0c9ed5-952a-4fe0-a767-ccfe6b754907","order_by":2,"name":"Han Cao","email":"","orcid":"https://orcid.org/0000-0002-9491-6188","institution":"Medical Data Science Center, Beijing Tsinghua Changgung Hospital, School of Clinical Medicine, Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Han","middleName":"","lastName":"Cao","suffix":""},{"id":254481999,"identity":"ec3fa40e-764d-4c18-bf8d-62daf1c29d12","order_by":3,"name":"Feifei Jin","email":"","orcid":"https://orcid.org/0000-0002-4991-0158","institution":"Peking University People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Feifei","middleName":"","lastName":"Jin","suffix":""},{"id":254482000,"identity":"6e4cf5d2-bd2d-4db2-85a2-0990b980b94d","order_by":4,"name":"Qiang Li","email":"","orcid":"","institution":"Department of Information Administration, School of Clinical Medicine, Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Qiang","middleName":"","lastName":"Li","suffix":""},{"id":254482001,"identity":"4dc6d757-2893-4233-8a49-5b4f5a503475","order_by":5,"name":"Mingkun Tang","email":"","orcid":"","institution":"Beijing Tsinghua Changgung Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mingkun","middleName":"","lastName":"Tang","suffix":""},{"id":254482002,"identity":"8799f185-5e4e-4501-8ebe-7d71e302a260","order_by":6,"name":"Chen Yao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIie3OIQvCQBTA8ZMLltNVh6B+hCeCxeBXmQy2MmHRYDBpcX0fYya1bTzQcrB6waAISwbBZBFvimCQU5vg/eEdd3A/7gjR6X40KqcGt22cL9ZnpPU96UUfE1hjcvIHG3du4Jb6HIlR9ICcFwrCHbsa8qy/DB2goUBiTg9QCLiCxF6blsbYjwQjlB2RgPCAFsYKkh4kuaALKb+T7lsi8ldGaMnnJJEfg8obYorMrrIVNiPhAIbcZRWe+UmgIOXUTk5siHVIcbf3V52aMbFn27OCNOKnA8ph+SZ+dfVRffR0oKqbOp1O979dAfDOVaDQe3AKAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-4224-5535","institution":"Peking University Clinical Research Institute, Peking University First Hospital","correspondingAuthor":true,"prefix":"","firstName":"Chen","middleName":"","lastName":"Yao","suffix":""},{"id":254482003,"identity":"471772a7-b059-416a-864d-169779d66e0d","order_by":7,"name":"Ping Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAApklEQVRIiWNgGAWjYDACCTCy4eFnbyBNS5qMZM8B0rQctjG44UCkDvnZvQdvfKg5z8Nwg4Hxw8ccIrQwzjmXbDnj2G0extkNzJIztxGhhVkix0yat+E2D7PMATZmXmK0sEG0nONhk0ggUgsPRMsBHh6itUhI5IH8kswjwXOwmTi/yM/IBYWYnb398eaDHz4SowXoNBiDsYEo9chaRsEoGAWjYBTgAACxXS/BEPxqdAAAAABJRU5ErkJggg==","orcid":"","institution":"Department of Cardiology, Beijing Tsinghua Changgung Hospital, School of Clinical Medicine, Tsinghua University","correspondingAuthor":true,"prefix":"","firstName":"Ping","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2023-11-21 15:41:17","currentVersionCode":3,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-3644810/v3","doiUrl":"https://doi.org/10.21203/rs.3.rs-3644810/v3","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1093/ehjdh/ztae066","type":"published","date":"2024-09-12T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":63648276,"identity":"aa597fe6-b688-47ce-a623-75de44ded3bd","added_by":"auto","created_at":"2024-08-30 14:25:20","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":185881,"visible":true,"origin":"","legend":"\u003cp\u003eComposition of the electronic source data repository (ESDR) system and description of the prompt design.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/b22543391317ee488a0d861a.png"},{"id":63648833,"identity":"4774cadc-5e53-465b-9f24-b80248df24b3","added_by":"auto","created_at":"2024-08-30 14:33:20","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2496509,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of departments from which the patient data were drawn. The percentages and numbers of patients in different departments are listed in the legend.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/c9bdf3bf93f4820d0ab34baa.png"},{"id":63648285,"identity":"b2c80a0d-4483-4771-8f6b-e09d85407ea7","added_by":"auto","created_at":"2024-08-30 14:25:20","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2633317,"visible":true,"origin":"","legend":"\u003cp\u003eData completeness across different fields in the free-text forms for all patients. (Note: The phrase \"free-text forms\" refers to the overall category, including four forms: lifestyle and behavior, family history, health status, and vital signs.)\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/d895d6b5ac87e5a7fe972722.png"},{"id":63648284,"identity":"b9ae3c11-ad12-4cfc-bf2a-f8d63b3e46fc","added_by":"auto","created_at":"2024-08-30 14:25:20","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1100217,"visible":true,"origin":"","legend":"\u003cp\u003eThe mean number of characters in the admission notes of patients from different departments.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/4f2bf4f040ac8a1fc1730a70.png"},{"id":63648834,"identity":"dab5e87d-02f3-41d7-8447-7a5dd7acce55","added_by":"auto","created_at":"2024-08-30 14:33:20","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":3594084,"visible":true,"origin":"","legend":"\u003cp\u003eThe accuracy of ChatGLM or LLaMA data extraction from different fields of the free-text forms for all patients.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/692c02f6cec86f8d51e4fbf8.png"},{"id":63648835,"identity":"70ecca7b-6069-435a-a20e-acde582d0149","added_by":"auto","created_at":"2024-08-30 14:33:20","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1873008,"visible":true,"origin":"","legend":"\u003cp\u003eThe mean accuracy of ChatGLM or LLaMA in extracting data from free-text forms of patients across different departments.