{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,9]],"date-time":"2025-12-09T06:39:40Z","timestamp":1765262380725,"version":"3.46.0"},"reference-count":29,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,12,9]],"date-time":"2025-12-09T00:00:00Z","timestamp":1765238400000},"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. Digit. Health"],"abstract":"<jats:p>\n                    Large language models (LLMs) are used increasingly in medicine, but their decision-making in cardiovascular risk attribution remains underexplored. This pilot study examined how an LLM apportioned relative cardiovascular risk across different demographic and clinical domains. A structured prompt set across six domains was developed, across general cardiovascular risk, body mass index (BMI), diabetes, depression, smoking, and hyperlipidaemia, and submitted in triplicate to ChatGPT 4.0 mini. For each domain, a neutral prompt assessed the LLM's risk attribution, while paired comparative prompts examined whether including the domain changed the LLM's decision of the higher-risk demographic group. The LLM attributed higher cardiovascular risk to men than women, and to Black rather than white patients, across most neutral prompts. In comparative prompts, the LLM's decision between sex changed in two of six domains: when depression was included, risk attribution was equal between men and women. It changed from females being at higher risk than males in scenarios without smoking, but changed to males being at higher risk than females when smoking was present. In contrast, race-based decisions of relative risk were stable across domains, as the LLM consistently judged Black patients to be higher-risk. Agreement across repeated runs was strong (ICC of 0.949, 95% CI: 0.819\u20130.992,\n                    <jats:italic>p<\/jats:italic>\n                    \u2009=\u2009&amp;lt;0.001). The LLM exhibited bias and variability across cardiovascular risk domains. Although decisions between males\/females sometimes changed when comorbidities were included, race-based decisions remained the same. This pilot study suggests careful evaluation of LLM clinical decision-making is needed, to avoid reinforcing inequities.\n                  <\/jats:p>","DOI":"10.3389\/fdgth.2025.1710594","type":"journal-article","created":{"date-parts":[[2025,12,9]],"date-time":"2025-12-09T06:30:37Z","timestamp":1765261837000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Uncovering bias and variability in how large language models attribute cardiovascular risk"],"prefix":"10.3389","volume":"7","author":[{"given":"Justine Tin Nok","family":"Chan","sequence":"first","affiliation":[]},{"given":"Ray Kin","family":"Kwek","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2025,12,9]]},"reference":[{"key":"B1","article-title":"Cardiovascular diseases (CVDs)","year":"2025"},{"key":"B2","doi-asserted-by":"publisher","first-page":"40","DOI":"10.7861\/fhj.2020-0233","article-title":"Implicit bias in healthcare: clinical practice, research and decision making","volume":"8","author":"Gopal","year":"2021","journal-title":"Future Healthc J"},{"key":"B3","doi-asserted-by":"publisher","first-page":"2385","DOI":"10.1016\/S0140-6736(21)00684-X","article-title":"The lancet women and cardiovascular disease commission: reducing the global burden by 2030","volume":"397","author":"Vogel","year":"2021","journal-title":"Lancet"},{"key":"B4","doi-asserted-by":"publisher","first-page":"e000300","DOI":"10.1136\/bmjmed-2022-000300","article-title":"Sex related inequalities in the management and prognosis of acute coronary syndrome in Switzerland: cross sectional study","volume":"1","author":"Huber","year":"2022","journal-title":"BMJ Med"},{"key":"B5","doi-asserted-by":"publisher","first-page":"101445","DOI":"10.1016\/j.lanepe.2025.101445","article-title":"Tolerating cardiovascular health disparities is a failure of science and society","volume":"56","year":"2025","journal-title":"Lancet Reg Health Eur"},{"key":"B6","doi-asserted-by":"publisher","first-page":"631","DOI":"10.3390\/bioengineering12060631","article-title":"Large language models in healthcare and medical applications: a review","volume":"12","author":"Maity","year":"2025","journal-title":"Bioengineering (Basel)"},{"key":"B7","doi-asserted-by":"publisher","first-page":"210","DOI":"10.1161\/CIRCRESAHA.121.320598","article-title":"Long-Term risk prediction for heart failure, disparities, and early prevention","volume":"130","author":"Shea","year":"2022","journal-title":"Circ Res"},{"key":"B8","doi-asserted-by":"publisher","first-page":"818","DOI":"10.1001\/jama.2023.15481","article-title":"AI In medicine\u2014JAMA\u2019s focus on clinical outcomes, patient-centered care, quality, and equity","volume":"330","author":"Khera","year":"2023","journal-title":"JAMA"},{"key":"B9","doi-asserted-by":"publisher","first-page":"1397","DOI":"10.1001\/jama.2020.9371","article-title":"The case