{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,20]],"date-time":"2026-06-20T05:45:13Z","timestamp":1781934313283,"version":"3.54.5"},"reference-count":44,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,12,12]],"date-time":"2025-12-12T00:00:00Z","timestamp":1765497600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Large language models such as ChatGPT hold promise as rapid \u201ccurbside consultation\u201d tools in laboratory medicine. However, their ability to generate consistent and clinically reliable reference intervals\u2014particularly in the absence of contextual clinical information\u2014remains uncertain.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Method<\/jats:title>\n                    <jats:p>This cross-sectional study evaluated whether three versions of ChatGPT (GPT-3.5-Turbo, GPT-4, GPT-4o) maintain repeatable reference-interval outputs when the prompt intentionally omits the interval, using reference interval variability as a stress-test for model consistency. Standardized prompts were submitted through 726,000 chatbot requests. A total of 246,842 reference intervals across 47 laboratory parameters were then analyzed for consistency using the coefficient of variation (CV) and regression models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      On average, the chatbots exhibited a CV of 26.50% (IQR: 7.35\u2013129.01%) for the lower limit and 15.82% (IQR: 4.50\u201345.30%) for the upper limit upon repetition. GPT-4 and GPT-4o demonstrated significantly lower CVs compared to GPT-3.5-Turbo. Reference intervals for poorly standardized parameters were particularly inconsistent across lower (\n                      <jats:italic>\u03b2<\/jats:italic>\n                      : 0.6; 95% CI: 0.35 to 0.86;\n                      <jats:italic>p<\/jats:italic>\n                      \u2009&amp;lt;\u20090.001) and upper limit (\u03b2: 0.5; 95% CI: 0.28 to 0.71;\n                      <jats:italic>p<\/jats:italic>\n                      \u2009&amp;lt;\u20090.001), while unit expressions also showed variability.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>While the newer ChatGPT versions tested demonstrate improved repeatability, diagnostically unacceptable variability persists, particularly for poorly standardized analytes. Mitigating this requires thoughtful prompt design (e.g., mandatory inclusion of reference intervals), global harmonization of laboratory standards, further model refinement, and robust regulatory oversight. Until then, AI chatbots should be restricted to professional use and trained to refuse laboratory interpretation when reference intervals are not provided by the user.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/frai.2025.1681979","type":"journal-article","created":{"date-parts":[[2025,12,12]],"date-time":"2025-12-12T06:32:45Z","timestamp":1765521165000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":4,"title":["ChatGPT and reference intervals: a comparative analysis of repeatability in GPT-3.5 Turbo, GPT-4, and GPT-4o"],"prefix":"10.3389","volume":"8","author":[{"given":"Annika","family":"Meyer","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Edgar","family":"Sch\u00f6mig","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thomas","family":"Streichert","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,12,12]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"e1362","DOI":"10.1002\/ctm2.1362","article-title":"A regulatory challenge for natural language processing (NLP)-based tools such as ChatGPT to be legally used for healthcare decisions. Where are we now?","volume":"13","author":"Baumgartner","year":"2023","journal-title":"Clin. Transl. Med."},{"key":"ref2","doi-asserted-by":"crossref","first-page":"610","DOI":"10.1145\/3442188.3445922","volume-title":"Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency","author":"Bender","year":"2021"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"1487","DOI":"10.3390\/diagnostics11081487","article-title":"Preparing Laboratories for Interconnected Health Care","volume":"11","author":"Bietenbeck","year":"2021","journal-title":"Diagnostics (Basel)."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"991","DOI":"10.1093\/clinchem\/hvad099","article-title":"Reference interval harmonization: harnessing the power of big data analytics to derive common reference intervals across populations and testing platforms","volume":"69","author":"Bohn","year":"2023","journal-title":"Clin. Chem."