Evaluating the role of ChatGPT in rehabilitation medicine: a narrative review
Open accessThe DOI matches its Crossref record; 48 of the 48 references it cites were found in a registry or an index.
Abstract
Chat Generative Pretrained Transformer (ChatGPT) has emerged as a sophisticated artificial intelligence (AI) language model in healthcare. This narrative review examines ChatGPT's current applications and limitations in rehabilitation medicine through analysing multiple studies. While demonstrating promising performance in structured tasks and basic medical guidance, significant challenges persist. These include inconsistent performance in complex clinical scenarios, limited regional adaptability, poor reference reliability, and inadequate safety considerations for special populations. Although innovative approaches like multi-agent systems show potential improvements in accuracy and interpretability, concerns regarding clinical responsibility, data security, and ethical implications remain crucial. As ChatGPT continues to evolve, its optimal integration into rehabilitation practice requires careful consideration of these limitations and appropriate professional oversight. This review aims to provide insights for healthcare professionals and policymakers in navigating the implementation of AI assistance in rehabilitation medicine, emphasizing the need for balanced integration while maintaining clinical safety and effectiveness.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures


Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“(7) contextual blindness, reflecting a loss of individualized patient understanding in modular AI workflows.”
From Ethical issues in multi-agent AI systems for healthcare: a narrative review (2026) · cites this paper as [37]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The review identifies “contextual blindness” as a significant barrier to safe deployment, where agents optimize for generalized clinical goals while missing the nuanced, value-laden realities of individual patients (37).”
From Ethical issues in multi-agent AI systems for healthcare: a narrative review (2026) · cites this paper as [37]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Contextual blindness and loss of nuance | Agent reasoning prioritizes generalized patterns and optimization goals, often failing to incorporate individual, cultural, or psychosocial context (37, 43).”
From Ethical issues in multi-agent AI systems for healthcare: a narrative review (2026) · cites this paper as [37]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“(2) error propagation and attribution difficulties, complicating accountability for clinical harm;”
From Ethical issues in multi-agent AI systems for healthcare: a narrative review (2026) · cites this paper as [37]
Experimental — accuracy not yet measuredThe check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“This asymmetric accountability, paired with automation bias, accelerates dependency and erodes critical engagement with AI-assisted decisions.”
From Ethical issues in multi-agent AI systems for healthcare: a narrative review (2026) · cites this paper as [37]
Experimental — accuracy not yet measured
1 more claim about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Patient Participation in Physical Therapy Goal Setting
Baker, S. M., Marshak, H. H., Rice, G. T., & Zimmerman, G. J. (2001) · Physical Therapy · doi:10.1093/ptj/81.5.1118 · Cited by 146 · cited 1×
“(8).”
Utzman, R. R., Riddle, D. L., & Jewell, D. V. (2007) · Physical Therapy · doi:10.2522/ptj.20060222 · Cited by 105
ChatGPT: The transformative influence of generative AI on science and healthcare
Varghese, J., & Chapiro, J. (2024) · Journal of Hepatology · doi:10.1016/j.jhep.2023.07.028 · Cited by 89
Özduran, E., Akkoç, İ., Büyükçoban, S., Erkin, Y., & Hancı, V. (2025) · Medicine · doi:10.1097/md.0000000000041780 · Cited by 76
Özduran, E., Hancı, V., Erkin, Y., Özbek, İ. C., & Abdulkarimov, V. (2025) · PeerJ · doi:10.7717/peerj.18847 · Cited by 73
Wolden, M., Hill, B., & Voorhees, S. F. (2019) · Physical Therapy · doi:10.1093/ptj/pzz145 · Cited by 71 · cited 1×
“The pass rate is closely linked to candidates’ academic backgrounds, GRE scores, and other factors (4), with an average first-time pass rate of 84.2% over the past five years (5).”
