{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T16:54:28Z","timestamp":1790873668342,"version":"4.1.0"},"reference-count":30,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JMIR Form Res"],"abstract":"Abstract<\/jats:title>\n \n Background<\/jats:title>\n Large language models use machine learning to produce natural language. These models have a range of potential applications in health care, such as patient education and diagnosis. However, evaluations of large language models in health care are still scarce.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aimed to (1) evaluate the accuracy and efficiency of automated fact-checking by 2 large language models and (2) illustrate a process through which a large language model might support a patient in redrafting a prompt to include key information needed for patient safety.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n A parallel comparison of 2 large language models and 3 human experts was conducted. A clinical scenario was devised in which a woman aged 23 years questions the safety of retinoid treatment for acne by sending prompts to 2 large language models (GPT-4o and OpenBioLLM-70B). GPT-4o and OpenBioLLM-70B were asked to suggest improvements to the patient\u2019s initial prompt to elicit key information for clinical decision-making. After the patient sent the revised prompt to the large language models, the models were then asked to fact-check the final response. To test the generalizability of automated fact-checking, a set of 20 clinical statements on disparate topics, mostly related to drug indications, contraindications, and side effects, was developed. The large language models also fact-checked these 20 medical statements. The results were compared against the evaluations of 3 clinical experts. The outcome measures were as follows: (1) percentage of accuracy of automated fact-checking, (2) time to complete fact-checking, and (3) a binary outcome for prompt redrafting (advising the patient to revise her prompt by naming her acne medication to address safety concerns).<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n For the scenario of a patient with acne, GPT-4o and OpenBioLLM-70B both had 86% agreement with the clinical experts\u2019 fact-checking. The large language models did not consistently convey the urgency of discontinuing isotretinoin treatment when pregnancy is suspected. In addition, the models did not adequately convey the importance of folic acid supplementation during pregnancy. For the set of 20 medical claims, GPT-4o fact-checking had 100% agreement with that of human experts, whereas OpenBioLLM-70B had 95% agreement. OpenBioLLM-70B diverged from human experts and GPT-4o on 1 question related to pediatric use of antihistamines. The expert fact-checks took a mean time of 18 (SD 3.74) minutes, GPT-4o took 42 seconds, and OpenBioLLM-70B took 33 minutes. The GPT-4o responses for the acne scenario had some inconsistencies but zero fabrication and no obvious omissions. In contrast, OpenBioLLM-70B omitted 1 key information item needed for patient safety.<\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n GPT-4o can interact with patients to improve the quality and comprehensiveness of the information contained in health-related prompts. GPT-4o and OpenBioLLM-70B can conduct efficient fact-checking that is close to the level of accuracy of human experts. Human experts need to perform additional checks for accuracy and safety.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/68223","type":"journal-article","created":{"date-parts":[[2026,4,15]],"date-time":"2026-04-15T18:51:01Z","timestamp":1776279061000},"page":"e68223-e68223","source":"Crossref","is-referenced-by-count":1,"title":["Fact-Checking Large Language Model Responses to a Health Care Prompt: Comparative Study"],"prefix":"10.2196","volume":"10","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7852-5606","authenticated-orcid":false,"given":"Padhraig","family":"Ryan","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-5643-6482","authenticated-orcid":false,"given":"Orla","family":"Davoren","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0917-6286","authenticated-orcid":false,"given":"Glyn","family":"Elwyn","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,4,15]]},"reference":[{"issue":"2","key":"R1","doi-asserted-by":"publisher","DOI":"10.1056\/AIcs2300145","article-title":"Using ChatGPT to facilitate truly