{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T22:54:22Z","timestamp":1781564062063,"version":"3.54.5"},"reference-count":51,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,4,14]],"date-time":"2026-04-14T00:00:00Z","timestamp":1776124800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0"},{"start":{"date-parts":[[2026,6,16]],"date-time":"2026-06-16T00:00:00Z","timestamp":1781568000000},"content-version":"vor","delay-in-days":63,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0"}],"funder":[{"DOI":"10.13039\/501100003710","name":"Korea Health Industry Development Institute","doi-asserted-by":"publisher","award":["RS-2023-KH134974"],"award-info":[{"award-number":["RS-2023-KH134974"]}],"id":[{"id":"10.13039\/501100003710","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100003725","name":"National Research Foundation of Korea","doi-asserted-by":"publisher","award":["RS-2024-00353051"],"award-info":[{"award-number":["RS-2024-00353051"]}],"id":[{"id":"10.13039\/501100003725","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["npj Digit. Med."],"DOI":"10.1038\/s41746-026-02611-8","type":"journal-article","created":{"date-parts":[[2026,4,14]],"date-time":"2026-04-14T02:57:51Z","timestamp":1776135471000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Large language model-augmented offline reinforcement learning framework for sepsis management in critical care"],"prefix":"10.1038","volume":"9","author":[{"given":"Yooseok","family":"Lim","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Byoungjun","family":"Jeon","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Seong-A","family":"Park","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jisoo","family":"Lee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sae Won","family":"Choi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chang Wook","family":"Jeong","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ho-Geol","family":"Ryu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hong Yeul","family":"Lee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hyun-Lim","family":"Yang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,4,14]]},"reference":[{"key":"2611_CR1","doi-asserted-by":"publisher","first-page":"801","DOI":"10.1001\/jama.2016.0287","volume":"315","author":"M Singer","year":"2016","unstructured":"Singer, M. et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). Jama 315, 801\u2013810 (2016).","journal-title":"Jama"},{"key":"2611_CR2","doi-asserted-by":"publisher","first-page":"380","DOI":"10.1016\/S2213-2600(14)70061-X","volume":"2","author":"J-L Vincent","year":"2014","unstructured":"Vincent, J.-L. et al. Assessment of the worldwide burden of critical illness: the intensive care over nations (ICON) audit. Lancet Respir. Med. 2, 380\u2013386 (2014).","journal-title":"Lancet Respir. Med."},{"key":"2611_CR3","doi-asserted-by":"publisher","first-page":"1546","DOI":"10.1056\/NEJMoa022139","volume":"348","author":"GS Martin","year":"2003","unstructured":"Martin, G. S., Mannino, D. M., Eaton, S. & Moss, M. The epidemiology of sepsis in the United States from 1979 through 2000. N. Engl. J. Med. 348, 1546\u20131554 (2003).","journal-title":"N. Engl. J. Med."},{"key":"2611_CR4","doi-asserted-by":"publisher","first-page":"1699","DOI":"10.1056\/NEJMra043632","volume":"355","author":"JA Russell","year":"2006","unstructured":"Russell, J. A. Management of sepsis. N. Engl. J. Med. 355, 1699\u20131713 (2006).","journal-title":"N. Engl. J. Med."},{"key":"2611_CR5","doi-asserted-by":"publisher","first-page":"e1063","DOI":"10.1097\/CCM.0000000000005337","volume":"49","author":"L Evans","year":"2021","unstructured":"Evans, L. et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Crit. Care Med. 49, e1063\u2013e1143 (2021).","journal-title":"Crit. Care Med."