{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,7,8]],"date-time":"2026-07-08T01:45:18Z","timestamp":1783475118602,"version":"3.55.0"},"posted":{"date-parts":[[2025,3,20]]},"group-title":"Health Informatics","reference-count":32,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,3,20]],"date-time":"2025-03-20T00:00:00Z","timestamp":1742428800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.medrxiv.org\/about\/FAQ#license"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,9,20]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>The temporal sequence of clinical events is crucial in outcomes research, yet standard machine learning (ML) approaches often overlook this aspect in electronic health records (EHRs), limiting predictive accuracy.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We introduce Temporal Learning with Dynamic Range (TLDR), a time-sensitive ML framework, to identify risk factors for post-acute sequelae of SARS-CoV-2 infection (PASC). Using longitudinal EHR data from over 85,000 patients in the Precision PASC Research Cohort (P2RC) from a large integrated academic medical center, we compare TLDR against a conventional atemporal ML model.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>TLDR demonstrated superior predictive performance, achieving a mean AUROC of 0.791 compared to 0.668 for the benchmark, marking an 18.4% improvement. Additionally, TLDR\u2019s mean PRAUC of 0.590 significantly outperformed the benchmark\u2019s 0.421, a 40.14% increase. The framework exhibited improved generalizability with a lower mean overfitting index (\u22120.028), highlighting its robustness. Beyond predictive gains, TLDR\u2019s use of time-stamped features enhanced interpretability, offering a more precise characterization of individual patient records.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Discussion<\/jats:title>\n                  <jats:p>TLDR effectively captures exposure\u2013outcome associations and offers flexibility in time-stamping strategies to suit diverse clinical research needs.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>TLDR provides a simple yet effective approach for integrating dynamic temporal windows into predictive modeling. It is available within the MLHO R package to support further exploration of recurrent treatment and exposure patterns in various clinical settings.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2025.03.19.25324272","type":"posted-content","created":{"date-parts":[[2025,3,20]],"date-time":"2025-03-20T15:35:24Z","timestamp":1742484924000},"source":"Crossref","is-referenced-by-count":1,"title":["Temporal Learning with Dynamic Range (TLDR) for Modeling Recurrent Exposure and Treatment Outcomes"],"prefix":"10.64898","author":[{"given":"Jingya","family":"Cheng","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4183-1287","authenticated-orcid":false,"given":"Jonas","family":"H\u00fcgel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiazi","family":"Tian","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alaleh","family":"Azhir","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shawn N.","family":"Murphy","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jeffrey G.","family":"Klann","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0204-8978","authenticated-orcid":false,"given":"Hossein","family":"Estiri","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2025092400250928000_2025.03.19.25324272v2.1","unstructured":"Vaswani A , Shazeer N , Parmar N , et al. Attention Is All You Need. NeurIPS. 2017."},{"key":"2025092400250928000_2025.03.19.25324272v2.2","doi-asserted-by":"publisher","DOI":"10.1213\/ANE.0000000000003511"},{"key":"2025092400250928000_2025.03.19.25324272v2.3","doi-asserted-by":"publisher","DOI":"10.1002\/(SICI)1097-0258(19990715)18:13<1707::AID-SIM138>3.0.CO;2-H"},{"key":"2025092400250928000_2025.03.19.25324272v2.4","doi-asserted-by":"publisher","DOI":"10.1186\/s12874-018-0499-5"},{"key":"2025092400250928000_2025.03.19.25324272v2.5","doi-asserted-by":"crossref","first-page":"1","DOI":"10.29115\/SP-2013-0023","article-title":"Measuring compliance in mobile longitudinal repeated-measures design study","volume":"6","year":"2013","journal-title":"Survey Practice"},{"key":"2025092400250928000_2025.03.19.25324272v2.6","doi-asserted-by":"crossref","first-page":"632","DOI":"10.1002\/sim.9635","article-title":"Modeling longitudinal change in biomarkers using data from a complex survey: HCHS\/SOL","volume":"42","year":"2023","journal-title":"Stat Med"},{"key":"2025092400250928000_2025.03.19.25324272v2.7","doi-asserted-by":"crossref","first-page":"772","DOI":"10.1093\/jamia\/ocaa288","article-title":"High-throughput phenotyping with temporal sequences","volume":"28","year":"2021","journal-title":"J Am Med Inform Assoc"},{"key":"2025092400250928000_2025.03.19.25324272v2.8","doi-asserted-by":"crossref","first-page":"104629","DOI":"10.1016\/j.ebiom.2023.104629","article-title":"Temporal characterization of Alzheimer\u2019s disease with sequences of clinical records","volume":"92","year":"2023","journal-title":"EBioMedicine"},{"key":"2025092400250928000_2025.03.19.25324272v2.9","doi-asserted-by":"publisher","DOI":"10.1016\/j.jbi.2021.103980"},{"key":"2025092400250928000_2025.03.19.25324272v2.10","doi-asserted-by":"crossref","first-page":"313","DOI":"10.1007\/s41666-023-00143-4","article-title":"Assessment of prediction tasks and time window selection in temporal modeling of EHR data: systematic review","volume":"7","year":"2023","journal-title":"J Healthc Inform