InfluentialAssessing the performance of AI chatbots in answering patients' common questions about low back pain
Scaff, S. P. S., dos Reis, F. J. J., Ferreira, G. E., Jacob, M. F., & Saragiotto, B. T. (2024) · Annals of the Rheumatic Diseases · doi:10.1136/ard-2024-226202 · Cited by 71
Accuracy of ChatGPT in Common Gastrointestinal Diseases: Impact for Patients and Providers
Kerbage, A., Kassab, J. T., Dahdah, J. E., Burke, C. A., Achkar, J., & Rouphael, C. (2023) · Clinical Gastroenterology and Hepatology · doi:10.1016/j.cgh.2023.11.008 · Cited by 70
Shrestha, N., Shen, Z., Zaidat, B., Duey, A. H., Tang, J. E., Ahmed, W., Hoang, T., Restrepo Mejia, M., Rajjoub, R., Markowitz, J. S., et al. (2024) · Spine · doi:10.1097/brs.0000000000004915 · Cited by 64
Temel, M. H., Erden, Y., & Bağcıer, F. (2024) · World Neurosurgery · doi:10.1016/j.wneu.2023.11.062 · Cited by 55 · cited 1×
“In the domain of complex disease information, findings on spinal cord injury are varied: a study based on GPT 3.5 noted significant issues in information quality and readability, requiring a university-level educational background for full comprehension (29).”
InfluentialRelation extraction using large language models: a case study on acupuncture point locations
Li, Y., Peng, X., Li, J., Zuo, X., Peng, S., Pei, D., Tao, C., Xu, H., & Hong, N. (2024) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocae233 · Cited by 49 · cited 1×
“In knowledge modeling, studies on acupuncture point relationship extraction reveal that a fine-tuned GPT 3.5 model achieved a micro-averaged F1 score of 0.92, markedly outperforming traditional deep learning models and other GPT versions, underscoring the importance of domain-specific fine-tuning (38).”
InfluentialAliyeva, A., Sarı, E., Alaskarov, E., & Nasirov, R. (2024) · Cureus · doi:10.7759/cureus.53897 · Cited by 46
Neo, J. R. E., Ser, J. S., & Tay, S. S. (2024) · Frontiers in Digital Health · doi:10.3389/fdgth.2024.1395501 · Cited by 44
Zhang, L., Tashiro, S., Mukaino, M., & Yamada, S. (2023) · Journal of Rehabilitation Medicine · doi:10.2340/jrm.v55.13373 · Cited by 43
Fish Intake May Affect Brain Structure and Improve Cognitive Ability in Healthy People
Kokubun, K., Nemoto, K., & Yamakawa, Y. (2020) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2020.00076 · Cited by 41 · cited 2×
“Given the global trend of population aging and the consequent increase in cognitive impairments such as mild cognitive impairment (MCI) (19), the accuracy and reliability of ChatGPT in cognitive assessment—particularly relative to established clinical tools— have become critical areas requiring…”
Yapar, D., Avcı, Y. D., Sonuvar, E. T., Eğerci, Ö. F., & Yapar, A. (2023) · Joint Diseases and Related Surgery · doi:10.52312/jdrs.2023.1402 · Cited by 40
GPT-4-based AI agents—the new expert system for detection of antimicrobial resistance mechanisms?
Giske, C. G., Bressan, M., Fiechter, F., Hinić, V., Mancini, S., Nolte, O., & Egli, A. (2024) · Journal of Clinical Microbiology · doi:10.1128/jcm.00689-24 · Cited by 35
Clinical reasoning using ChatGPT: Is it beyond credibility for physiotherapists use?
