{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T09:02:32Z","timestamp":1789981352510,"version":"4.0.1"},"reference-count":20,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T00:00:00Z","timestamp":1789948800000},"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. Pharmacol."],"abstract":"<jats:sec>\n                    <jats:title>Background\/Objectives<\/jats:title>\n                    <jats:p>Polypharmacy in intensive care units (ICUs) substantially increases the risk of potential drug\u2013drug interactions (pDDIs), which may lead to adverse drug events and worsen clinical outcomes. Conventional rule-based clinical decision support systems (CDSSs), although widely used, are limited by static logic and restricted contextual reasoning. Recently, large language models (LLMs) such as ChatGPT have emerged as potential adjunctive tools capable of generating clinically interpretable outputs; however, their diagnostic reliability in safety-critical settings remains unclear.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>This cross-sectional study aimed to evaluate the diagnostic performance of ChatGPT-5 in detecting clinically significant pDDIs in ICU patients, using the UpToDate Drug Interaction Checker as the reference comparator. Data were collected from 101 adult ICU patients on a predefined index date. The model referred to as \u201cChatGPT-5\u201d corresponds to the version available via the OpenAI API in September 2025. A standardized prompt was applied for each patient to ensure consistency. Diagnostic performance metrics, including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), overall accuracy, and Cohen\u2019s \u03ba coefficient, were calculated.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>ChatGPT-5 identified a higher number of pDDIs per patient compared with UpToDate (median [IQR]: 6 vs. 0 [0\u20132]). The mean number of pDDIs per patient was also higher for ChatGPT-5 (7.2 \u00b1 4.1) compared with UpToDate (1.1 \u00b1 1.8). The diagnostic performance analysis was conducted at the patient level. Of the 101 included patients, 17 without a category D or X interaction in the UpToDate reference assessment were excluded from this analysis, leaving 84 patients. Sensitivity, specificity, and overall accuracy were 69.8%, 52.4%, and 65.5%, respectively, with a positive predictive value of 81.5% and negative predictive value of 36.7%. In a conservative full-cohort scenario analysis, treating the 17 excluded patients as discordant cases reduced overall exact agreement from 65.5% (55\/84) to 54.5% (55\/101). Agreement between the two systems was poor (\u03ba = 0.19), indicating considerable variability in classification.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Although ChatGPT-5 demonstrated limited diagnostic performance and the ability to generate clinically interpretable explanations, its low specificity and limited agreement with a rule-based system highlight important safety concerns. These findings suggest that LLMs may serve as complementary tools rather than standalone solutions in ICU pharmacovigilance. Further multicenter validation studies are required before clinical integration.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fphar.2026.1911849","type":"journal-article","created":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T08:19:34Z","timestamp":1789978774000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Evaluating a large language model (ChatGPT-5) for detecting potential drug\u2013drug interactions in intensive care: a cross-sectional comparative study with a clinical decision support system"],"prefix":"10.3389","volume":"17","author":[{"given":"\u0130lkay","family":"Ceylan","sequence":"first","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Bursa Y\u00fcksek Ihtisas Training and Research Hospital, University of Health Sciences","place":["Bursa, T\u00fcrkiye"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Serpil","family":"Ekin","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Bursa Y\u00fcksek Ihtisas Training and Research Hospital, University of Health Sciences","place":["Bursa, T\u00fcrkiye"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Buket","family":"\u00d6zyaprak","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Bursa Y\u00fcksek Ihtisas Training and Research Hospital, University of Health Sciences","place":["Bursa, T\u00fcrkiye"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Derful","family":"G\u00fclen","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Bursa Y\u00fcksek Ihtisas Training and Research Hospital, University of Health Sciences","place":["Bursa, T\u00fcrkiye"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,21]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"137","DOI":"10.2147\/DHPS.S425858","article-title":"Evaluating the sensitivity, specificity, and accuracy of ChatGPT-3.5, ChatGPT-4, bing AI, and bard against conventional drug-drug interactions clinical tools","volume":"15","author":"Al-Ashwal","year":"2023","journal-title":"Drug Healthc. Patient Saf."},{"key":"B2","doi-asserted-by":"publisher","first-page":"439","DOI":"10.1016\/S0140-6736(23)02465-0","article-title":"The effect of computerised decision support alerts tailored to intensive care on the administration of high-risk drug combinations, and their monitoring: a cluster randomised stepped-wedge trial","volume":"403","author":"Bakker","year":"2024","journal-title":"Lancet"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1142","DOI":"10.1002\/cpt.3585","article-title":"Chat GPT vs. clinical decision support systems in the analysis of drug-drug interactions","volume":"117","author":"Bischof","year":"2025","journal-title":"Clin. Pharmacol. Ther."