{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:52:05Z","timestamp":1791525125625,"version":"4.3.4"},"reference-count":23,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T00:00:00Z","timestamp":1791504000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Clinical Foundations Models (FMs) hold the potential to exhibit generalist medical AI capabilities, and the ability to reduce the need for explicit feature engineering. However, their evaluation often relies on locally-limited datasets, raising questions regarding their generalizability across diverse global healthcare systems, populations, and clinical practices. This study assesses the cross-continental transferability of a US-trained FM, CLMBR-T-base, applied to a large, distinct, Italian hospital cohort to predict Length of Stay (LoS).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We utilized a retrospective EHR dataset from IRCCS Maugeri, Pavia, Italy, comprising 99,884 admissions for 63,030 unique patients (2012\u20132025). To ensure interoperability, raw data were harmonized using the OMOP Common Data Model and converted to the Medical Event Data Standard (MEDS). Patient embeddings were generated zero-shot using the pretrained CLMBR-T-Base model. We benchmarked supervised regression models (Gradient Boosting, Multilayer Perceptron) trained on these embeddings against traditional count-based baselines to predict log-transformed LoS. We built different models trained on the complete hospital admissions set and stratified across four clinical macro-categories (Surgical, Rehabilitation, Oncology, Generic). Embedding cohesion was validated using latent space analysis (HDBSCAN clustering).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The best overall model utilizing FM embeddings achieved an R\n                      <jats:sup>2<\/jats:sup>\n                      of 0.705. Performance varied by department type, proving lower in Rehabilitation (R\n                      <jats:sup>2<\/jats:sup>\n                      \u2009=\u20090.36 for the best count-based model and R\n                      <jats:sup>2<\/jats:sup>\n                      \u2009=\u20090.33 for the best embedding-based model) and Oncology (R\n                      <jats:sup>2<\/jats:sup>\n                      \u2009=\u20090.17 for the best count-based model and R\n                      <jats:sup>2<\/jats:sup>\n                      \u2009=\u20090.16 for the best embedding-based model). Latent space analysis confirmed that the FM embeddings preserved high semantic transferability, autonomously organizing Italian patients into clinically coherent groups.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Clinical FMs trained on US data exhibit generalizability, effectively capturing clinical semantics across distinct international healthcare architectures when mediated by rigorous data standardization pipelines (OMOP\/MEDS).<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fdgth.2026.1956993","type":"journal-article","created":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:34:44Z","timestamp":1791524084000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Cross-continental generalizability of EHR foundation model: predicting length of stay in an Italian hospital using US-trained embeddings"],"prefix":"10.3389","volume":"8","author":[{"given":"Giovanna","family":"Nicora","sequence":"first","affiliation":[{"name":"Department of Electrical, Computer and Biomedical Engineering, University of Pavia","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Francesca","family":"Negri","sequence":"additional","affiliation":[{"name":"Department of Electrical, Computer and Biomedical Engineering, University of Pavia","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Valentina","family":"Tibollo","sequence":"additional","affiliation":[{"name":"IRCCS ICS Maugeri Hospital","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alberto","family":"Malovini","sequence":"additional","affiliation":[{"name":"IRCCS ICS Maugeri Hospital","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Arianna","family":"Dagliati","sequence":"additional","affiliation":[{"name":"Department of Electrical, Computer and Biomedical Engineering, University of Pavia","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lucia","family":"Sacchi","sequence":"additional","affiliation":[{"name":"Department of Electrical, Computer and Biomedical Engineering, University of Pavia","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Federico","family":"Sottotetti","sequence":"additional","affiliation":[{"name":"IRCCS ICS Maugeri Hospital","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Paola","family":"Baiardi","sequence":"additional","affiliation":[{"name":"IRCCS ICS Maugeri Hospital","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Riccardo","family":"Bellazzi","sequence":"additional","affiliation":[{"name":"Department of Electrical, Computer and Biomedical Engineering, University of Pavia","place":["Pavia, Italy"]},{"name":"IRCCS ICS Maugeri Hospital","place":["Pavia, Italy"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,10,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"104507","DOI":"10.1016\/j.ijmedinf.2021.104507","article-title":"Technological progress in electronic health record system optimization: systematic review of systematic literature