},{"key":"2611_CR6","doi-asserted-by":"publisher","DOI":"10.1186\/s13054-020-03204-x","volume":"24","author":"Y Li","year":"2020","unstructured":"Li, Y., Li, H. & Zhang, D. Timing of norepinephrine initiation in patients with septic shock: a systematic review and meta-analysis. Crit. Care 24, 488 (2020).","journal-title":"Crit. Care"},{"key":"2611_CR7","doi-asserted-by":"publisher","first-page":"625","DOI":"10.1007\/s00134-016-4675-y","volume":"43","author":"PE Marik","year":"2017","unstructured":"Marik, P. E., Linde-Zwirble, W. T., Bittner, E. A., Sahatjian, J. & Hansell, D. Fluid administration in severe sepsis and septic shock, patterns and outcomes: an analysis of a large national database. Intensive Care Med. 43, 625\u2013632 (2017).","journal-title":"Intensive Care Med."},{"key":"2611_CR8","doi-asserted-by":"publisher","first-page":"2158","DOI":"10.1097\/CCM.0000000000000520","volume":"42","author":"J Waechter","year":"2014","unstructured":"Waechter, J. et al. Interaction between fluids and vasoactive agents on mortality in septic shock: a multicenter, observational study. Crit. Care Med. 42, 2158\u20132168 (2014).","journal-title":"Crit. Care Med."},{"key":"2611_CR9","doi-asserted-by":"publisher","first-page":"e351","DOI":"10.1097\/CCM.0000000000005363","volume":"50","author":"HJ Yeo","year":"2022","unstructured":"Yeo, H. J. et al. Vasopressor initiation within 1\u2009hour of fluid loading is associated with increased mortality in septic shock patients: analysis of national registry data. Crit. Care Med. 50, e351\u2013e360 (2022).","journal-title":"Crit. Care Med."},{"key":"2611_CR10","doi-asserted-by":"publisher","first-page":"337","DOI":"10.1038\/s41746-024-01316-0","volume":"7","author":"P Jayaraman","year":"2024","unstructured":"Jayaraman, P., Desman, J., Sabounchi, M., Nadkarni, G. N. & Sakhuja, A. A primer on reinforcement learning in medicine for clinicians. NPJ Digit. Med. 7, 337 (2024).","journal-title":"NPJ Digit. Med."},{"key":"2611_CR11","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-023-00893-w","volume":"6","author":"H Lee","year":"2023","unstructured":"Lee, H. et al. Development and validation of a reinforcement learning model for ventilation control during emergence from general anesthesia. npj Digit. Med. 6, 145 (2023).","journal-title":"npj Digit. Med."},{"key":"2611_CR12","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-024-01335-x","volume":"7","author":"HY Lee","year":"2024","unstructured":"Lee, H. Y. et al. Reinforcement learning model for optimizing dexmedetomidine dosing to prevent delirium in critically ill patients. npj Digit. Med. 7, 325 (2024).","journal-title":"npj Digit. Med."},{"key":"2611_CR13","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-023-00986-6","volume":"7","author":"A Boussina","year":"2024","unstructured":"Boussina, A. et al. Impact of a deep learning sepsis prediction model on quality of care and survival. NPJ Digit. Med. 7, 14 (2024).","journal-title":"NPJ Digit. Med."},{"key":"2611_CR14","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-025-59643-z","volume":"16","author":"S Rajendran","year":"2025","unstructured":"Rajendran, S. et al. Multicenter target trial emulation to evaluate corticosteroids for sepsis stratified by predicted organ dysfunction trajectory. Nat. Commun. 16, 4450 (2025).","journal-title":"Nat. Commun."},{"key":"2611_CR15","doi-asserted-by":"publisher","first-page":"245","DOI":"10.1038\/s43856-024-00665-x","volume":"4","author":"Y Choi","year":"2024","unstructured":"Choi, Y. et al. Deep reinforcement learning extracts the optimal sepsis treatment policy from treatment records. Commun. Med. 4, 245 (2024).","journal-title":"Commun. Med."},{"key":"2611_CR16","doi-asserted-by":"publisher","first-page":"1716","DOI":"10.1038\/s41591-018-0213-5","volume":"24","author":"M Komorowski","year":"2018","unstructured":"Komorowski, M., Celi, L. A., Badawi, O., Gordon, A. C. & Faisal, A. A. The Artificial Intelligence Clinician learns optimal treatment strategies for sepsis in intensive care. Nat. Med. 24, 1716\u20131720 (2018).","journal-title":"Nat. Med"},{"key":"2611_CR17","doi-asserted-by":"crossref","unstructured":"Liang, W. & Jia, J. Reinforcement learning using neural networks in estimating an optimal dynamic treatment regime in patients with sepsis. Comput. Methods Programs Biomed. 