Res"},{"key":"2025092400250928000_2025.03.19.25324272v2.11","doi-asserted-by":"publisher","DOI":"10.1093\/jamia\/ocac216"},{"key":"2025092400250928000_2025.03.19.25324272v2.12","doi-asserted-by":"crossref","first-page":"100051","DOI":"10.1016\/j.patter.2020.100051","article-title":"Transitive sequencing medical records for mining predictive and inter-pretable temporal representations","volume":"1","year":"2020","journal-title":"Patterns"},{"key":"2025092400250928000_2025.03.19.25324272v2.13","doi-asserted-by":"crossref","unstructured":"Estiri H , Vasey S , Murphy SN . Transitive sequential pattern mining for discrete clinical data. In: Artificial Intelligence in Medicine. Springer; 2020:414\u2013424.","DOI":"10.1007\/978-3-030-59137-3_37"},{"key":"2025092400250928000_2025.03.19.25324272v2.14","first-page":"14604582221101526","article-title":"Trajectory mining in kidney cancer","volume":"28","year":"2022","journal-title":"Health Informatics J"},{"key":"2025092400250928000_2025.03.19.25324272v2.15","first-page":"1642","article-title":"Temporal characterization and visualization of revolving therapy-events in lung cancer patients","volume":"316","year":"2024","journal-title":"Stud Health Technol Inform"},{"key":"2025092400250928000_2025.03.19.25324272v2.16","doi-asserted-by":"crossref","first-page":"102576","DOI":"10.1016\/j.artmed.2023.102576","article-title":"Frequent temporal patterns of biomarkers in sepsis","volume":"143","year":"2023","journal-title":"Artif Intell Med"},{"key":"2025092400250928000_2025.03.19.25324272v2.17","doi-asserted-by":"crossref","first-page":"107156","DOI":"10.1016\/j.compbiomed.2023.107156","article-title":"Early prediction of late-onset sepsis with ML","volume":"163","year":"2023","journal-title":"Comput Biol Med"},{"key":"2025092400250928000_2025.03.19.25324272v2.18","doi-asserted-by":"crossref","first-page":"102210","DOI":"10.1016\/j.eclinm.2023.102210","article-title":"Long COVID temporal sub-phenotypes via distributed representation learning","volume":"64","year":"2023","journal-title":"eClinicalMedicine"},{"key":"2025092400250928000_2025.03.19.25324272v2.19","doi-asserted-by":"crossref","first-page":"5322","DOI":"10.1038\/s41598-021-84781-x","article-title":"Individualized prediction of COVID-19 adverse outcomes with MLHO","volume":"11","year":"2021","journal-title":"Sci Rep"},{"key":"2025092400250928000_2025.03.19.25324272v2.20","doi-asserted-by":"crossref","unstructured":"Azhir A , H\u00fcgel J , Tian J , et al. Precision phenotyping for PASC cohort curation. Med (N Y). 2024;100532.","DOI":"10.1016\/j.medj.2024.10.009"},{"key":"2025092400250928000_2025.03.19.25324272v2.21","unstructured":"Cheng J , Tian J , H\u00fcgel J , Estiri H. MLHO. GitHub, 2021."},{"key":"2025092400250928000_2025.03.19.25324272v2.22","doi-asserted-by":"crossref","first-page":"8520","DOI":"10.1016\/j.eswa.2015.07.007","article-title":"Feature selection using Joint Mutual Information Maximisation","volume":"42","year":"2015","journal-title":"Expert Syst Appl"},{"key":"2025092400250928000_2025.03.19.25324272v2.23","doi-asserted-by":"crossref","first-page":"107","DOI":"10.1016\/j.ijar.2021.01.003","article-title":"Feature selection with fuzzy joint mutual information","volume":"132","year":"2021","journal-title":"Int J Approx Reason"},{"key":"2025092400250928000_2025.03.19.25324272v2.24","unstructured":"Yang H , Moody J. Feature selection based on joint mutual information. ICSC Symposium. 1999:22\u201325."},{"key":"2025092400250928000_2025.03.19.25324272v2.25","year":"2019","journal-title":"Publicly available clinical BERT embeddings"},{"key":"2025092400250928000_2025.03.19.25324272v2.26","unstructured":"Clark K , Luong M-T , Le QV , et al. ELECTRA: Pre-training text encoders as discriminators rather than generators. ICLR. 2020."},{"key":"2025092400250928000_2025.03.19.25324272v2.27","doi-asserted-by":"crossref","first-page":"86","DOI":"10.1038\/s41746-021-00455-y","article-title":"Med-BERT: pretrained contextualized embeddings on large-scale structured electronic health records for disease prediction","volume":"4","year":"2021","journal-title":"NPJ Digit Med"},{"key":"2025092400250928000_2025.03.19.25324272v2.28","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-30657-1"},{"key":"2025092400250928000_2025.03.19.25324272v2.29","first-page":"e281","article-title":"Foresight: GPT for modeling patient timelines from EHR","volume":"6","year":"2024","journal-title":"Lancet Digit Health"},{"key":"2025092400250928000_2025.03.19.25324272v2.30","doi-asserted-by":"crossref","first-page":"105442","DOI":"10.1016\/j.ebiom.2024.105442","article-title":"Transformer model for diagnosis of suspected lung cancer in primary care EHR","volume":"110","year":"2024","journal-title":"EBioMedicine"},{"issue":"4","key":"2025092400250928000_2025.03.19.25324272v2.31","first-page":"303","article-title":"Computer-based consultations in clinical therapeutics: explanation and rule acquisition capabilities of the MYCIN system","volume":"7","year":"1974","journal-title":"Comput Biomed Res"},{"key":"2025092400250928000_2025.03.19.25324272v2.32","doi-asserted-by":"crossref","unstructured":"Berner ES , La Lande TJ . Overview of clinical decision support systems. In: Berner ES , ed. Clinical Decision Support Systems: Theory and Practice. 2nd ed. New York: Springer; 2007:3\u201322.","DOI":"10.1007\/978-0-387-38319-4_1"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.03.19.25324272","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:31:29Z","timestamp":1768509089000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.03.19.25324272"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,3,20]]},"references-count":32,"URL":"https:\/\/doi.org\/10.1101\/2025.03.19.25324272","relation":{},"subject":[],"published":{"date-parts":[[2025,3,20]]},"subtype":"preprint"}}