{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,22]],"date-time":"2026-09-22T03:14:34Z","timestamp":1790046874031,"version":"4.0.1"},"reference-count":29,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec sec-type=\"background\">\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Heart failure with preserved ejection fraction (HFpEF) is a major clinical manifestation of cardiac amyloidosis, a condition frequently underdiagnosed due to its nonspecific symptomatology. Electronic health records (EHRs) offer a promising avenue for supporting early symptom detection through natural language processing. However, identifying relevant clinical cues within unstructured narratives, particularly in Spanish, remains a significant challenge due to the scarcity of annotated corpora and domain-specific models. This study proposes and evaluates a Transformer-based natural language processing framework for automated detection of HFpEF-related symptoms in Spanish EHRs.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"objective\">\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>The aim of this study is to assess the feasibility of leveraging unstructured clinical narratives to support early identification of heart failure phenotypes indicative of cardiac amyloidosis. It also examines how domain-specific language models and clinically guided optimization strategies can improve the reliability, sensitivity, and generalizability of symptom detection in real-world EHRs.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"methods\">\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      A novel corpus of 15,304 Spanish clinical documents was manually annotated and validated by cardiology experts. The corpus was derived from the records of 262 patients (173 with suspected cardiac amyloidosis and 89 without). In total, 8 Transformer-based language models were evaluated, including general-purpose models, biomedical-specialized variants, and Longformers. Three clinically motivated optimization strategies were implemented to align models\u2019 behavior with different diagnostic priorities: maximizing area under the curve (AUC) to enhance overall discrimination, optimizing\n                      <jats:italic>F<\/jats:italic>\n                      <jats:sub>1<\/jats:sub>\n                      -score to balance sensitivity and precision, and prioritizing sensitivity to minimize false negatives. These strategies were independently applied during the fine-tuning of the models to assess their impact on performance under different clinical constraints. To ensure robust evaluation, testing was conducted on a dataset composed exclusively of previously unseen patients, allowing performance to be assessed under realistic and generalizable conditions.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"results\">\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      All models achieved high performance, with AUC values above 0.940. The best-performing model,\n                      <jats:italic>Longformer Biomedical-clinical<\/jats:italic>\n                      , reached an AUC of 0.987,\n                      <jats:italic>F<\/jats:italic>\n                      <jats:sub>1<\/jats:sub>\n                      -score of 0.985, sensitivity of 0.987, and specificity of 0.987 on the test dataset. Models optimized for sensitivity reduced the false-negative rate to under 3%, a key threshold for clinical safety. Comparative analyses confirmed that domain-adapted, long-sequence models are better suited for the semantic and structural complexity of Spanish clinical texts than general-purpose models.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"conclusions\">\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Transformer-based models can reliably detect HFpEF-related symptoms from Spanish EHRs, even in the presence of class imbalance and substantial linguistic complexity. The results show that combining domain-specific pretraining with long-context modeling architectures and clinically aligned optimization strategies leads to substantial gains in classification performance, particularly in sensitivity. These models not only achieve high accuracy and generalization on unseen patients but also demonstrate robustness in handling the semantic nuances and narrative structure of real-world clinical documentation. These findings support the potential deployment of Transformer-based systems as effective screening tools to prioritize patients at risk for cardiac amyloidosis in Spanish-speaking health care settings.