266, 108754 (2025).","DOI":"10.1016\/j.cmpb.2025.108754"},{"key":"2611_CR18","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2024.102811","volume":"150","author":"Y Wang","year":"2024","unstructured":"Wang, Y. et al. Clinical knowledge-guided deep reinforcement learning for sepsis antibiotic dosing recommendations. Artif. Intell. Med. 150, 102811 (2024).","journal-title":"Artif. Intell. Med."},{"key":"2611_CR19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41746-025-01485-6","volume":"8","author":"PC Nauka","year":"2025","unstructured":"Nauka, P. C. et al. Challenges with reinforcement learning model transportability for sepsis treatment in emergency care. npj Digit. Med. 8, 1\u20135 (2025).","journal-title":"npj Digit. Med."},{"key":"2611_CR20","doi-asserted-by":"publisher","DOI":"10.3389\/fdgth.2021.637386","volume":"3","author":"J Song","year":"2021","unstructured":"Song, J. et al. An effective multimodal image fusion method using MRI and PET for Alzheimer\u2019s disease diagnosis. Front. Digit. Health 3, 637386 (2021).","journal-title":"Front. Digit. Health"},{"key":"2611_CR21","doi-asserted-by":"publisher","first-page":"1227","DOI":"10.1021\/acsptsci.4c00686","volume":"8","author":"G Paraskar","year":"2025","unstructured":"Paraskar, G., Bhattacharya, S. & Kuttiappan, A. The role of proteomics and genomics in the development of colorectal cancer diagnostic tools and potential new treatments. ACS Pharmacol. Transl. Sci. 8, 1227\u20131250 (2025).","journal-title":"ACS Pharmacol. Transl. Sci."},{"key":"2611_CR22","doi-asserted-by":"crossref","unstructured":"Kocak, O. et al. Automated detection and classification of sleep apnea types using electrocardiogram (ECG) and electroencephalogram (EEG) features. Adv. Electrocardiograms-Clin. Appl. 211\u2013230 (2012).","DOI":"10.5772\/22782"},{"key":"2611_CR23","doi-asserted-by":"crossref","unstructured":"Teles, A. S., de Moura, I. R., Silva, F., Roberts, A. & Stahl, D. EHR-based prediction modelling meets multimodal deep learning: a systematic review of structured and textual data fusion methods. Inform. Fusion 118, 102981 (2025).","DOI":"10.1016\/j.inffus.2025.102981"},{"key":"2611_CR24","doi-asserted-by":"crossref","unstructured":"Sinha, S., Kumari, K., Kumar, A. & Mishra, N. Multi-modal data fusion for predictive analytics in healthcare: a hierarchical attention framework for early sepsis prediction. in 2025 6th International Conference for Emerging Technology (INCET) 1\u20136 (IEEE, 2025).","DOI":"10.1109\/INCET64471.2025.11139921"},{"key":"2611_CR25","doi-asserted-by":"publisher","first-page":"369","DOI":"10.1183\/20734735.018016","volume":"12","author":"A Mathioudakis","year":"2016","unstructured":"Mathioudakis, A., Rousalova, I., Gagnat, A. A., Saad, N. & Hardavella, G. How to keep good clinical records. Breathe 12, 369\u2013373 (2016).","journal-title":"Breathe"},{"key":"2611_CR26","first-page":"218","volume":"2014","author":"P Raghavan","year":"2014","unstructured":"Raghavan, P., Chen, J. L., Fosler-Lussier, E. & Lai, A. M. How essential are unstructured clinical narratives and information fusion to clinical trial recruitment?. AMIA Summits Transl. Sci. Proc. 2014, 218 (2014).","journal-title":"AMIA Summits Transl. Sci. Proc."},{"key":"2611_CR27","first-page":"103652","volume":"37","author":"A Nie","year":"2024","unstructured":"Nie, A. et al. OPERA: automatic offline policy evaluation with re-weighted aggregates of multiple estimators. Adv. Neural Inf. Process. Syst. 37, 103652\u2013103680 (2024).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"2611_CR28","unstructured":"Jiang, N. & Li, L. Doubly robust off-policy value evaluation for reinforcement learning. in International conference on machine learning 652\u2013661 (PMLR, 2016)."