MIMIC-IV-Ext-22MCTS: A 22 Millions-Event Temporal Clinical Time-Series Dataset with Relative Timestamp for Risk Prediction

MIMIC-IV-Ext-22MCTS: A 22 Millions-Event Temporal Clinical Time-Series Dataset with Relative Timestamp for Risk Prediction | 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 MIMIC-IV-Ext-22MCTS: A 22 Millions-Event Temporal Clinical Time-Series Dataset with Relative Timestamp for Risk Prediction jing wang, Xing Niu, Juyong Kim, Jie Shen, Tong Zhang, Jeremy Weiss This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6347897/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 Clinical risk prediction based on machine learning algorithms plays a vital role in modern healthcare. A crucial component in developing a reliable prediction model is collecting high-quality time series clinical events. In this work, we release such a dataset that consists of 22,588,586 Clinical Time Series events, which we term MIMIC-IV-Ext-22MCTS. Our source data are discharge sum- maries selected from the well-known yet unstructured MIMIC-IV-Note [1]. We then extract clinical events as short text span from the discharge summaries, along with the timestamps of these events as temporal information. The general- purpose MIMIC-IV-Note pose specific challenges for our work: it turns out that the discharge summaries are too lengthy for typical natural language models to process, and the clinical events of interest often are not accompanied with explicit timestamps. Therefore, we propose a new framework that works as follows: 1) we break each discharge summary into manageably small text chunks; 2) we apply contextual BM25 and contextual semantic search to retrieve chunks that have a high potential of containing clinical events; and 3) we carefully design prompts to teach the recently released Llama-3.1-8B [4] model to identify or infer temporal information of the chunks. We show that the obtained dataset is so informative and transparent that standard models fine-tuned on our dataset are achieving sig- nificant improvements in healthcare applications. In particular, the BERT model fine-tuned based on our dataset achieves 10% improvement in accuracy on medi- cal question answering task, and 3% improvement in clinical trial matching task compared with the classic BERT. The GPT-2 model, fine-tuned on our dataset, produces more clinically reliable results for clinical questions. Health sciences/Diseases Health sciences/Health care/Public health Clinical event Temporal information Time series MIMIC Contextual BM25 Contextual semantic search Natural language processing Large language model Question answering Clinical trial Full Text Additional Declarations There is NO Competing Interest. Supplementary Files datasetsubset.csv a subset of dataset 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-6347897","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":451195894,"identity":"4b6719e9-1dcf-44e6-8728-62b6359c1ada","order_by":0,"name":"jing wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3ElEQVRIiWNgGAWjYBACxmYYi72x8QGDwQFStPAcPmxAlBYEkEhLk2BgIEILczvv4ReMbXZ58g45ZtUFBXcY5N17DAg4jC/NgrEtudjwwBmz2zMMnjEYnjlDSAuPmQHjNubEjY09Zrd5DA4zGM5ISyBGS33iRiCjmFgtxg8Ytx1OnM/GlsYM0iIvkXyAoC0Mif+OJ27gYT4sDdTCY8BzGL8Ww/4zxh8+nKlOnD//YeNnnj+H5eTbGxvwa2lgYJMAuR0WiTwEY1MeGDUfwAyY0XDGKBgFo2AUjAIoAAD/mUgH6xwW6AAAAABJRU5ErkJggg==","orcid":"","institution":"NLM","correspondingAuthor":true,"prefix":"","firstName":"jing","middleName":"","lastName":"wang","suffix":""},{"id":451195895,"identity":"ad5034ab-45fc-44dd-91be-609672c84612","order_by":1,"name":"Xing Niu","email":"","orcid":"","institution":"Amazon Web Services","correspondingAuthor":false,"prefix":"","firstName":"Xing","middleName":"","lastName":"Niu","suffix":""},{"id":451195896,"identity":"fe59cdd2-a5f6-4bdb-b6be-abae60a36ccf","order_by":2,"name":"Juyong Kim","email":"","orcid":"","institution":"Carnegie Mellon University","correspondingAuthor":false,"prefix":"","firstName":"Juyong","middleName":"","lastName":"Kim","suffix":""},{"id":451195897,"identity":"1dcc5573-9f60-45ae-a45c-6f507227fc7b","order_by":3,"name":"Jie Shen","email":"","orcid":"","institution":"Stevens Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Shen","suffix":""},{"id":451195898,"identity":"c8267531-2cb5-47da-82d8-dcda96dcd18d","order_by":4,"name":"Tong Zhang","email":"","orcid":"","institution":"Hong Kong University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Tong","middleName":"","lastName":"Zhang","suffix":""},{"id":451195899,"identity":"ffc951e7-00a3-4be7-809e-4872cf3e0525","order_by":5,"name":"Jeremy