},{"key":"B4","doi-asserted-by":"publisher","first-page":"280","DOI":"10.1186\/s13052-025-02073-w","article-title":"Will artificial intelligence improve residents' quality of life without compromising healthcare quality? a pediatric point-of-view","volume":"51","author":"Corsello","year":"2025","journal-title":"Ital. J. Pediatr."},{"key":"B5","doi-asserted-by":"publisher","first-page":"5801","DOI":"10.26355\/eurrev_202109_26798","article-title":"Evaluation of potential drug-drug interactions in intensive care unit","volume":"25","author":"Dagdelen","year":"2021","journal-title":"Eur. Rev. Med. Pharmacol. Sci."},{"key":"B6","doi-asserted-by":"publisher","first-page":"100655","DOI":"10.1016\/j.rcsop.2025.100655","article-title":"Comparative evaluation of artificial intelligence platforms and drug interaction screening databases using real-world patient data","volume":"20","author":"Domi\u00e1n","year":"2025","journal-title":"Explor Res. Clin. Soc. Pharm."},{"key":"B7","doi-asserted-by":"publisher","first-page":"1386","DOI":"10.1093\/ajhp\/zxaf038","article-title":"The role of artificial intelligence in emergency medicine pharmacy practice","volume":"82","author":"Edwards","year":"2025","journal-title":"Am. J. Health Syst. Pharm."},{"key":"B8","article-title":"Designing hybrid clinical decision support with generative AI: safety and explainability considerations","volume-title":"Intersection of Artificial Intelligence, Data Science, and Cutting-Edge Technologies: From Concepts to Applications in Smart Environment (ICAISE 2024)","author":"Farhaoui","year":"2025"},{"key":"B9","doi-asserted-by":"publisher","first-page":"e233909","DOI":"10.1001\/jamahealthforum.2023.3909","article-title":"How to safely integrate large language models into health care","volume":"4","author":"Gottlieb","year":"2023","journal-title":"JAMA Health Forum"},{"key":"B10","doi-asserted-by":"publisher","first-page":"1411","DOI":"10.1093\/jamia\/ocae076","article-title":"The use of artificial intelligence to optimize medication alerts generated by clinical decision support systems: a scoping review","volume":"31","author":"Graafsma","year":"2024","journal-title":"J. Am. Med. Inf. Assoc."},{"key":"B11","doi-asserted-by":"publisher","first-page":"93","DOI":"10.4274\/tybd.galenos.2022.32932","article-title":"Potential drug-drug interactions in intensive care units in Turkey: a point prevalence study","volume":"21","author":"G\u00fclen","year":"2023","journal-title":"Turk. J. Intensive Care"},{"key":"B12","doi-asserted-by":"publisher","first-page":"1617","DOI":"10.1093\/jamia\/ocaf139","article-title":"Effect of electronic drug-drug interaction alerts on patient and clinician outcomes: a systematic review","volume":"32","author":"Holbrook","year":"2025","journal-title":"J. Am. Med. Inf. Assoc."},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2503.05777","article-title":"Medical hallucinations in foundation models and their impact on healthcare","author":"Kim","year":"2025","journal-title":"arXiv Prepr."},{"key":"B14","doi-asserted-by":"publisher","first-page":"1237","DOI":"10.1093\/jamia\/ocad072","article-title":"Using AI-generated suggestions from ChatGPT to optimize clinical decision support","volume":"30","author":"Liu","year":"2023","journal-title":"J. Am. Med. Inf. Assoc."},{"key":"B15","doi-asserted-by":"publisher","first-page":"1898","DOI":"10.1007\/s10439-023-03234-w","article-title":"Artificial intelligence in intensive care medicine: toward a ChatGPT\/GPT-4 way?","volume":"51","author":"Lu","year":"2023","journal-title":"Ann. Biomed. Eng."},{"key":"B16","doi-asserted-by":"publisher","first-page":"72","DOI":"10.1186\/s12911-024-02459-6","article-title":"Assessing the research landscape and clinical utility of large language models: a scoping review","volume":"24","author":"Park","year":"2024","journal-title":"BMC Med. Inf. Decis. Mak."},{"key":"B17","doi-asserted-by":"publisher","first-page":"290","DOI":"10.7399\/fh.13180","article-title":"Drug-drug interactions in an intensive care unit and comparison of updates in two databases. Interacciones farmacol\u00f3gicas en una unidad de cuidados intensivos y comparaci\u00f3n de las actualizaciones de dos bases de datos","volume":"46","author":"Riera","year":"2022","journal-title":"Farm Hosp."},{"key":"B18","doi-asserted-by":"publisher","first-page":"12851","DOI":"10.3390\/app152312851","article-title":"Comparing artificial intelligence (ChatGPT, Gemini, DeepSeek) and oral surgeons in detecting clinically relevant drug\u2013drug interactions in dental therapy","volume":"15","author":"Tayeb","year":"2025","journal-title":"Appl. Sci."},{"key":"B19","doi-asserted-by":"publisher","first-page":"100564","DOI":"10.1016\/j.rcsop.2025.100564","article-title":"Exploring potential drug-drug interactions in discharge prescriptions: ChatGPT's effectiveness in assessing those interactions","volume":"17","author":"Thapa","year":"2025","journal-title":"Explor Res. Clin. Soc. Pharm."},{"key":"B20","doi-asserted-by":"publisher","first-page":"103933","DOI":"10.1016\/j.iccn.2024.103933","article-title":"Evaluating the effectiveness of a clinical decision support system (AI-Antidelirium) to improve Nurses\u2019 adherence to delirium guidelines in the intensive care unit","volume":"87","author":"Zhang","year":"2025","journal-title":"Intensive Crit. Care Nurs."}],"container-title":["Frontiers in Pharmacology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fphar.2026.1911849\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T08:19:35Z","timestamp":1789978775000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fphar.2026.1911849\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,21]]},"references-count":20,"alternative-id":["10.3389\/fphar.2026.1911849"],"URL":"https:\/\/doi.org\/10.3389\/fphar.2026.1911849","relation":{},"ISSN":["1663-9812"],"issn-type":[{"value":"1663-9812","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,21]]},"article-number":"1911849"}}