},{"key":"2611_CR29","unstructured":"Gordon, G. J. Approximate solutions to Markov decision processes (Carnegie Mellon University, 1999)."},{"key":"2611_CR30","first-page":"3014","volume":"27","author":"AR Mahmood","year":"2014","unstructured":"Mahmood, A. R., Van Hasselt, H. P. & Sutton, R. S. Weighted importance sampling for off-policy learning with linear function approximation. Adv. Neural Inform. Process. Syst. 27, 3014\u20133022 (2014).","journal-title":"Adv. Neural Inform. Process. Syst."},{"key":"2611_CR31","first-page":"1179","volume":"33","author":"A Kumar","year":"2020","unstructured":"Kumar, A., Zhou, A., Tucker, G. & Levine, S. Conservative q-learning for offline reinforcement learning. Adv. Neural Inf. Process. Syst. 33, 1179\u20131191 (2020).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"2611_CR32","doi-asserted-by":"publisher","unstructured":"Kaushik, P., Kummetha, S., Moodley, P. & Bapi, R. S. A conservative q-learning approach for handling distribution shift in sepsis treatment strategies. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2203.13884 (2022).","DOI":"10.48550\/arXiv.2203.13884"},{"key":"2611_CR33","doi-asserted-by":"publisher","first-page":"2037","DOI":"10.1109\/TAI.2025.3610390","volume":"7","author":"N Fang","year":"2026","unstructured":"Fang, N., Liu, G. & Gong, W. Offline inverse constrained reinforcement learning for safe-critical decision making in healthcare. IEEE Trans. Artif. Intell. 7, 2037\u20132046 (2026).","journal-title":"IEEE Trans. Artif. Intell."},{"key":"2611_CR34","doi-asserted-by":"publisher","DOI":"10.3389\/fphar.2025.1514445","volume":"16","author":"A Chase","year":"2025","unstructured":"Chase, A. et al. Large language models management of complex medication regimens: a case-based evaluation. Front. Pharmacol. 16, 1514445 (2025).","journal-title":"Front. Pharmacol."},{"key":"2611_CR35","first-page":"695","volume":"690","author":"C-C Liu","year":"2025","unstructured":"Liu, C.-C., Yao, H.-R., Chang, D.-C. & Frieder, O. TreatRAG: a framework for personalized treatment recommendation. Proc. Nineteenth ACM Conf. Recommender Syst. 690, 695 (2025).","journal-title":"Proc. Nineteenth ACM Conf. Recommender Syst."},{"key":"2611_CR36","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/sdata.2016.35","volume":"3","author":"AE Johnson","year":"2016","unstructured":"Johnson, A. E. et al. MIMIC-III, a freely accessible critical care database. Sci. data 3, 1\u20139 (2016).","journal-title":"Sci. data"},{"key":"2611_CR37","doi-asserted-by":"crossref","unstructured":"Johnson, A. E. et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci. Data 10, 1 (2023).","DOI":"10.1038\/s41597-023-01945-2"},{"key":"2611_CR38","doi-asserted-by":"crossref","unstructured":"Sutton, R. S. & Barto, A. G. Reinforcement learning: An introduction (MIT press Cambridge, 1998).","DOI":"10.1109\/TNN.1998.712192"},{"key":"2611_CR39","doi-asserted-by":"publisher","unstructured":"Levine, S., Kumar, A., Tucker, G. & Fu, J. Offline reinforcement learning: tutorial, review, and perspectives on open problems. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2005.01643 (2020).","DOI":"10.48550\/arXiv.2005.01643"},{"key":"2611_CR40","first-page":"24261","volume":"34","author":"IO Tolstikhin","year":"2021","unstructured":"Tolstikhin, I. O. et al. Mlp-mixer: an all-mlp architecture for vision. Adv. Neural Inf. Process. Syst. 34, 24261\u201324272 (2021).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"2611_CR41","doi-asserted-by":"publisher","first-page":"63","DOI":"10.1007\/s44230-025-00093-7","volume":"5","author":"R Tu","year":"2025","unstructured":"Tu, R. et al. Offline safe reinforcement learning for sepsis treatment: tackling variable-length episodes with sparse rewards. Hum.-Centric Intell. Syst. 5, 63\u201376 (2025).","journal-title":"Hum.-Centric Intell. Syst."