informed medical consent","volume":"1","author":"Mirza","journal-title":"NEJM AI"},{"key":"R2","doi-asserted-by":"publisher","DOI":"10.2196\/47564","article-title":"User intentions to use ChatGPT for self-diagnosis and health-related purposes: cross-sectional survey study","volume":"10","author":"Shahsavar","journal-title":"JMIR Hum Factors"},{"key":"R3","doi-asserted-by":"publisher","DOI":"10.2196\/56797","article-title":"ChatGPT use among pediatric health care providers: cross-sectional survey study","volume":"8","author":"Kisvarday","journal-title":"JMIR Form Res"},{"issue":"1","key":"R4","doi-asserted-by":"publisher","DOI":"10.1136\/bmjhci-2024-101102","article-title":"Generative artificial intelligence in primary care: an online survey of UK general practitioners","volume":"31","author":"Blease","journal-title":"BMJ Health Care Inform"},{"issue":"6681","key":"R5","doi-asserted-by":"publisher","DOI":"10.1126\/science.adn9602","article-title":"Toward the eradication of medical diagnostic errors","volume":"383","author":"Topol","journal-title":"Science"},{"issue":"4","key":"R6","doi-asserted-by":"publisher","first-page":"556","DOI":"10.2214\/AJR.23.29493","article-title":"Accuracy of information provided by ChatGPT regarding liver cancer surveillance and diagnosis","volume":"221","author":"Cao","journal-title":"AJR Am J Roentgenol"},{"key":"R7","doi-asserted-by":"publisher","DOI":"10.2196\/48659","article-title":"Assessing the utility of ChatGPT throughout the entire clinical workflow: development and usability study","volume":"25","author":"Rao","journal-title":"J Med Internet Res"},{"issue":"7","key":"R8","doi-asserted-by":"publisher","first-page":"1488","DOI":"10.1108\/OIR-02-2024-0111","article-title":"Scientific evidence and specific context: leveraging large language models for health fact-checking","volume":"48","author":"Ni","journal-title":"Online Inf Rev"},{"key":"R9","doi-asserted-by":"crossref","unstructured":"Zarharan M Wullschleger P Pilehvar MT Foster J Kia BB . Tell me why: explainable public health fact-checking with large language models. arXiv. Preprint posted online on May 15, 2024. [doi: 10.48550\/ARXIV.2405.09454]","DOI":"10.18653\/v1\/2024.trustnlp-1.21"},{"issue":"6","key":"R10","doi-asserted-by":"publisher","first-page":"e379","DOI":"10.1016\/S2589-7500(24)00060-8","article-title":"The effect of using a large language model to respond to patient messages","volume":"6","author":"Chen","journal-title":"Lancet Digit Health"},{"key":"R11","doi-asserted-by":"crossref","unstructured":"Vladika J Hacajova I Matthes F . Step-by-step fact verification system for medical claims with explainable reasoning. arXiv. Preprint posted online on Feb 20, 2025. [doi: 10.48550\/arXiv.2502.14765]","DOI":"10.18653\/v1\/2025.naacl-short.68"},{"key":"R12","unstructured":"Dorfner FJ Dada A Busch F et al. Biomedical large languages models seem not to be superior to generalist models on unseen medical data. arXiv. Preprint posted online on Aug 25, 2024. [doi: 10.48550\/ARXIV.2408.13833]"},{"key":"R13","doi-asserted-by":"crossref","unstructured":"Pal A Umapathi LK Sankarasubbu M . Med-HALT: medical domain hallucination test for large language models. Presented at: Proceedings of the 27th Conference on Computational Natural Language Learning (CoNLL); Dec 6-7, 2023; Singapore, Singapore. URL: https:\/\/aclanthology.org\/2023.conll-1.21.pdf [Accessed 28-01-2026]","DOI":"10.18653\/v1\/2023.conll-1.21"},{"key":"R14","unstructured":"Zhang C Yang R Zhang Z et al. Atomic calibration of LLMs in long-form generations. Preprint posted online on Oct 17, 2024. [doi: 10.48550\/arXiv.2410.13246]"},{"key":"R15","doi-asserted-by":"publisher","DOI":"10.3389\/frai.2024.1341697","article-title":"The perils and promises of fact-checking with large language models","volume":"7","author":"Quelle","journal-title":"Front Artif Intell"},{"issue":"7454","key":"R16","doi-asserted-by":"publisher","first-page":"1490","DOI":"10.1136\/bmj.328.7454.1490","article-title":"Grading quality of evidence and strength of recommendations","volume":"328","author":"Atkins","journal-title":"BMJ"},{"key":"R17","doi-asserted-by":"publisher","DOI":"10.2196\/54704","article-title":"A preliminary checklist (METRICS) to standardize the design and reporting of studies on generative artificial intelligence-based models in health care education and practice: development study involving a literature review","volume":"13","author":"Sallam","journal-title":"Interact