CoChat
Sign Up
Cited by 9Artificial Intelligence in Healthcare and EducationArtificial Intelligence ApplicationsMachine Learning in Healthcare

From chat to act: large language model agents and agentic AI as the next frontier of AI in rheumatology

Open access
Madrid-García, Alfredo, Benavent, Diego, Merino-Barbancho, Beatriz·2025·EULAR Rheumatology Open
DOI 10.1016/j.ero.2025.06.012PMC PMC13292459
Open this paper in CoChat — free account

Abstract

Objectives: Large language models (LLMs) have begun to influence rheumatology, yet their static knowledge and hallucination risks limit their potential. Retrieval-augmented generation mitigates some limitations, but complex rheumatologic care demands real-time data access, multistep reasoning, and tool usage that exceed standard LLM capabilities. The objective of this study is to explore how agentic artificial intelligence (AI) can address the limitations of current LLM applications in rheumatology. Methods: We conducted a viewpoint analysis of the capabilities of agentic AI systems, focusing on their technical foundations, current use cases in healthcare, and relevance to the specific demands of rheumatologic care. Results: Agentic AI extends LLMs with planning, memory, and the ability to interact with external tools, enabling execution of complex tasks. These capabilities offer promising applications in rheumatology, including personalized treatment planning, automated literature synthesis, and clinical decision support. Conclusions: Agentic AI systems represent a necessary evolution to meet the complexity of rheumatologic care. Regulatory, ethical and technical challenges must be overcome before agentic systems can be safely deployed in routine rheumatologic care.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1016/j.ero.2025.06.012

Cites

8 of 8 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Multimodal multi-agent models demonstrated superior diagnostic performance by integrating imaging and text more effectively than classical CNN-based or text-only systems.

    “while multimodal multiagent models demonstrated superior diagnostic performance by integrating imaging and text more effectively than classical CNN-based or text-only systems [23][24].”

    From Multi-Agent Retrieval Augmented Generation for Clinical Decision Support: A Systematic Review and Integrative Conceptual Framework (2026) · cites this paper as [23]

    The Introduction’s “Agents and agentic AI” section reports that combining several LLM outputs can improve diagnostic accuracy, and Table 2 describes a potential vision-language application integrating images with clinical metadata. Neither passage demonstrates that multimodal multi-agent models outperform CNN-based or text-only diagnostic systems; Table 2 presents an expected application, not a comparative result. No page numbers are provided in our copy. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Multimodal agent specialisation makes clear how radiological patterns and textual evidence relate to the overall prediction.

    “This made it clear how radiological patterns and textual evidence worked in relation to the overall prediction [23].”

    From Multi-Agent Retrieval Augmented Generation for Clinical Decision Support: A Systematic Review and Integrative Conceptual Framework (2026) · cites this paper as [23]

    Table 2 (AI agents in rheumatology; page unavailable) describes a vision-language model that integrates images with clinical metadata to produce a diagnostic score and probability. The Responsible AI agents section discusses explainability in general, but neither passage says that multimodal agent specialisation makes clear how radiological patterns and textual evidence each relate to an overall prediction. (LLM, conf=0.78)

    Experimental — accuracy not yet measured

  • Multimodal systems add transparency by demonstrating how image features are connected to text retrieval to draw a conclusion.

    “Multimodal [23][24] also added transparency by demonstrating how the image features were connected to text retrieval in order to draw a conclusion.”

    From Multi-Agent Retrieval Augmented Generation for Clinical Decision Support: A Systematic Review and Integrative Conceptual Framework (2026) · cites this paper as [23]

    The Introduction discusses multimodal rheumatology data, while “Responsible AI agents” says explainability can make an agent’s decisions transparent. Neither passage describes a multimodal system demonstrating how image features connect to text retrieval to reach a conclusion. Searches for image features, text retrieval, and transparency found no passage addressing that mechanism. (LLM, conf=0.74)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 44

