{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,28]],"date-time":"2026-06-28T11:55:13Z","timestamp":1782647713254,"version":"3.54.5"},"reference-count":33,"publisher":"Oxford University Press (OUP)","issue":"10","license":[{"start":{"date-parts":[[2021,8,1]],"date-time":"2021-08-01T00:00:00Z","timestamp":1627776000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"funder":[{"DOI":"10.13039\/100000092","name":"National Library of Medicine","doi-asserted-by":"publisher","award":["R15LM013209"],"award-info":[{"award-number":["R15LM013209"]}],"id":[{"id":"10.13039\/100000092","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2021,9,18]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>Clinical notes contain an abundance of important, but not-readily accessible, information about patients. Systems that automatically extract this information rely on large amounts of training data of which there exists limited resources to create. Furthermore, they are developed disjointly, meaning that no information can be shared among task-specific systems. This bottleneck unnecessarily complicates practical application, reduces the performance capabilities of each individual solution, and associates the engineering debt of managing multiple information extraction systems.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Materials and Methods<\/jats:title>\n                  <jats:p>We address these challenges by developing Multitask-Clinical BERT: a single deep learning model that simultaneously performs 8 clinical tasks spanning entity extraction, personal health information identification, language entailment, and similarity by sharing representations among tasks.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>We compare the performance of our multitasking information extraction system to state-of-the-art BERT sequential fine-tuning baselines. We observe a slight but consistent performance degradation in MT-Clinical BERT relative to sequential fine-tuning.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Discussion<\/jats:title>\n                  <jats:p>These results intuitively suggest that learning a general clinical text representation capable of supporting multiple tasks has the downside of losing the ability to exploit dataset or clinical note-specific properties when compared to a single, task-specific model.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>We find our single system performs competitively with all state-the-art task-specific systems while also benefiting from massive computational benefits at inference.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/jamia\/ocab126","type":"journal-article","created":{"date-parts":[[2021,6,4]],"date-time":"2021-06-04T19:19:39Z","timestamp":1622834379000},"page":"2108-2115","source":"Crossref","is-referenced-by-count":47,"title":["MT-clinical BERT: scaling clinical information extraction with multitask learning"],"prefix":"10.1093","volume":"28","author":[{"given":"Andriy","family":"Mulyar","sequence":"first","affiliation":[{"name":"Computer Science Department, Virginia Commonwealth University, Richmond, Virginia, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8011-9850","authenticated-orcid":false,"given":"Ozlem","family":"Uzuner","sequence":"additional","affiliation":[{"name":"Information Sciences and Technology, George Mason University, Fairfax, Virginia, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bridget","family":"McInnes","sequence":"additional","affiliation":[{"name":"Information Sciences and Technology, George Mason University, Fairfax, Virginia, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2021,8,1]]},"reference":[{"issue":"5","key":"2021091817524117300_ocab126-B1","doi-asserted-by":"crossref","first-page":"552","DOI":"10.1136\/amiajnl-2011-000203","article-title":"2010 i2b2\/va challenge on concepts, assertions, and relations in clinical text","volume":"18","author":"Uzuner","year":"2011","journal-title":"J Am Med Inform Assoc"},{"issue":"5","key":"2021091817524117300_ocab126-B2","doi-asserted-by":"crossref","first-page":"806","DOI":"10.1136\/amiajnl-2013-001628","article-title":"Evaluating temporal relations in clinical text: 2012 i2b2 challenge","volume":"20","author":"Sun","year":"2013","journal-title":"J Am Med Inform Assoc"},{"issue":"1","key":"2021091817524117300_ocab126-B3","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1093\/jamia\/ocz166","article-title":"2018 n2c2 shared task on adverse drug events and medication extraction in electronic health records","volume":"27","author":"Henry","year":"2020","journal-title":"J Am Med Inform Assoc"},{"key":"2021091817524117300_ocab126-B4","doi-asserted-by":"crossref","first-page":"S11","DOI":"10.1016\/j.jbi.2015.06.007","article-title":"Automated systems for the de-identification of longitudinal clinical narratives: Overview of 2014 i2b2\/uthealth shared task track 1","volume":"58","author":"Stubbs","year":"2015","journal-title":"J Biomed Inform"},{"key":"2021091817524117300_ocab126-B5","volume-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Devlin"},{"key":"2021091817524117300_ocab126-B6","first-page":"5753","author":"Yang","year":"2019"},{"key":"2021091817524117300_ocab126-B7","doi-asserted-by":"crossref","first-page":"160035","DOI":"10.1038\/sdata.2016.35","article-title":"Mimic-iii, a freely accessible critical care database","volume":"3","author":"Johnson","year":"2016","journal-title":"Scientific Data"},{"key":"2021091817524117300_ocab126-B8","volume-title":"Proceedings of the 2nd Clinical