J Med Res"},{"key":"R18","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-024-00881-z","article-title":"Factuality challenges in the era of large language models and opportunities for fact-checking","volume":"6","author":"Augenstein","journal-title":"Nat Mach Intell"},{"issue":"8","key":"R19","doi-asserted-by":"publisher","DOI":"10.1016\/j.patter.2021.100308","article-title":"Did chatbots miss their \u201cApollo Moment\u201d? Potential, gaps, and lessons from using collaboration assistants during COVID-19","volume":"2","author":"Srivastava","journal-title":"Patterns (N Y)"},{"issue":"7","key":"R20","doi-asserted-by":"publisher","first-page":"1030","DOI":"10.1253\/circj.CJ-23-0308","article-title":"Evaluation of the accuracy of ChatGPT in answering clinical questions on the Japanese Society of Hypertension Guidelines","volume":"87","author":"Kusunose","journal-title":"Circ J"},{"issue":"5","key":"R21","doi-asserted-by":"publisher","first-page":"292","DOI":"10.1136\/bmjebm-2023-112651","article-title":"Meet generative AI\u2026 your new shared decision-making assistant","volume":"29","author":"Elwyn","journal-title":"BMJ Evid Based Med"},{"key":"R22","doi-asserted-by":"crossref","unstructured":"Lin S Hilton J Evans O . TruthfulQA: measuring how models mimic human falsehoods. Presented at: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics; May 22-27, 2022; Dublin, Ireland. URL: https:\/\/aclanthology.org\/2022.acl-long.229.pdf [Accessed 28-01-2026]","DOI":"10.18653\/v1\/2022.acl-long.229"},{"issue":"7","key":"R23","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2024.22399","article-title":"Large language model-based responses to patients' in-basket messages","volume":"7","author":"Small","journal-title":"JAMA Netw Open"},{"issue":"12","key":"R24","doi-asserted-by":"publisher","first-page":"4180","DOI":"10.1109\/TFUZZ.2024.3431710","article-title":"A hybrid framework integrating LLM and ANFIS for explainable fact-checking","volume":"33","author":"Bangerter","journal-title":"IEEE Trans Fuzzy Syst"},{"issue":"1","key":"R25","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-024-01157-x","article-title":"The ethics of ChatGPT in medicine and healthcare: a systematic review on Large Language Models (LLMs)","volume":"7","author":"Haltaufderheide","journal-title":"NPJ Digit Med"},{"issue":"1","key":"R26","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-024-01282-7","article-title":"Medical large language models are susceptible to targeted misinformation attacks","volume":"7","author":"Han","journal-title":"NPJ Digit Med"},{"issue":"6","key":"R27","doi-asserted-by":"publisher","first-page":"968","DOI":"10.1016\/j.berh.2017.05.001","article-title":"Clinical guidelines for low back pain: a critical review of consensus and inconsistencies across three major guidelines","volume":"30","author":"O\u2019Connell","journal-title":"Best Pract Res Clin Rheumatol"},{"issue":"3","key":"R28","doi-asserted-by":"publisher","first-page":"330","DOI":"10.1016\/j.cjca.2023.10.016","article-title":"A systematic review of clinical practice guidelines and consensus statements for cardiac rehabilitation delivery: consensus, divergence, and important knowledge gaps","volume":"40","author":"Cotie","journal-title":"Can J Cardiol"},{"issue":"8","key":"R29","doi-asserted-by":"publisher","first-page":"1665","DOI":"10.1093\/jamia\/ocae142","article-title":"Using large language model to guide patients to create efficient and comprehensive clinical care message","volume":"31","author":"Liu","journal-title":"J Am Med Inform Assoc"},{"issue":"15","key":"R30","doi-asserted-by":"publisher","first-page":"1320","DOI":"10.1001\/jama.2023.27861","article-title":"Comparative analysis of multimodal large language model performance on clinical vignette questions","volume":"331","author":"Han","journal-title":"JAMA"}],"container-title":["JMIR Formative Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/formative.jmir.org\/2026\/1\/e68223","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T17:54:49Z","timestamp":1782842089000},"score":1,"resource":{"primary":{"URL":"https:\/\/formative.jmir.org\/2026\/1\/e68223"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,15]]},"references-count":30,"URL":"https:\/\/doi.org\/10.2196\/68223","relation":{},"ISSN":["2561-326X"],"issn-type":[{"value":"2561-326X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,15]]},"article-number":"v10i2e68223"}}