  1. [1]Zhang K. Meng X. Yan X. Ji J. Liu J. Xu H. Revolutionizing health care: the transformative impact of large language models in medicine J Med Internet Res [Internet] 27 2025 e59069 [cited 2025 May 15]: [15 p.]. Available from: https://www.jmir.org/2025/1/e59069 PMC11751657 39773666
  2. [2]Liu F. Zhou H. Gu B. Zou X. Huang J. Wu J. Application of large language models in medicine Nat Rev Bioeng 3 6 2025 445 464 10.1038/s44222-025-00279-5 · DOI · Paper page
  3. [3]Madrid-García A. Merino-Barbancho B. Rodríguez-González A. Fernández-Gutiérrez B. Rodríguez-Rodríguez L. Menasalvas-Ruiz E. Understanding the role and adoption of artificial intelligence techniques in rheumatology research: an in-depth review of the literature Semin Arthritis Rheum 61 2023 152213 10.1016/j.semarthrit.2023.152213 [cited 2025 May 15 ] [24 p.] Available from: https://www.sciencedirect.com/science/article/pii/S0049017223000550 37315379 · DOI
  4. [4]Benavent D. Madrid-García A. Large language models and rheumatology: are we there yet? Rheumatol Adv Pract 2024 rkae119 10.1093/rap/rkae119 40256630 PMC12007598 · DOI
  5. [5]Madrid-García A. Rosales-Rosado Z. Freites-Nuñez D. Pérez-Sancristóbal I. Pato-Cour E. Plasencia-Rodríguez C. Harnessing ChatGPT and GPT-4 for evaluating the rheumatology questions of the Spanish access exam to specialized medical training Sci Rep 13 1 2023 22129 10.1038/s41598-023-49483-6 PMC10719375 38092821 · DOI
  6. [6]Madrid-García A. Benavent D. Merino-Barbancho B. Rosales-Rosado Z. Merino-Barbancho B. Freites-Núnez D. Optimizing the clinical application of rheumatology guidelines using large language models: a retrieval-augmented generation framework integrating ACR and EULAR recommendations medRxiv 2025.04.10.25325588 2025 [Preprint] [cited 2025 May 15]: [52 p.]. Available from: https://www.medrxiv.org/content/10.1101/2025.04.10.25325588v2 PMC13292420 42368617 · DOI
  7. [7]Kremer P. Schiebisch H. Lechner F. Haase I. Cirkel L. Kötter I. Comparative analysis of large language models and traditional diagnostic decision support systems for rare rheumatic disease identification EULAR Rheumatol Open 1 2 2025 51 59 10.1016/j.ero.2025.04.007 PMC13292275 42368386 · DOI · Paper page
  8. [8]Bürgisser N. Chalot E. Mehouachi S. Buclin C.P. Lauper K. Courvoisier D.S. Large language models for accurate disease detection in electronic health records: the examples of crystal arthropathies RMD Open 10 2024 e005003 [cited 2025 May 15 ] [7 p.]. Available from: https://rmdopen.bmj.com/content/10/4/e005003 PMC11664341 39794274
  9. [9]Nair S. Lushington G.H. Purushothaman M. Rubin B. Jupe E. Gattam S. Prediction of lupus classification criteria via generative AI medical record profiling BioTech (Basel) 14 1 2025 15 10.3390/biotech14010015 40227336 PMC11940096 · DOI · Paper page
  10. [10]Munguía-Realpozo P. Mendoza-Pinto C. Etchegaray-Morales I. Ramírez-Lara E. Solis-Poblano J.C. Méndez-Martínez S. Evaluating large language models as a supplementary patient information resource on antimalarial use in systemic lupus erythematosus Lupus 34 4 2025 374 380 10.1177/09612033251324501 40013495 · DOI
  11. [11]Kim Y. Jeong H. Chen S. Ramírez-Lara E. Solis-Poblano J.C. Méndez-Martínez S. Medical hallucinations in foundation models and their impact on healthcare arXiv:2503.05777 2025 [Preprint][cited 2025 May 15]: [95 p.]. Available from: https://arxiv.org/abs/2503.05777 · arXiv
  12. [12]Mousavi S.M. Alghisi S. Riccardi G. DyKnow: dynamically verifying time-sensitive factual knowledge in LLMs arXiv:2404.08700 [Preprint] 2025 [cited 2025 May 15]: [16 p.]. Available from: https://arxiv.org/abs/2404.08700 · arXiv
  13. [14]Lewis P. Perez E. Piktus A. Petroni F. Karpukhin V. Goyal N. Retrieval-augmented generation for knowledge-intensive NLP tasks Larochelle H. Ranzato M. Hadsell R. Balcan M.F. Lin H. Advances in Neural Information Processing Systems 2020. Advances in Neural Information Processing Systems 33 (NeurIPS 2020) 2020 Curran Associates, Inc London 339459 339474 December 6-12, 2020
  14. [15]McNamara K.J. Marpu RP. Demystifying AI agents: the final generation of intelligence arXiv:2505.09932 [Preprint] 2025 [cited 2025 May 15]: [10 p.]. Available from: https://arxiv.org/abs/2505.09932 · arXiv