Bilika, P., Stefanouli, V., Στριμπάκος, Ν., & Kapreli, E. (2023) · Physiotherapy Theory and Practice · doi:10.1080/09593985.2023.2291656 · Cited by 32
Fahy, S., Niemann, M., Böhm, P., Winkler, T., & Oehme, S. (2024) · Journal of Personalized Medicine · doi:10.3390/jpm14050495 · Cited by 32
ChatGPT in the development of medical questionnaires. The example of the low back pain
Coraci, D., Maccarone, M. C., Regazzo, G., Accordi, G., Papathanasiou, J. V., & Masiero, S. (2023) · European Journal of Translational Myology · doi:10.4081/ejtm.2023.12114 · Cited by 31 · cited 1×
“A low back pain (LBP) evaluation questionnaire created with ChatGPT 3.5 showed promising results in multidimensional pain assessment, though improvement is still needed in social relationship assessment (39).”
InfluentialPerceptions of Artificial Intelligence and ChatGPT by Speech-Language Pathologists and Students
Austin, J., Benas, K., Caicedo, S., Imiolek, E., Piekutowski, A., & Ghanim, I. (2025) · American Journal of Speech-Language Pathology · doi:10.1044/2024_ajslp-24-00218 · Cited by 31
Hodson, N., & Williamson, S. (2024) · JMIR AI · doi:10.2196/52500 · Cited by 29 · cited 1×
“…the “catch, check, and change” cognitive framework, indicating a degree of capability in identifying and addressing maladaptive thinking patterns (21), Nevertheless, its utility substantially diminishes when handling complex, individualized psychological cases, underscoring the persistent…”
Mittal, K., & Dhar, M. (2023) · Journal of The Indian Academy of Geriatrics · doi:10.4103/jiag.jiag_27_23 · Cited by 25 · cited 1×
“…of rehabilitation program development for elderly patients, researchers, through systematic parameter optimization and expert evaluation, have validated ChatGPT’s capacity to generate multidimensional rehabilitation plans encompassing physical therapy, occupational therapy, and speech therapy (25).”
Tabanlı, A., & Demirkıran, N. D. (2025) · World Neurosurgery · doi:10.1016/j.wneu.2025.123755 · Cited by 25
Sawamura, S., Bito, T., Ando, T., Masuda, K., Kameyama, S., & Ishida, H. (2024) · Journal of Physical Therapy Science · doi:10.1589/jpts.36.234 · Cited by 23
Chung, S. M., & Chang, M. C. (2024) · BMJ Health & Care Informatics · doi:10.1136/bmjhci-2023-101006 · Cited by 22
Artificial Intelligence in Scoliosis Classification: An Investigation of Language-Based Models
Fabijan, A., Polis, B., Fabijan, R., Zakrzewski, K., Nowosławska, E., & Zawadzka-Fabijan, A. (2023) · Journal of Personalized Medicine · doi:10.3390/jpm13121695 · Cited by 21
Han, J. S., Lim, J. H., Kim, Y., Aliyeva, A., Seo, J., Lee, J., & Park, S. N. (2024) · JMIR mhealth and uhealth · doi:10.2196/50292 · Cited by 21 · cited 1×
“For instance, Han et al. demonstrated that a chat-based mobile auditory training program significantly improved hearing rehabilitation outcomes in experienced hearing aid users, highlighting the practical utility of conversational AI in patient-driven rehabilitation processes (43).”
Arbel, Y., Gimmon, Y., & Shmueli, L. (2025) · Physical Therapy · doi:10.1093/ptj/pzaf010 · Cited by 21 · cited 1×
“Expert evaluations identified 25% of ChatGPT’s responses as “completely incorrect”, highlighting considerable risk when dealing with complex reasoning tasks (47).”
McBee, J. C., Han, D. Y., Liu, L., Ma, L., Adjeroh, D., Xu, D. H., & Hu, G. (2023) · medRxiv · doi:10.1101/2023.07.23.23292452 · Cited by 16
Muluk, S. Y., & Olcucu, N. (2024) · Cureus · doi:10.7759/cureus.63580 · Cited by 15
Performance of ChatGPT on the National Korean Occupational Therapy Licensing Examination
Lee, S.-A., Heo, S., & Park, J. (2024) · Digital Health · doi:10.1177/20552076241236635 · Cited by 14 · cited 1×
“Although it did not meet the passing standard, it performed well on anatomy and physiology questions, which are less influenced by cultural factors, and displayed strong multilingual adaptability (Figure 1) (6).”