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.2196\/76433","type":"journal-article","created":{"date-parts":[[2025,5,26]],"date-time":"2025-05-26T04:00:08Z","timestamp":1748232008000},"page":"e76433-e76433","source":"Crossref","is-referenced-by-count":4,"title":["Multicriteria Optimization of Language Models for Heart Failure With Preserved Ejection Fraction Symptom Detection in Spanish Electronic Health Records: Comparative Modeling Study"],"prefix":"10.2196","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5329-9622","authenticated-orcid":false,"given":"Jacinto","family":"Mata","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0697-4044","authenticated-orcid":false,"given":"Victoria","family":"Pach\u00f3n","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7015-170X","authenticated-orcid":false,"given":"Ana","family":"Manovel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7551-2401","authenticated-orcid":false,"given":"Manuel J","family":"Ma\u00f1a","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3464-2944","authenticated-orcid":false,"given":"Manuel","family":"de la Villa","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2025,7,17]]},"reference":[{"issue":"18","key":"R1","doi-asserted-by":"publisher","first-page":"1810","DOI":"10.1016\/j.jacc.2023.01.049","article-title":"Heart failure with preserved ejection fraction: JACC scientific statement","volume":"81","author":"Borlaug","journal-title":"J Am Coll Cardiol"},{"issue":"6","key":"R2","doi-asserted-by":"publisher","first-page":"1128","DOI":"10.1002\/ehf2.12518","article-title":"Transthyretin cardiac amyloidosis: an update on diagnosis and treatment","volume":"6","author":"Yamamoto","journal-title":"ESC Heart Fail"},{"issue":"13","key":"R3","doi-asserted-by":"publisher","first-page":"1201","DOI":"10.1056\/NEJMra2302038","article-title":"Artificial intelligence and machine learning in clinical medicine, 2023","volume":"388","author":"Haug","journal-title":"N Engl J Med"},{"issue":"12","key":"R4","doi-asserted-by":"publisher","first-page":"909","DOI":"10.1136\/heartjnl-2021-319769","article-title":"Systematic review of current natural language processing methods and applications in cardiology","volume":"108","author":"Reading Turchioe","journal-title":"Heart"},{"issue":"9","key":"R5","doi-asserted-by":"publisher","first-page":"2137","DOI":"10.1093\/jamia\/ocae054","article-title":"A comparative analysis of Spanish Clinical encoder-based models on NER and classification tasks","volume":"31","author":"Garc\u00eda Subies","journal-title":"J Am Med Inform Assoc"},{"key":"R6","unstructured":"Miranda-Escalada A Farr\u00e9-Maduell E Krallinger M . NUBes: a Spanish corpus of negation and uncertainty in the biomedical domain. Presented at: Twelfth Language Resources and Evaluation Conference; Marseille, France. May 2020.URL: https:\/\/aclanthology.org\/2020.lrec-1.708 [Accessed 12-07-2025]"},{"key":"R7","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2018.09.034","article-title":"Clinical text classification research trends: systematic literature review and open issues","volume":"116","author":"Mujtaba","journal-title":"Expert Syst Appl"},{"issue":"11","key":"R8","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2024.43925","article-title":"Natural language processing of clinical documentation to assess functional status in patients with heart failure","volume":"7","author":"Adejumo","journal-title":"JAMA Netw Open"},{"key":"R9","doi-asserted-by":"crossref","unstructured":"Liu Y Tan Z Zhang Z et al. A natural language processing-based approach for early detection of heart failure onset using electronic health records. medRxiv. Preprint posted online on Apr 6, 2025. [doi: 10.1101\/2025.04.04.25325211]","DOI":"10.1016\/j.knosys.2025.114102"},{"issue":"1","key":"R10","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCHEARTFAILURE.124.012514","article-title":"Natural language processing to adjudicate heart failure hospitalizations in global clinical trials","volume":"18","author":"Marti-Castellote","journal-title":"Circ Heart Fail"},{"issue":"2","key":"R11","doi-asserted-by":"publisher","first-page":"287","DOI":"10.1016\/j.hfc.2021.12.002","article-title":"Advances in machine learning approaches to heart failure with preserved ejection fraction","volume":"18","author":"Ahmad","journal-title":"Heart Fail Clin"},{"issue":"1","key":"R12","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-34294-6","article-title":"Heart disease risk factors detection from electronic health records using advanced NLP and deep learning techniques","volume":"13","author":"Houssein","journal-title":"Sci Rep"},{"issue":"1","key":"R13","doi-asserted-by":"publisher","DOI":"10.1186\/s40537-024-00903-y","article-title":"Adapting transformer-based language models for heart disease detection and risk factors