\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/14880652250f0243b8ec6334.png"},{"id":63648832,"identity":"cb4b43a5-6842-479d-af1c-b37f513edb0b","added_by":"auto","created_at":"2024-08-30 14:33:20","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":19221,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaryfile1promptexample.docx","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/97b57d0bc1782980337c9c89.docx"},{"id":63648278,"identity":"9087b78e-5307-4b80-a7a8-785a076f591f","added_by":"auto","created_at":"2024-08-30 14:25:20","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":26850,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary file 2 Part electronic Case Report Form\u003c/p\u003e","description":"","filename":"Supplementaryfile2PartelectronicCaseReportForm.docx","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/f79a91299fbda944dcf6fd02.docx"},{"id":63648282,"identity":"90b7cb64-22b0-408e-a79e-502a81444e68","added_by":"auto","created_at":"2024-08-30 14:25:20","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":255819,"visible":true,"origin":"","legend":"\u003cp\u003eGraphical Abstract\u003c/p\u003e","description":"","filename":"GraphicalAbstract.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/bbbe7c9e40962e5d6efa34af.png"},{"id":63648280,"identity":"c221a419-d331-41ce-9962-4693918ddc6c","added_by":"auto","created_at":"2024-08-30 14:25:20","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":2124764,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 1 \u003c/strong\u003eThe percentage of time saved by ChatGLM-Assisted compared with manual processes from three field types for patients in various departments.\u003c/p\u003e","description":"","filename":"SupplementaryFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-3644810/v3/9eaa4a5d594da83140a6a00b.png"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eEnhancing Real-World Data Extraction in Clinical Research: Evaluating the Impact of the Implementation of Large Language Models in Hospital Settings\u003c/strong\u003e\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"Beijing Tsinghua Changgung Hospital","isAcceptedByJournal":true,"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":"Challenge, Data extraction, Electronic health records, Interoperability, Large language models, Prompt","lastPublishedDoi":"10.21203/rs.3.rs-3644810/v3","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3644810/v3","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eAims\u003c/h2\u003e \u003cp\u003eThis study aims to assess the feasibility and impact of the implementation of the ChatGLM for real-world data (RWD) extraction in hospital settings. The primary focus of this research is on the effectiveness of ChatGLM-driven data extraction compared with that of manual processes associated with the electronic source data repository (ESDR) system.\u003c/p\u003e\u003ch2\u003eMethods and results\u003c/h2\u003e \u003cp\u003eThe researchers developed the ESDR system, which integrates ChatGLM, electronic case report forms (eCRFs) and electronic health records (EHRs). The LLaMA (Large Language Model Meta AI) model was also deployed to compare the extraction accuracy of ChatGLM in free-text forms. A single-center retrospective cohort study served as a pilot case. Five eCRF forms of 63 subjects, including free-text forms and discharge medication, were evaluated.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eData collection involved electronic medical and prescription records collected from 13 departments. The ChatGLM-assisted process was associated with an estimated efficiency improvement of 80.7% in the eCRF data transcription time. The initial manual input accuracy for free-text forms was 99.59%, the ChatGLM data extraction accuracy was 77.13%, and the LLaMA data extraction accuracy was 43.86%. The challenges associated with the use of ChatGLM focus on prompt design, prompt output consistency, prompt output verification, and integration with hospital information systems.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThe main contribution of this study is to validate the use of ESDR tools to address the interoperability and transparency challenges of using ChatGLM for RWD extraction in Chinese hospital settings.\u003c/p\u003e","manuscriptTitle":"Enhancing Real-World Data Extraction in Clinical Research: Evaluating the Impact of the Implementation of Large Language Models in Hospital Settings","msid":"","msnumber":"","nonDraftVersions":[{"code":3,"date":"2024-08-30 14:25:15","doi":"10.21203/rs.3.rs-3644810/v3","editorialEvents":[{"type":"communityComments","content":2}],"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}},{"code":2,"date":"2023-11-29 15:40:04","doi":"10.21203/rs.3.rs-3644810/v2","editorialEvents":[{"type":"communityComments","content":0}],"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}},{"code":1,"date":"2023-11-28 19:46:18","doi":"10.21203/rs.3.rs-3644810/v1","editorialEvents":[{"type":"communityComments","content":0}],"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":"0e1164f5-6bb7-4616-ac12-16b24a3f5800","owner":[],"postedDate":"August 30th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":26879672,"name":"Medical Informatics"}],"tags":[],"updatedAt":"2024-09-23T15:17:31+00:00","versionOfRecord":{"articleIdentity":"rs-3644810","link":"https://doi.org/10.1093/ehjdh/ztae066","journal":{"identity":"european-heart-journal-digital-health","isVorOnly":true,"title":"European Heart Journal - Digital Health"},"publishedOn":"2024-09-12 00:00:00","publishedOnDateReadable":"September 12th, 2024"},"versionCreatedAt":"2024-08-30 14:25:15","video":"","vorDoi":"10.1093/ehjdh/ztae066","vorDoiUrl":"https://doi.org/10.1093/ehjdh/ztae066","workflowStages":[]},"version":"v3","identity":"rs-3644810","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3644810","identity":"rs-3644810","version":["v3"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}