Li, Z., Zhang, J.-F., Zhou, W., Zheng, J.-J., & Xia, Y.-S. (2024) · Scientific Reports · doi:10.1038/s41598-024-58514-9 · Cited by 14
Yang, S., & Chang, M. C. (2024) · Medicine · doi:10.1097/md.0000000000038445 · Cited by 14 · cited 1×
“Evaluations indicate that while ChatGPT 4.0 provides accurate information in 75% of common patient queries—91.7% considered safe, and 75% practically useful (17) — these findings predominantly reflect routine clinical queries rather than more complex, individualized therapeutic scenarios.”
InfluentialMcBee, J. C., Han, D. Y., Liu, L., Ma, L., Adjeroh, D. A., Xu, D., & Hu, G. (2024) · JMIR Medical Education · doi:10.2196/51157 · Cited by 14 · cited 1×
“More innovatively, the introduction of the PanelGPT model, which simulates multidisciplinary team discussions involving roles such as physical therapists, psychologists, and nutritionists, highlights ChatGPT’s potential for round-the-clock support and personalized recommendations (41).”
The Role of Artificial Intelligence in the Primary Prevention of Common Musculoskeletal Diseases
Muluk, S. Y., & Olcucu, N. (2024) · Cureus · doi:10.7759/cureus.65372 · Cited by 14
Evaluating cognitive performance: Traditional methods vs. ChatGPT
Fei, X., Tang, Y., Zhang, J., Zhou, Z., Yamamoto, I., & Zhang, Y. (2024) · Digital Health · doi:10.1177/20552076241264639 · Cited by 14 · cited 1×
“Comparative studies employing conventional cognitive assessment tools (MMSE, MoCA) reveal that GPT 4.0 aligns closely with clinician evaluations among cognitively healthy individuals, notably in memory assessment, exhibiting minimal discrepancies (20).”
InfluentialAssessment of the reliability and usability of ChatGPT in response to spinal cord injury questions
Özcan, F., Atar, M. Ö., Köroğlu, Ö., & Yilmaz, B. Ö. (2024) · Journal of Spinal Cord Medicine · doi:10.1080/10790268.2024.2361551 · Cited by 13
Is there any room for ChatGPT AI bot in speech-language pathology?
Birol, N. Y., Çiftci, H. B., Yılmaz, A., Çağlayan, A., & Alkan, F. (2025) · European Archives of Oto-Rhino-Laryngology · doi:10.1007/s00405-025-09295-y · Cited by 12
Hollman, J. H., Cloud, B. A., Krause, D. A., & Calley, D. Q. (2024) · Physical Therapy · doi:10.1093/ptj/pzae006 · Cited by 11
Menekşeoğlu, A. K., & İş, E. E. (2024) · Revista da Associação Médica Brasileira · doi:10.1590/1806-9282.20240241 · Cited by 11
Gallano, G., Giglio, A., & Ferre, A. (2025) · Journal of Voice · doi:10.1016/j.jvoice.2025.04.010 · Cited by 9
Low Performance of ChatGPT on Echocardiography Board Review Questions
Kangiszer, G., Mahtani, A. U., Pintea, M., Jacobs, C., Sragovicz, H., Nguyen, T., Yeturu, S., Lieberman, M., Waldman, C. E., Bhavnani, S. P., et al. (2023) · JACC. Cardiovascular imaging · doi:10.1016/j.jcmg.2023.09.004 · Cited by 8
Enhancing Diabetes Management for Older Patients: The Potential Role of ChatGPT
Tanaka, K., Okazaki, H., Omura, T., Kamihara, T., & Tokuda, H. (2024) · Geriatrics and gerontology international/Geriatrics & gerontology international · doi:10.1111/ggi.14933 · Cited by 7 · cited 1×
“In practical scenarios such as insulin injection guidance, GPT 4.0 delivers comprehensive step-by-step instructions, yet remains somewhat lacking in safety considerations tailored to older adults (24).”