reviews","volume":"152","author":"Negro-Calduch","year":"2021","journal-title":"Int J Med Inform"},{"key":"B2","doi-asserted-by":"publisher","first-page":"7359","DOI":"10.1109\/JBHI.2025.3595140","article-title":"Improving clinical foundation models with multi-modal learning and domain adaptation for chronic disease prediction","volume":"30","author":"Hou","year":"2025","journal-title":"IEEE J Biomed Health Inform"},{"key":"B3","author":"Bommasani","year":"2022"},{"key":"B4","doi-asserted-by":"publisher","first-page":"1190","DOI":"10.1093\/jamia\/ocag033","article-title":"Systematic review of foundation models for structured electronic health records","volume":"33","author":"Guo","year":"2026","journal-title":"J Am Med Inform Assoc"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41746-024-01166-w","article-title":"A multi-center study on the adaptability of a shared foundation model for electronic health records","volume":"7","author":"Guo","year":"2024","journal-title":"npj Digit Med"},{"key":"B6","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1109\/RBME.2024.3496744","article-title":"Foundation model for advancing healthcare: challenges, opportunities and future directions","volume":"18","author":"He","year":"2025","journal-title":"IEEE Rev Biomed Eng"},{"key":"B7","doi-asserted-by":"publisher","first-page":"nwaf050","DOI":"10.1093\/nsr\/nwaf050","article-title":"Generative artificial intelligence: a historical perspective","volume":"12","author":"He","year":"2025","journal-title":"Natl Sci Rev"},{"key":"B8","author":"Liu","year":"2024"},{"key":"B9","doi-asserted-by":"publisher","first-page":"1877","DOI":"10.48550\/arXiv.2005.14165","article-title":"Language models are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Adv Neural Inf Process Syst"},{"key":"B10","doi-asserted-by":"publisher","first-page":"3129","DOI":"10.1038\/s41591-024-03185-2","article-title":"A generalist vision\u2013language foundation model for diverse biomedical tasks","volume":"30","author":"Zhang","year":"2024","journal-title":"Nat Med"},{"key":"B11","author":"Kim","year":"2025"},{"key":"B12","doi-asserted-by":"publisher","first-page":"259","DOI":"10.1038\/s41586-023-05881-4","article-title":"Foundation models for generalist medical artificial intelligence","volume":"616","author":"Moor","year":"2023","journal-title":"Nature"},{"key":"B13","doi-asserted-by":"publisher","first-page":"3767","DOI":"10.1038\/s41598-023-30820-8","article-title":"EHR foundation models improve robustness in the presence of temporal distribution shift","volume":"13","author":"Guo","year":"2023","journal-title":"Sci Rep"},{"key":"B14","author":"Bedi","year":"2024"},{"key":"B15","doi-asserted-by":"publisher","first-page":"69","DOI":"10.1038\/s41746-022-00614-9","article-title":"Machine learning generalizability across healthcare settings: insights from multi-site COVID-19 screening","volume":"5","author":"Yang","year":"2022","journal-title":"npj Digit Med"},{"key":"B16","doi-asserted-by":"publisher","first-page":"67125","DOI":"10.52202\/075280-2933","article-title":"EHRSHOT: an EHR benchmark for few-shot evaluation of foundation models","volume":"36","author":"Wornow","year":"2023","journal-title":"Adv Neural Inf Process Syst"},{"key":"B17","doi-asserted-by":"publisher","first-page":"103637","DOI":"10.1016\/j.jbi.2020.103637","article-title":"Language models are an effective representation learning technique for electronic health record data","volume":"113","author":"Steinberg","year":"2021","journal-title":"J Biomed Inform"},{"key":"B18","author":"Wornow","year":"2023"},{"key":"B19","doi-asserted-by":"publisher","first-page":"e48297","DOI":"10.2196\/48297","article-title":"Machine learning\u2013enabled clinical information systems using fast healthcare interoperability resources data standards: scoping review","volume":"11","author":"Balch","year":"2023","journal-title":"JMIR Med Inform"},{"key":"B20","doi-asserted-by":"publisher","first-page":"15","DOI":"10.13063\/2327-9214.1110","article-title":"Transforming the premier perspective\u00ae hospital database into the observational medical outcomes partnership (OMOP) common data model","volume":"2","author":"Makadia","year":"2017","journal-title":"EGEMS (Wash DC)"},{"key":"B21","author":"Arnrich","year":"2024"},{"key":"B22","doi-asserted-by":"publisher","first-page":"62","DOI":"10.1186\/s12911-022-01802-z","article-title":"Network analytics and machine learning for predicting length of stay in elderly patients with chronic diseases at point of admission","volume":"22","author":"Hu","year":"2022","journal-title":"BMC Med Inform Decis Mak"},{"key":"B23","doi-asserted-by":"publisher","first-page":"860","DOI":"10.1186\/s12913-024-11238-y","article-title":"Predicting hospital length of stay using machine learning on a large open health dataset","volume":"24","author":"Jain","year":"2024","journal-title":"BMC Health Serv Res"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1956993\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:34:45Z","timestamp":1791524085000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1956993\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,9]]},"references-count":23,"alternative-id":["10.3389\/fdgth.2026.1956993"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2026.1956993","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,9]]},"article-number":"1956993"}}