},{"key":"ref5","doi-asserted-by":"publisher","first-page":"1158","DOI":"10.1515\/cclm-2023-0355","article-title":"Potentials and pitfalls of ChatGPT and natural-language artificial intelligence models for the understanding of laboratory medicine test results. An assessment by the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) working group on artificial intelligence (WG-AI)","volume":"61","author":"Cadamuro","year":"2023","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref6","volume-title":"rio: A Swiss-army knife for data file I\/O","author":"Chan","year":"2023"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"2140","DOI":"10.1515\/cclm-2024-0009","article-title":"The impact of physiological variations on personalized reference intervals and decision limits: an in-depth analysis","volume":"62","author":"Coskun","year":"2024","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref8","year":""},{"key":"ref9","doi-asserted-by":"publisher","first-page":"40","DOI":"10.1007\/s43678-023-00616-w","article-title":"Repeatability, reproducibility, and diagnostic accuracy of a commercial large language model (ChatGPT) to perform emergency department triage using the Canadian triage and acuity scale","volume":"26","author":"Franc","year":"2024","journal-title":"CJEM"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"e55648","DOI":"10.2196\/55648","article-title":"Accuracy of a commercial large language model (ChatGPT) to perform disaster triage of simulated patients using the simple triage and rapid treatment (START) protocol: gage repeatability and reproducibility study","volume":"26","author":"Franc","year":"2024","journal-title":"J. Med. Internet Res."},{"key":"ref11","doi-asserted-by":"publisher","first-page":"27699","DOI":"10.48550\/arXiv.2301.13867","article-title":"Mathematical capabilities of chatgpt","volume":"36","author":"Frieder","year":"2024","journal-title":"Adv. Neural Inf. Proces. Syst."},{"key":"ref12","volume-title":"Viridis(lite)\u2014Colorblind-friendly color maps for R","author":"Garnier","year":"2024"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"915","DOI":"10.1515\/cclm-2013-0081","article-title":"Progress towards standardization: an IFCC scientific division perspective","volume":"51","author":"Gillery","year":"2013","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref14","doi-asserted-by":"publisher","first-page":"1122","DOI":"10.1093\/clinchem\/hvae093","article-title":"ChatGPT vs medical professional: Analyzing responses to laboratory medicine questions on social media","volume":"70","author":"Girton","year":"2024","journal-title":"Clin. Chem."},{"key":"ref15","doi-asserted-by":"publisher","first-page":"57","DOI":"10.1016\/j.cca.2006.06.009","article-title":"Managing transferability of laboratory data","volume":"374","author":"Guidi","year":"2006","journal-title":"Clin. Chim. Acta"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"341","DOI":"10.1515\/cclm-2016-1112","article-title":"Critical comments to a recent EFLM recommendation for the review of reference intervals","volume":"55","author":"Haeckel","year":"2017","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref17","first-page":"250","article-title":"Recommendations on measurement units\u2014why and how","volume":"30","author":"Hansen","year":"2019","journal-title":"EJIFCC"},{"key":"ref18","author":"Hu","year":"2023"},{"key":"ref19","first-page":"835","article-title":"ChatGPT-exploring its role in clinical chemistry","volume":"53","author":"Ibrahim","year":"2023","journal-title":"Ann. Clin. Lab. Sci."},{"key":"ref20","year":"2025"},{"key":"ref21","year":""},{"key":"ref22","year":""},{"key":"ref23","first-page":"169","article-title":"Standardisation of reference intervals: an Australasian view","volume":"28","author":"Jones","year":"2007","journal-title":"Clin. Biochem. Rev."},{"key":"ref24","doi-asserted-by":"publisher","first-page":"752","DOI":"10.1515\/CCLM.2004.127","article-title":"Clinical interpretation of reference intervals and reference limits. A plea for assay harmonization","volume":"42","author":"Klee","year":"2004","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref25","doi-asserted-by":"publisher","first-page":"38","DOI":"10.1515\/cclm-2017-0920","article-title":"An update report on the harmonization of adult reference intervals in Australasia","volume":"57","author":"Koerbin","year":"2018","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref26","doi-asserted-by":"publisher","first-page":"1233","DOI":"10.1056\/NEJMsr2214184","article-title":"Benefits, limits, and risks of GPT-4 as an AI chatbot for medicine","volume":"388","author":"Lee","year":"2023","journal-title":"N. Engl. J. Med."},{"key":"ref27","doi-asserted-by":"publisher","first-page":"e60807","DOI":"10.2196\/60807","article-title":"Performance of ChatGPT across different versions in medical licensing examinations worldwide: systematic review and Meta-analysis","volume":"26","author":"Liu","year":"2024","journal-title":"J. Med. Internet Res."