  15. [16]Ferrag M.A. Tihanyi N. Debbah M. From LLM Reasoning to autonomous AI agents: a comprehensive review arXiv:2504.19678 [Preprint] 2025 [cited 2025 May 15]: [44 p.]. Available from: https://arxiv.org/abs/2504.19678 · arXiv
  16. [18]Veluwenkamp H. Buijsman S. Design for operator contestability: control over autonomous systems by introducing defeaters AI Ethics 5 2025 3699 3711 10.1007/s43681-025-00657-0 · DOI
  17. [20]Zheng T. Deng Z. Tsang H.T. From automation to autonomy: a survey on large language models in scientific discovery arXiv:2505.13259 2025 [Preprint] 2025 [cited 2025 May 15]: [16 p.]. Available from: https://arxiv.org/abs/2505.13259 · arXiv
  18. [23]Schneider J. Generative to agentic AI: survey, conceptualization, and challenges arXiv:2504.18875 [Preprint] 2025 [cited 2025 May 15]: [54 p.]. Available from: https://arxiv.org/abs/2504.18875 · arXiv
  19. [24]Wang W. Ma Z. Wang Z. Wu C. Ji J. Chen W. A survey of LLM-based agents in medicine: how far are we from Baymax? arXiv:2502.11211 [Preprint] 2025 [cited 2025 May 15]: [14 p.]. Available from: https://arxiv.org/abs/2502.11211 · arXiv
  20. [25]Sapkota R. Roumeliotis K.I. Karkee M. AI agents vs. agentic AI: a conceptual taxonomy, applications and challenge arXiv:2505.10468 [Preprint] 2025 [cited 2025 May 15]: [36 p.]. Available from: https://arxiv.org/abs/2505.10468 · arXiv
  21. [26]Barabucci G. Shia V. Chu E. Harack B. Laskowski K. Fu N. Combining multiple large language models improves diagnostic accuracy NEJM AI 1 11 2024 10.1056/AIcs2400502 · DOI · Paper page
  22. [27]Kim Y. Park C. Jeong H. Chan Y.S. Xu X. McDuff D. MDAgents: an adaptive collaboration of LLMs for medical decision-making Adv Neural Inf Process Syst 37 2024 79410 79452
  23. [29]Li H. Pan W. Rajendran S. Zang C. Wang F. TrialGenie: empowering clinical trial design with agentic intelligence and real world data medRxiv 2025.04.17.25326033 [Preprint] 2025 [cited 2025 May 15]: [34 p.]. Available from: https://www.medrxiv.org/content/10.1101/2025.04.17.25326033 PMC13341881 42414290 · DOI
  24. [30]Song K. Trotter A. Chen JY. LLM agent swarm for hypothesis-driven drug discovery arXiv:2504.17967 [Preprint] 2025 [cited 2025 May 15]: [15 p.]. Available from: https://arxiv.org/abs/2504.17967 · arXiv
  25. [31]Zhang G. Song C. Liu L. Wang Q. Li C. TransAgent: dynamizing transcriptional regulation analysis via multi-omics-aware AI agent bioRxiv 2025.04.27.650826 [Preprint] 2025 [cited 2025 May 15]: [18 p.]. Available from: https://www.biorxiv.org/content/10.1101/2025.04.27.650826 · DOI
  26. [32]Qu Y. Huang K. Cousins H. Johnson W.A. Yin D. Shah M. CRISPR-GPT: an LLM agent for automated design of gene-editing experiments bioRxiv 2024.04.25.591003 [Preprint] 2024 [Preprint] 2024. [cited 2025 May 15]: [18 p.]. Available from: https://www.biorxiv.org/content/10.1101/2025.04.27.650826v1 · DOI
  27. [34]Zeng F. Lyu Z. Li Q. Li X. Enhancing LLMs for impression generation in radiology reports through a multi-agent system arXiv:2412.06828 [Preprint] 2024 [cited 2025 May 15]: [12 p.]. Available from: https://arxiv.org/abs/2412.06828 · arXiv
  28. [35]Das R. Maheswari K. Siddiqui S. Improved precision oncology question-answering using agentic LLM medRxiv 2024.09.20.24314076 [Preprint] 2024 [cited 2025 May 15]: [39 p.]. Available from: https://www.medrxiv.org/content/10.1101/2024.09.20.24314076 · DOI
  29. [36]Chan T.K. Dinh N-D. ENTAgents: AI agents for complex knowledge otolaryngology medRxiv 2025.01.01.25319863 [Preprint] 2025 [cited 2025 May 15]: [93 p.]. Available from: https://www.medrxiv.org/content/10.1101/2025.01.01.25319863 · DOI · Paper page
  30. [38]Maarseveen T.D. Glas H.K. Veris-van Dieren J. van den Akker E. Knevel R. Improving musculoskeletal care with AI enhanced triage through data driven screening of referral letters NPJ Digit Med 8 1 2025 98 10.1038/s41746-025-01495-4 39948271 PMC11825706 · DOI · Paper page
  31. [39]Bayala Y.L.T. Zabsonré/Tiendrebeogo W.J.S. Ouedraogo D-D Kaboré F. Sougué C. Yameogo A.R. Performance of the large language models in African rheumatology: a diagnostic test accuracy study of ChatGPT-4, Gemini, Copilot, and Claude artificial intelligence BMC Rheumatol 9 1 2025 54 10.1186/s41927-025-00512-z 40380276 PMC12083132 · DOI