Weiss","email":"","orcid":"","institution":"NLM","correspondingAuthor":false,"prefix":"","firstName":"Jeremy","middleName":"","lastName":"Weiss","suffix":""}],"badges":[],"createdAt":"2025-03-31 22:00:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6347897/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6347897/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91374114,"identity":"92a36c1a-a41a-4167-bc35-f3b03c814883","added_by":"auto","created_at":"2025-09-15 19:33:45","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":445214,"visible":true,"origin":"","legend":"Article File","description":"","filename":"cmtxdata.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6347897/v1_covered_35ae383e-a643-46fe-a3c9-f7729b3e2ab0.pdf"},{"id":89967369,"identity":"afc05f36-6b13-464c-84b1-7ca5543381b2","added_by":"auto","created_at":"2025-08-27 04:03:51","extension":"csv","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":35032,"visible":true,"origin":"","legend":"a subset of dataset","description":"","filename":"datasetsubset.csv","url":"https://assets-eu.researchsquare.com/files/rs-6347897/v1/d8eb238fb750f22d543b254c.csv"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"MIMIC-IV-Ext-22MCTS: A 22 Millions-Event Temporal Clinical Time-Series Dataset with Relative Timestamp for Risk Prediction","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"Clinical event, Temporal information, Time series, MIMIC, Contextual BM25, Contextual semantic search, Natural language processing, Large language model, Question answering, Clinical trial","lastPublishedDoi":"10.21203/rs.3.rs-6347897/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6347897/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eClinical risk prediction based on machine learning algorithms plays a vital role in modern healthcare. A crucial component in developing a reliable prediction model is collecting high-quality time series clinical events. In this work, we release such a dataset that consists of 22,588,586 Clinical Time Series events, which we term MIMIC-IV-Ext-22MCTS. Our source data are discharge sum- maries selected from the well-known yet unstructured MIMIC-IV-Note [1]. We then extract clinical events as short text span from the discharge summaries, along with the timestamps of these events as temporal information. The general- purpose MIMIC-IV-Note pose specific challenges for our work: it turns out that the discharge summaries are too lengthy for typical natural language models to process, and the clinical events of interest often are not accompanied with explicit timestamps. Therefore, we propose a new framework that works as follows: 1) we break each discharge summary into manageably small text chunks; 2) we apply contextual BM25 and contextual semantic search to retrieve chunks that have a high potential of containing clinical events; and 3) we carefully design prompts to teach the recently released Llama-3.1-8B [4] model to identify or infer temporal information of the chunks. We show that the obtained dataset is so informative and transparent that standard models fine-tuned on our dataset are achieving sig- nificant improvements in healthcare applications. In particular, the BERT model fine-tuned based on our dataset achieves 10% improvement in accuracy on medi- cal question answering task, and 3% improvement in clinical trial matching task compared with the classic BERT. The GPT-2 model, fine-tuned on our dataset, produces more clinically reliable results for clinical questions.\u003c/p\u003e","manuscriptTitle":"MIMIC-IV-Ext-22MCTS: A 22 Millions-Event Temporal Clinical Time-Series Dataset with Relative Timestamp for Risk Prediction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-27 03:55:47","doi":"10.21203/rs.3.rs-6347897/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":"2f52eace-124c-4ac0-8ffd-d94c0ca14ee9","owner":[],"postedDate":"August 27th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":48004919,"name":"Health sciences/Diseases"},{"id":48004920,"name":"Health sciences/Health care/Public health"}],"tags":[],"updatedAt":"2025-09-15T19:25:37+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-27 03:55:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6347897","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6347897","identity":"rs-6347897","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}