{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,8,13]],"date-time":"2026-08-13T12:27:34Z","timestamp":1786624054332,"version":"build-2736575974"},"posted":{"date-parts":[[2026,5,21]]},"group-title":"Emergency Medicine","reference-count":27,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,5,21]],"date-time":"2026-05-21T00:00:00Z","timestamp":1779321600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"name":"ARIA Foundation"},{"award":["KL2 TR001862"],"award-info":[{"award-number":["KL2 TR001862"]}],"id":[{"id":"https:\/\/ror.org\/04pw6fb54","id-type":"ROR","asserted-by":"publisher"}]},{"award":["UL1 TR001863"],"award-info":[{"award-number":["UL1 TR001863"]}],"id":[{"id":"https:\/\/ror.org\/04pw6fb54","id-type":"ROR","asserted-by":"publisher"}]},{"award":["5T15LM007056-40"],"award-info":[{"award-number":["5T15LM007056-40"]}],"id":[{"id":"https:\/\/ror.org\/0060t0j89","id-type":"ROR","asserted-by":"publisher"}]},{"id":[{"id":"https:\/\/ror.org\/038cgyc59","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,8,11]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>Emergency department (ED) risk models predict whether a patient will revisit within a fixed window, but not what the visit will entail. We evaluated whether a medical event foundation model predicts the timing, sequence, and severity of downstream care after ED discharge for abdominal pain.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Materials and Methods<\/jats:title>\n                  <jats:p>Retrospective cohort study in the de-identified Epic Cosmos network. From 150,030 eligible adults, we analyzed a random sample of 3,000, generating 50 simulated 30-day trajectories per patient with Curiosity, a generative medical event foundation model, and compared these with gradient boosted tree baselines. Endpoints were discrimination for three outcome classes (any revisit, discharge-revisit, admit-revisit) at three horizons, 36-class trajectory accuracy, transition-level calibration, and subgroup performance.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>Curiosity\u2019s 30-day admit-revisit AUROC was 0.83 (95% CI 0.79-0.87) versus 0.76 (0.71-0.81) for XGBoost (P=.002). Differences were significant at five of nine outcome-horizon pairs but not for discharge-revisit at any horizon. Its most likely trajectory (of 36) matched the observed in 45.9% versus 43.0% (P&lt;.001); median absolute calibration error across 45 transitions was 1.30 percentage points. Discrimination and calibration were poorer for Hispanic or Latino, self-pay, and other-race patients.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Discussion<\/jats:title>\n                  <jats:p>Without task-specific training, one pretrained foundation model matched or exceeded a tuned supervised baseline, with significant advantages on the primary endpoint and trajectory prediction but none for discharge-revisit. Subgroup gaps must be resolved before clinical use.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>A medical event foundation model can represent post-ED-discharge revisit risk as a calibrated trajectory rather than a single probability; prospective evaluation and remediation of subgroup performance gaps remain necessary.<\/jats:p>\n                <\/jats:sec>","DOI":"10.64898\/2026.05.18.26353199","type":"posted-content","created":{"date-parts":[[2026,5,21]],"date-time":"2026-05-21T22:45:16Z","timestamp":1779403516000},"source":"Crossref","is-referenced-by-count":0,"title":["Evaluation of a Generative Medical Event Foundation Model for Predicting Post-Discharge Trajectories in Emergency Department Abdominal Pain"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-5969-4502","authenticated-orcid":false,"given":"Kent A.","family":"McCann","sequence":"first","affiliation":[{"name":"Department of Emergency Medicine, Yale School of Medicine, New Haven, CT"},{"name":"Veterans Affairs Connecticut Healthcare System, West Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2564-7754","authenticated-orcid":false,"given":"Donald