InfluentialChatGPT for speech-impaired assistance
Bhamidipaty, V., Botchu, B., Bhamidipaty, D. L., Guntoory, I., & Iyengar, K. P. (2025) · Disability and Rehabilitation Assistive Technology · doi:10.1080/17483107.2025.2483300 · Cited by 7
Parental education in pediatric dysphagia: A comparative analysis of three large language models
Alyanak, B., Dede, B. T., Bağcıer, F., & Akaltun, M. S. (2025) · Journal of Pediatric Gastroenterology and Nutrition · doi:10.1002/jpn3.70069 · Cited by 6
Harris, S. M., Roddey, T., Shelton, T., Bailey, L. B., Brewer, W., Ellison, J. B., Wang, W., & Gleeson, P. B. (2022) · Musculoskeletal Science and Practice · doi:10.1016/j.msksp.2022.102666 · Cited by 2
ChatGPT and Transcranial Direct Current Stimulation for Chronic Pain.
Silva‐Filho, E., & Pegado, R. (2024) · PubMed Central · Cited by 2
Cited by
Ethical issues in multi-agent AI systems for healthcare: a narrative review
Xie, Z., Wang, H., Dai, L., Wang, Z., Song, H., & Qian, J. (2026) · Frontiers in Public Health · doi:10.3389/fpubh.2026.1792627 · Cited by 11
“Such frameworks might encompass shared liability models, wherein vendors are responsible for technical precision within specified confidence intervals, while clinicians maintain ultimate authority over clinical judgments (37).”
InfluentialBai, T., Jiang, K., Yu, Y., Qie, S., Wang, C., Wang, B., & Zhang, W. (2026) · Future Internet · doi:10.3390/fi18020095 · Cited by 2
“Leveraging their exceptional natural language understanding and generation capabilities, LLMs have been widely adopted for rehabilitation-related textual tasks [84].”
Koluman, A., Yiğitbay, A., Çiftçi, E. A., Çiftçi, M. U., Ziroğlu, N., & Kural, C. (2026) · Journal of Medicine and Palliative Care · doi:10.47582/jompac.1884791 · Cited by 2
Generative AI and Large Language Models in Rehabilitation: A Scoping Review
Cha, S. (2026) · Life · doi:10.3390/life16081310 · Cited by 2
Cali, A., Yörükoğlu, M. E., Açar, G., & Safran, E. (2026) · Frontiers in Medicine · doi:10.3389/fmed.2026.1935702
“This finding is consistent with previous literature indicating that LLM performance varies substantially according to the complexity of the clinical scenario (7, 8).”
Influential
Metadata
- Authors
- Luo, M., Duan, Z., Gao, J., Sun, Y., Chen, L., & Feng, X.
- Affiliations
- Henan University of Chinese Medicine (CN)
- First Affiliated Hospital of Henan University of Traditional Chinese Medicine (CN)
- Corresponding author
- Chen, Lidian
- Feng, Xiaodong
- Venue
- Frontiers in Digital Health
- Publisher
- Frontiers Media SA
- Published
- 2025-10-29
- Volume
- 7
- Pages
- 1618510-1618510
- ISSN
- 2673-253X (ISSN-L)
- PMC
- PMC12605377
- OpenAlex
- W4415650567
- Cited by
- 5
- References
- 47
- Citations by year
- 2026: 5
- FWCI
- 0.30
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Digital Health (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Keywords
- ChatGPT, large language models, rehabilitation medicine, artificial intelligence in healthcare, clinical decision support, multi-agent systems, clinical safety, data security, ethical implications, clinical responsibility, regional adaptability
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.