Natural Language Processing Workshop","author":"Alsentzer","year":"2019"},{"issue":"1","key":"2021091817524117300_ocab126-B9","doi-asserted-by":"crossref","first-page":"41","DOI":"10.1023\/A:1007379606734","article-title":"Multitask learning","volume":"28","author":"Caruana","year":"1997","journal-title":"Mach Learn"},{"key":"2021091817524117300_ocab126-B10","doi-asserted-by":"crossref","first-page":"120","DOI":"10.1016\/j.patrec.2020.05.031","article-title":"Multi-task learning for natural language processing in the 2020s: where are we going?","volume":"136","author":"Worsham","year":"2020","journal-title":"Pattern Recogn Lett"},{"key":"2021091817524117300_ocab126-B11","article-title":"A survey on multi-task learning","author":"Zhang","year":"2017","journal-title":"arXiv preprint arXiv:1707.08114"},{"key":"2021091817524117300_ocab126-B12","article-title":"Massively multitask networks for drug discovery","author":"Ramsundar","year":"2015","journal-title":"arXiv preprint arXiv:1502.02072"},{"key":"2021091817524117300_ocab126-B13","author":"Yan","year":"2019"},{"key":"2021091817524117300_ocab126-B14","volume-title":"Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics","author":"Liu","year":"28\u2013 2, 2019;"},{"key":"2021091817524117300_ocab126-B15","author":"Raffel"},{"issue":"1","key":"2021091817524117300_ocab126-B16","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s12859-017-1776-8","article-title":"A neural network multi-task learning approach to biomedical named entity recognition","volume":"18","author":"Crichton","year":"2017","journal-title":"BMC Bioinform"},{"issue":"10","key":"2021091817524117300_ocab126-B17","doi-asserted-by":"crossref","first-page":"1745","DOI":"10.1093\/bioinformatics\/bty869","article-title":"Cross-type biomedical named entity recognition with deep multi-task learning","volume":"35","author":"Wang","year":"2019","journal-title":"Bioinformatics"},{"key":"2021091817524117300_ocab126-B18","article-title":"Mt-bioner: Multi-task learning for biomedical named entity recognition using deep bidirectional transformers","author":"Khan","year":"2020","journal-title":"arXiv preprint arXiv:2001.08904"},{"key":"2021091817524117300_ocab126-B19","author":"Mehmood"},{"key":"2021091817524117300_ocab126-B20","article-title":"Analyzing the effect of multi-task learning for biomedical named entity recognition","author":"Akdemir","year":"2020","journal-title":"arXiv Preprint arXiv:2011.00425"},{"issue":"1","key":"2021091817524117300_ocab126-B21","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s12859-016-1414-x","article-title":"A neural joint model for entity and relation extraction from biomedical text+","volume":"18","author":"Li","year":"2017","journal-title":"BMC Bioinform"},{"issue":"S10","key":"2021091817524117300_ocab126-B22","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s12911-019-0995-5","article-title":"Family history information extraction via deep joint learning","volume":"19","author":"Shi","year":"2019","journal-title":"BMC Med Inform Decis Mak"},{"key":"2021091817524117300_ocab126-B23","article-title":"An empirical study of multi-task learning on BERT for biomedical text mining","author":"Peng","year":"2020","journal-title":"arXiv:2005.02799 [cs]"},{"key":"2021091817524117300_ocab126-B24","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1007\/s00521-018-3699-3","article-title":"Empirical evaluation of multi-task learning in deep neural networks for natural language processing","author":"Li","year":"2020","journal-title":"Neural Comput Appl"},{"key":"2021091817524117300_ocab126-B25","author":"Collobert","year":"5\u20139, 2008;  ,"},{"key":"2021091817524117300_ocab126-B26"},{"key":"2021091817524117300_ocab126-B27","article-title":"Lessons from natural language inference in the clinical domain","author":"Romanov","year":"2018","journal-title":"arXiv preprint arXiv:1808.06752"},{"key":"2021091817524117300_ocab126-B28","first-page":"310","volume-title":"AMIA Annu Symp Proc","author":"Ben Abacha","year":"2016"},{"key":"2021091817524117300_ocab126-B29","author":"N\u00e9v\u00e9ol","year":"31, 2014; ,"},{"issue":"01","key":"2021091817524117300_ocab126-B30","first-page":"216","article-title":"Aggregating UMLs semantic types for reducing conceptual complexity","volume":"84","author":"McCray","year":"2001","journal-title":"Stud Health Technol Inform"},{"key":"2021091817524117300_ocab126-B31","author":"Ruder","year":"2019"},{"key":"2021091817524117300_ocab126-B32","author":"Dodge","year":"2020"},{"key":"2021091817524117300_ocab126-B33","author":"Giorgi","year":"2019"}],"container-title":["Journal of the American Medical Informatics Association"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/academic.oup.com\/jamia\/article-pdf\/28\/10\/2108\/40408880\/ocab126.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"http:\/\/academic.oup.com\/jamia\/article-pdf\/28\/10\/2108\/40408880\/ocab126.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,9,18]],"date-time":"2021-09-18T17:53:56Z","timestamp":1631987636000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jamia\/article\/28\/10\/2108\/6333354"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,8,1]]},"references-count":33,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2021,8,1]]},"published-print":{"date-parts":[[2021,9,18]]}},"URL":"https:\/\/doi.org\/10.1093\/jamia\/ocab126","relation":{},"ISSN":["1527-974X"],"issn-type":[{"value":"1527-974X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2021,10,1]]},"published":{"date-parts":[[2021,8,1]]}}}