S.","family":"Wright","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Yale School of Medicine, New Haven, CT"},{"name":"Veterans Affairs Connecticut Healthcare System, West Haven, CT"},{"name":"Department of Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4446-2488","authenticated-orcid":false,"given":"Mark S.","family":"Iscoe","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Yale School of Medicine, New Haven, CT"},{"name":"Department of Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6509-9537","authenticated-orcid":false,"given":"Edward R.","family":"Melnick","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Yale School of Medicine, New Haven, CT"},{"name":"Veterans Affairs Connecticut Healthcare System, West Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8005-7327","authenticated-orcid":false,"given":"Lucila","family":"Ohno-Machado","sequence":"additional","affiliation":[{"name":"Department of Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1034-7628","authenticated-orcid":false,"given":"Daniella","family":"Meeker","sequence":"additional","affiliation":[{"name":"Department of Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT"},{"name":"Yale New Haven Health System, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8248-0567","authenticated-orcid":false,"given":"Arjun K.","family":"Venkatesh","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Yale School of Medicine, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0435-7029","authenticated-orcid":false,"given":"Rohit B.","family":"Sangal","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Yale School of Medicine, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1754-5133","authenticated-orcid":false,"given":"Andrew J.","family":"Loza","sequence":"additional","affiliation":[{"name":"Veterans Affairs Connecticut Healthcare System, West Haven, CT"},{"name":"Department of Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT"},{"name":"Yale New Haven Health System, New Haven, CT"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026081304300547000_2026.05.18.26353199v2.1","doi-asserted-by":"crossref","first-page":"e281","DOI":"10.1016\/S2589-7500(24)00025-6","article-title":"Foresight\u2014a generative pretrained transformer for modelling of patient timelines using electronic health records: a retrospective modelling study","volume":"6","year":"2024","journal-title":"Lancet Digit Health"},{"key":"2026081304300547000_2026.05.18.26353199v2.2","doi-asserted-by":"crossref","unstructured":"Renc P , Jia Y , Samir AE , et al. Zero shot health trajectory prediction using transformer. NPJ Digit Med 2024;7:256.","DOI":"10.1038\/s41746-024-01235-0"},{"key":"2026081304300547000_2026.05.18.26353199v2.3","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2508.12104"},{"key":"2026081304300547000_2026.05.18.26353199v2.4","doi-asserted-by":"publisher","DOI":"10.1111\/acem.15017"},{"key":"2026081304300547000_2026.05.18.26353199v2.5","doi-asserted-by":"crossref","first-page":"e2350522","DOI":"10.1001\/jamanetworkopen.2023.50522","article-title":"Concordance in medical urgency classification of discharge diagnoses and reasons for visit","volume":"7","year":"2024","journal-title":"JAMA Netw Open"},{"key":"2026081304300547000_2026.05.18.26353199v2.6","doi-asserted-by":"publisher","DOI":"10.1016\/j.jemermed.2018.12.025"},{"key":"2026081304300547000_2026.05.18.26353199v2.7","doi-asserted-by":"publisher","DOI":"10.1136\/emermed-2015-204754"},{"key":"2026081304300547000_2026.05.18.26353199v2.8","doi-asserted-by":"publisher","DOI":"10.1136\/bmjqs-2021-014130"},{"key":"2026081304300547000_2026.05.18.26353199v2.9","doi-asserted-by":"publisher","DOI":"10.1136\/emermed-2018-208343"},{"key":"2026081304300547000_2026.05.18.26353199v2.10","doi-asserted-by":"crossref","unstructured":"Lee YC , Ng CJ , Hsu CC , et al. Machine learning models for predicting unscheduled return visits to an emergency department: a scoping review. BMC Emerg Med 2024;24:20.","DOI":"10.1186\/s12873-024-00939-6"},{"key":"2026081304300547000_2026.05.18.26353199v2.11","doi-asserted-by":"crossref","unstructured":"Hsu