extraction","volume":"11","author":"Houssein","journal-title":"J Big Data"},{"key":"R14","doi-asserted-by":"publisher","DOI":"10.1016\/j.ebiom.2025.105657","article-title":"Identification of heart failure subtypes using transformer-based deep learning modelling: a population-based study of 379,108 individuals","volume":"114","author":"Fan","journal-title":"EBioMedicine"},{"issue":"1","key":"R15","doi-asserted-by":"publisher","first-page":"146","DOI":"10.1093\/bib\/bbz130","article-title":"An extensive review of tools for manual annotation of documents","volume":"22","author":"Neves","journal-title":"Brief Bioinform"},{"key":"R16","unstructured":"Prodigy: an annotation tool for AI, machine learning & NLP. URL: https:\/\/prodi.gy\/ [Accessed 12-07-2025]"},{"key":"R17","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2024.3349952","article-title":"A survey of text classification with transformers: how wide? How large? How long? How accurate? How expensive? How safe?","volume":"12","author":"Fields","journal-title":"IEEE Access"},{"key":"R18","doi-asserted-by":"publisher","DOI":"10.1145\/3439726","article-title":"Deep learning-based text classification: a comprehensive review","author":"Minaee","journal-title":"ACM Comput Surv"},{"issue":"2","key":"R19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3605943","article-title":"Recent advances in natural language processing via large pre-trained language models: a survey","volume":"56","author":"Min","journal-title":"ACM Comput Surv"},{"key":"R20","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2024.102900","article-title":"Transformers and large language models in healthcare: a review","volume":"154","author":"Nerella","journal-title":"Artif Intell Med"},{"key":"R21","unstructured":"Beltagy I Peters ME Cohan A . Longformer: the long-document transformer. arXiv. Preprint posted online on Apr 10, 2020. [doi: 10.48550\/arXiv.2004.05150]"},{"key":"R22","unstructured":"Ca\u00f1ete J Chaperon G Fuentes R Ho JH Kang H P\u00e9rez J . Spanish pre-trained BERT model and evaluation data. Preprint posted online on Aug 6, 2020. [doi: 10.48550\/arXiv.2308.02976]"},{"key":"R23","unstructured":"Liu Y Ott M Goyal N Du J . RoBERTa: a robustly optimized BERT pretraining approach. arXiv. Preprint posted online on Jul 26, 2019. [doi: 10.48550\/arXiv.1907.11692]"},{"key":"R24","unstructured":"Guti\u00e9rrez-Fandi\u00f1o A Armengol-Estap\u00e9 J P\u00e0mies M Llop-Palao J Silveira-Ocampo J et al. MarIA: Spanish language models. arXiv. Preprint posted online on Jul 15, 2022. [doi: 10.26342\/2022-68-3]"},{"key":"R25","unstructured":"Carrino CP Armengol-Estap\u00e9 J Guti\u00e9rrez-Fandi\u00f1o A Llop-Palao J P\u00e0mies M et al. Biomedical and clinical language models for Spanish: on the benefits of domain-specific pretraining in a mid-resource scenario. arXiv. Preprint posted online on Sep 8, 2021. [doi: 10.48550\/arXiv.2109.03570]"},{"key":"R26","doi-asserted-by":"crossref","unstructured":"Carrino CP Llop J P\u00e0mies M et al. Pretrained biomedical language models for clinical NLP in Spanish. Presented at: Proceedings of the 21st Workshop on Biomedical Language Processing; 193-199; Dublin, Ireland. May 26, 2022.[doi: 10.18653\/v1\/2022.bionlp-1.19]","DOI":"10.18653\/v1\/2022.bionlp-1.19"},{"issue":"1","key":"R27","doi-asserted-by":"publisher","DOI":"10.1186\/s12874-022-01665-y","article-title":"A comparative study on deep learning models for text classification of unstructured medical notes with various levels of class imbalance","volume":"22","author":"Lu","journal-title":"BMC Med Res Methodol"},{"issue":"4","key":"R28","doi-asserted-by":"publisher","first-page":"195","DOI":"10.4103\/tjem.tjem_182_23","article-title":"Receiver operating characteristic curve analysis in diagnostic accuracy studies: a guide to interpreting the area under the curve value","volume":"23","author":"\u00c7orbac\u0131o\u011flu","journal-title":"Turk J Emerg Med"},{"issue":"10","key":"R29","doi-asserted-by":"publisher","first-page":"2081","DOI":"10.1016\/j.arth.2022.10.031","article-title":"External validation of natural language processing algorithms to extract common data elements in THA operative notes","volume":"38","author":"Wyles","journal-title":"J Arthroplasty"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jmir.org\/2025\/1\/e76433","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T18:07:57Z","timestamp":1782842877000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2025\/1\/e76433"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,17]]},"references-count":29,"URL":"https:\/\/doi.org\/10.2196\/76433","relation":{},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,17]]},"article-number":"v27i18e76433"}}