},{"key":"ref28","doi-asserted-by":"publisher","first-page":"e50965","DOI":"10.2196\/50965","article-title":"Comparison of the performance of GPT-3.5 and GPT-4 with that of medical students on the written German medical licensing examination: observational study","volume":"10","author":"Meyer","year":"","journal-title":"JMIR Med. Educ."},{"key":"ref29","doi-asserted-by":"publisher","first-page":"2425","DOI":"10.1515\/cclm-2024-0246","article-title":"Comparison of ChatGPT, Gemini, and Le chat with physician interpretations of medical laboratory questions from an online health forum","volume":"62","author":"Meyer","year":"","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref30","doi-asserted-by":"publisher","first-page":"110276","DOI":"10.1016\/j.clinimag.2024.110276","article-title":"Encouragement vs. liability: how prompt engineering influences ChatGPT-4's radiology exam performance","volume":"115","author":"Nguyen","year":"2024","journal-title":"Clin. Imaging"},{"key":"ref31","author":"O'Connor","year":"2023"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"741","DOI":"10.1515\/cclm-2013-0075","article-title":"Harmonization in laboratory medicine: the complete picture","volume":"51","author":"Plebani","year":"2013","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref33","doi-asserted-by":"publisher","first-page":"184","DOI":"10.3109\/10408363.2015.1116851","article-title":"Harmonization in laboratory medicine: requests, samples, measurements and reports","volume":"53","author":"Plebani","year":"2016","journal-title":"Crit. Rev. Clin. Lab. Sci."},{"key":"ref34","doi-asserted-by":"publisher","first-page":"1579","DOI":"10.1515\/cclm-2017-0865","article-title":"Harmonization in laboratory medicine: more than clinical chemistry?","volume":"56","author":"Plebani","year":"2018","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref35","doi-asserted-by":"publisher","first-page":"185","DOI":"10.1515\/cclm-2022-1122","article-title":"Standardization and harmonization in laboratory medicine: not only for clinical chemistry measurands","volume":"61","author":"Plebani","year":"2023","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref36","doi-asserted-by":"publisher","first-page":"e47564","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","year":"2023","journal-title":"JMIR Hum. Factors"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"949","DOI":"10.1515\/cclm-2015-0016","article-title":"Performance criteria of the post-analytical phase","volume":"53","author":"Sikaris","year":"2015","journal-title":"Clin. Chem. Lab. Med."},{"key":"ref38","volume-title":"gtsummary: presentation-ready data summary and analytic result tables","author":"Sjoberg","year":"2022"},{"key":"ref39","volume-title":"R: A language and environment for statistical computing","year":"2022"},{"key":"ref40","doi-asserted-by":"publisher","first-page":"907S","DOI":"10.3945\/ajcn.115.110387","article-title":"Current practices and challenges in the standardization and harmonization of clinical laboratory tests1223","volume":"104","author":"Vesper","year":"2016","journal-title":"Am. J. Clin. Nutr."},{"key":"ref41","doi-asserted-by":"publisher","first-page":"e53216","DOI":"10.2196\/53216","article-title":"Investigating the impact of prompt engineering on the performance of large language models for standardizing obstetric diagnosis text: comparative study","volume":"8","author":"Wang","year":"2024","journal-title":"JMIR Form Res."},{"key":"ref42","doi-asserted-by":"publisher","first-page":"1686","DOI":"10.21105\/joss.01686","article-title":"Welcome to the {tidyverse}","volume":"4","author":"Wickham","year":"2019","journal-title":"J. Open Source Softw."},{"key":"ref43","doi-asserted-by":"publisher","first-page":"525","DOI":"10.1097\/JCMA.0000000000001092","article-title":"Comparative analysis of GPT-3.5 and GPT-4.0 in Taiwan's medical technologist certification: a study in artificial intelligence advancements","volume":"87","author":"Yang","year":"2024","journal-title":"J. Chin. Med. Assoc."},{"key":"ref44","doi-asserted-by":"publisher","first-page":"1238","DOI":"10.1093\/clinchem\/hvad106","article-title":"AI chatbots in clinical laboratory medicine: foundations and trends","volume":"69","author":"Yang","year":"2023","journal-title":"Clin. Chem."}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1681979\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,12]],"date-time":"2025-12-12T06:32:47Z","timestamp":1765521167000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1681979\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,12]]},"references-count":44,"alternative-id":["10.3389\/frai.2025.1681979"],"URL":"https:\/\/doi.org\/10.3389\/frai.2025.1681979","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,12]]},"article-number":"1681979"}}