CC , Chu C-CJ , Lin CH , et al. A machine learning model for predicting unscheduled 72 h return visits to the emergency department by patients with abdominal pain. Diagnostics (Basel ) 2022;12:82.","DOI":"10.3390\/diagnostics12010082"},{"key":"2026081304300547000_2026.05.18.26353199v2.12","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1016\/j.jemermed.2026.04.007","article-title":"Socioeconomic disparities in 72-hour emergency department return visits (bouncebacks): a multicenter analysis","volume":"87","year":"2026","journal-title":"J Emerg Med"},{"key":"2026081304300547000_2026.05.18.26353199v2.13","first-page":"832","article-title":"High-risk return visits to United States emergency departments, 2010\u20132018","volume":"23","year":"2022","journal-title":"West J Emerg Med"},{"key":"2026081304300547000_2026.05.18.26353199v2.14","doi-asserted-by":"publisher","DOI":"10.1055\/S-0041-1731004"},{"key":"2026081304300547000_2026.05.18.26353199v2.15","doi-asserted-by":"publisher","DOI":"10.1055\/a-2842-3407"},{"key":"2026081304300547000_2026.05.18.26353199v2.16","doi-asserted-by":"publisher","DOI":"10.1111\/acem.70085"},{"key":"2026081304300547000_2026.05.18.26353199v2.17","doi-asserted-by":"publisher","DOI":"10.64898\/2026.08.04.26359713"},{"key":"2026081304300547000_2026.05.18.26353199v2.18","doi-asserted-by":"crossref","unstructured":"Chen T , Guestrin C . XGBoost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, NY: Association for Computing Machinery 2016:785\u201394.","DOI":"10.1145\/2939672.2939785"},{"key":"2026081304300547000_2026.05.18.26353199v2.19","doi-asserted-by":"crossref","unstructured":"Akiba T , Sano S , Yanase T , et al. Optuna: a next-generation hyperparameter optimization framework. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York, NY: Association for Computing Machinery 2019:2623\u201331.","DOI":"10.1145\/3292500.3330701"},{"key":"2026081304300547000_2026.05.18.26353199v2.20","doi-asserted-by":"publisher","DOI":"10.2307\/2531595"},{"key":"2026081304300547000_2026.05.18.26353199v2.21","doi-asserted-by":"crossref","first-page":"31","DOI":"10.1145\/375360.375365","article-title":"A guided tour to approximate string matching","volume":"33","year":"2001","journal-title":"ACM Comput Surv"},{"key":"2026081304300547000_2026.05.18.26353199v2.22","doi-asserted-by":"crossref","unstructured":"Efron B , Tibshirani RJ . An Introduction to the Bootstrap. New York, NY: Chapman and Hall\/CRC 1994.","DOI":"10.1201\/9780429246593"},{"key":"2026081304300547000_2026.05.18.26353199v2.23","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2014.2346248"},{"key":"2026081304300547000_2026.05.18.26353199v2.24","doi-asserted-by":"publisher","DOI":"10.1136\/bmj-2023-078378"},{"key":"2026081304300547000_2026.05.18.26353199v2.25","doi-asserted-by":"crossref","first-page":"e100859","DOI":"10.1136\/bmjhci-2023-100859","article-title":"Prediction of high-risk emergency department revisits from a machine-learning algorithm: a proof-of-concept study","volume":"31","year":"2024","journal-title":"BMJ Health Care Inform"},{"key":"2026081304300547000_2026.05.18.26353199v2.26","doi-asserted-by":"crossref","unstructured":"Hsu SY , Jhu JY , Gao JW , et al. Deep learning to predict emergency department revisit using static and dynamic features (Deep Revisit): development and validation study. BioData Min 2025;18:88.","DOI":"10.1186\/s13040-025-00509-x"},{"key":"2026081304300547000_2026.05.18.26353199v2.27","doi-asserted-by":"publisher","DOI":"10.2196\/74053"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.05.18.26353199","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,13]],"date-time":"2026-08-13T11:30:24Z","timestamp":1786620624000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.64898\/2026.05.18.26353199"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,21]]},"references-count":27,"URL":"https:\/\/doi.org\/10.64898\/2026.05.18.26353199","relation":{},"subject":[],"published":{"date-parts":[[2026,5,21]]},"subtype":"preprint"}}