},{"key":"2611_CR42","doi-asserted-by":"crossref","unstructured":"Nambiar, M. et al. Deep offline reinforcement learning for real-world treatment optimization applications. in Proc. 29th ACM SIGKDD conference on knowledge discovery and data mining 4673\u20134684 (ACM, 2023).","DOI":"10.1145\/3580305.3599800"},{"key":"2611_CR43","doi-asserted-by":"publisher","unstructured":"Team, G. et al. Gemma 3 technical report. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2503.19786 (2025).","DOI":"10.48550\/arXiv.2503.19786"},{"key":"2611_CR44","doi-asserted-by":"publisher","first-page":"515","DOI":"10.1016\/j.ijmedinf.2010.03.004","volume":"79","author":"K Natarajan","year":"2010","unstructured":"Natarajan, K., Stein, D., Jain, S. & Elhadad, N. An analysis of clinical queries in an electronic health record search utility. Int. J. Med. Inform. 79, 515\u2013522 (2010).","journal-title":"Int. J. Med. Inform."},{"key":"2611_CR45","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1093\/jamia\/ocy155","volume":"26","author":"TK Colicchio","year":"2019","unstructured":"Colicchio, T. K. & Cimino, J. J. Clinicians\u2019 reasoning as reflected in electronic clinical note-entry and reading\/retrieval: a systematic review and qualitative synthesis. J. Am. Med. Inform. Assoc. 26, 172\u2013184 (2019).","journal-title":"J. Am. Med. Inform. Assoc."},{"key":"2611_CR46","unstructured":"Tang, S. & Wiens, J. Model selection for offline reinforcement learning: Practical considerations for healthcare settings. in Machine Learning for Healthcare Conference 2\u201335 (PMLR, 2021)."},{"key":"2611_CR47","first-page":"10025","volume":"37","author":"T Udagawa","year":"2023","unstructured":"Udagawa, T., Kiyohara, H., Narita, Y., Saito, Y. & Tateno, K. Policy-adaptive estimator selection for off-policy evaluation. Proc. AAAI Conf. Artif. Intell. 37, 10025\u201310033 (2023).","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"2611_CR48","unstructured":"Fujimoto, S., Meger, D. & Precup, D. Off-policy deep reinforcement learning without exploration. in International conference on machine learning 2052\u20132062 (PMLR, 2019)."},{"key":"2611_CR49","doi-asserted-by":"publisher","unstructured":"Touvron, H. et al. Llama: Open and efficient foundation language models. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2302.13971 (2023).","DOI":"10.48550\/arXiv.2302.13971"},{"key":"2611_CR50","doi-asserted-by":"publisher","first-page":"4171","DOI":"10.18653\/v1\/N19-1423","volume":"ume 1","author":"J Devlin","year":"2019","unstructured":"Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. Bert: Pre-training of deep bidirectional transformers for language understanding. Proc. 2019 Conf. North Am. chapter Assoc. Comput. Linguist.: Hum. Lang. Technol. ume 1, 4171\u20134186 (2019).","journal-title":"Proc. 2019 Conf. North Am. chapter Assoc. Comput. Linguist.: Hum. Lang. Technol."},{"key":"2611_CR51","doi-asserted-by":"crossref","unstructured":"Alsentzer, E. et al. Publicly available clinical BERT embeddings. in Proc. 2nd clinical natural language processing workshop 72\u201378 (Association for Computational Linguistics, 2019).","DOI":"10.18653\/v1\/W19-1909"}],"container-title":["npj Digital Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02611-8","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02611-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02611-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T22:18:43Z","timestamp":1781561923000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02611-8"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,14]]},"references-count":51,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["2611"],"URL":"https:\/\/doi.org\/10.1038\/s41746-026-02611-8","relation":{},"ISSN":["2398-6352"],"issn-type":[{"value":"2398-6352","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,14]]},"assertion":[{"value":"5 November 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"27 March 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"14 April 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors declare no competing interests.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"457"}}