  32. [40]Creagh A.P. Hamy V. Yuan H. Mertes G. Tomlinson R. Chen W.H. Digital health technologies and machine learning augment patient reported outcomes to remotely characterise rheumatoid arthritis NPJ Digit Med 7 1 2024 33 10.1038/s41746-024-01013-y 38347090 PMC10861520 · DOI · Paper page
  33. [41]Patel D. Raut G. Cheetirala S.N. AI agents in modern healthcare: from foundation to pioneer—a comprehensive review and implementation roadmap for impact and integration in clinical settings 202503.1352 [Preprint] 2025 [cited 2025 May 15]: [27 p.]. Available from: https://www.preprints.org/manuscript/202503.1352
  34. [42]Taxter AJ, Natter MD. Using the electronic health record to enhance care in pediatric rheumatology. Rheumatic Disease Clinics of North America. 2022;48(1):245-58. doi:10.1016/j.rdc.2021.08.004 34798950 · DOI
  35. [43]Chaffer T.J. Charles von Goins I.I. Okusanya B. Cotlage D. Goldston J. Decentralized governance of autonomous AI agents arXiv:2412.17114 [Preprint] 2025 [cited 2025 May 15]: [23 p.]. Available from: https://arxiv.org/abs/2412.17114 · arXiv
  36. [44]Kim J. Shin B. Chung J. Rhu M. The cost of dynamic reasoning: demystifying AI agents and test-time scaling from an AI infrastructure perspective arXiv:2506.04301 [Preprint] 2025 [cited XXX]: [XX p.]. Available from: XXX · arXiv
  37. [45]Murugesan S. The rise of agentic AI: implications, concerns, and the path forward IEEE Intell Syst 40 2 2025 8 14 10.1109/MIS.2025.3544940 · DOI
  38. [47]Ong J.C.L. Chang SY-H William W. Butte A.J. Shah N.H. Chew L.S. Medical ethics of large language models in medicine NEJM AI 1 7 2024 10.1056/AIra2400038 · DOI
  39. [48]Desai D.R. Riedl MO. Responsible AI Agents arXiv:2502.18359 [Preprint] 2025 [cited 2025 May 15]: [51 p.]. Available from: https://arxiv.org/abs/2502.18359 · arXiv
  40. [51]Ranjan R. Gupta S. Singh SN. Fairness in agentic AI: a unified framework for ethical and equitable multi-agent system arXiv:2502.07254 [Preprint] 2025 [cited 2025 May 15]: [12 p.]. Available from: https://arxiv.org/abs/2502.07254 · arXiv
  41. [52]Hammond L. Chan A. Clifton J. Hoelscher-Obermaier J. Khan A. McLean E. Multi-agent risks from advanced AI arXiv:2502.14143 [Preprint] 2025 [cited 2025 May 15]: [96 p.]. Available from: https://arxiv.org/abs/2502.14143 · arXiv
  42. [53]Liu B. Li X. Zhang J. Wang J. He T. Hong S. Advances and challenges in foundation agents: from brain-inspired intelligence to evolutionary, collaborative, and safe systems arXiv:2504.01990 [Preprint] 2025 [cited 2025 May 15]: [264 p.]. Available from: https://arxiv.org/abs/2504.01990 · arXiv
  43. [54]Singh K. Ngu W. Bias-aware agent: enhancing fairness in AI-driven knowledge retrieval Companion Proceedings of the ACM on Web Conference 2025 2025 May 8 Association for Computing Machinery;2025 New York 1705 1712
  44. [55]Osogami T. AI agents should be regulated based on the extent of their autonomous operations arXiv:2503.04750 [Preprint] 2025 [cited 2025 May 15]: [22 p.]. Available from: https://arxiv.org/abs/2503.04750 · arXiv

Cited by

Sorted by citations

Authors
Madrid-García, A., Benavent, D., & Merino-Barbancho, B.
Affiliations
Bellvitge University Hospital (ES)
Corresponding author
Madrid-García, Alfredo
ORCID
Venue
EULAR Rheumatology Open
Publisher
Elsevier BV
Published
2025-10
Volume
1
Issue
3
Pages
147-156
ISSN
3050-7081 (ISSN-L)
DOI
10.1016/j.ero.2025.06.012
PMC
PMC13292459
OpenAlex
W4412723643
Cited by
9
References
8
Citations by year
2026: 7 · 2025: 2
FWCI
0.53
Type
journal article
Language
en
Open access
Licence
Locations
  • EULAR Rheumatology Open (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
Topics
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
Keywords
agentic AI, large language models, rheumatology, retrieval-augmented generation, hallucinations, tool use, clinical decision support, personalized treatment planning, artificial intelligence in healthcare, ethical and regulatory challenges

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review