for algorithmic stewardship for artificial intelligence and machine learning technologies","volume":"324","author":"Eaneff","year":"2020","journal-title":"JAMA"},{"key":"B10","doi-asserted-by":"publisher","first-page":"81","DOI":"10.1038\/s41746-020-0288-5","article-title":"Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare","volume":"3","author":"Cirillo","year":"2020","journal-title":"NPJ Digit Med"},{"key":"B11","first-page":"835","article-title":"Healthcare bias in AI: a systematic literature review","author":"Moldovan","year":"2025"},{"key":"B12","doi-asserted-by":"publisher","first-page":"100145","DOI":"10.1016\/j.chbah.2025.100145","article-title":"Gender biases within artificial intelligence and ChatGPT: evidence, sources of biases and solutions","volume":"4","author":"Ho","year":"2025","journal-title":"Comput Hum Behav"},{"key":"B13","article-title":"Towards expert-level medical question answering with large language models","author":"Singhal","year":"2023"},{"key":"B14","doi-asserted-by":"publisher","first-page":"16492","DOI":"10.1038\/s41598-023-43436-9","article-title":"Comparing ChatGPT and GPT-4 performance in USMLE soft skill assessments","volume":"13","author":"Brin","year":"2023","journal-title":"Sci Rep"},{"key":"B15","doi-asserted-by":"publisher","first-page":"e54242","DOI":"10.2196\/54242","article-title":"Gender bias in AI\u2019s perception of cardiovascular risk","volume":"26","author":"Achtari","year":"2024","journal-title":"J Med Internet Res"},{"key":"B16","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1186\/s41043-025-01003-0","article-title":"Metabolic health and cardiovascular disease across BMI categories: nHANES findings","volume":"44","author":"Tu","year":"2025","journal-title":"J Health Popul Nutr"},{"key":"B17","doi-asserted-by":"publisher","first-page":"221","DOI":"10.1038\/s42256-024-00976-7","article-title":"What large language models know and what people think they know","volume":"7","author":"Steyvers","year":"2025","journal-title":"Nat Mach Intell"},{"key":"B18","doi-asserted-by":"publisher","first-page":"e64348","DOI":"10.2196\/64348","article-title":"Token probabilities to mitigate large language models overconfidence in answering medical questions: quantitative study","volume":"27","author":"Bentegeac","year":"2025","journal-title":"J Med Internet Res"},{"key":"B19","article-title":"R: The R Project for Statistical Computing","year":""},{"key":"B20","doi-asserted-by":"publisher","first-page":"3590","DOI":"10.1038\/s41591-024-03258-2","article-title":"A toolbox for surfacing health equity harms and biases in large language models","volume":"30","author":"Pfohl","year":"2024","journal-title":"Nat Med"},{"key":"B21","doi-asserted-by":"publisher","first-page":"39.1","DOI":"10.1145\/3641289","article-title":"A survey on evaluation of large language models","volume":"15","author":"Chang","year":"2024","journal-title":"ACM Trans Intell Syst Technol"},{"key":"B22","doi-asserted-by":"publisher","first-page":"57","DOI":"10.1186\/s12939-025-02419-0","article-title":"Evaluating and addressing demographic disparities in medical large language models: a systematic review","volume":"24","author":"Omar","year":"2025","journal-title":"Int J Equity Health"},{"key":"B23","doi-asserted-by":"publisher","first-page":"582","DOI":"10.37765\/ajmc.2022.89261","article-title":"Improving risk stratification using AI and social determinants of health","volume":"28","author":"Carroll","year":"2022","journal-title":"Am J Manag Care"},{"key":"B24","doi-asserted-by":"publisher","first-page":"539","DOI":"10.1093\/eurjpc\/zwae384","article-title":"Temporal trend and attributable risk factors of cardiovascular disease burden for adults 55 years and older in 204 countries\/territories from 1990 to 2021: an analysis for the global burden of disease study 2021","volume":"32","author":"Wang","year":"2025","journal-title":"Eur J Prev Cardiol"},{"key":"B25","article-title":"Challenges in Testing Large Language Model Based Software: A Faceted Taxonomy","author":"Dobslaw","year":"2025"},{"key":"B26","doi-asserted-by":"publisher","first-page":"708","DOI":"10.3390\/biom11050708","article-title":"Susceptibility of women to cardiovascular disease and the prevention potential of mind\u2013body intervention by changes in neural circuits and cardiovascular physiology","volume":"11","author":"Yang","year":"2021","journal-title":"Biomolecules"},{"key":"B27","article-title":"Breaking heart attack stereotypes","year":""},{"key":"B28","doi-asserted-by":"publisher","first-page":"2594","DOI":"10.1016\/j.numecd.2022.08.002","article-title":"Differences in the risk of cardiovascular disease across ethnic groups: uK biobank observational study","volume":"32","author":"Razieh","year":"2022","journal-title":"Nutr Metab Cardiovasc Dis"},{"key":"B29","doi-asserted-by":"publisher","first-page":"782","DOI":"10.1161\/CIRCRESAHA.121.319811","article-title":"Social determinants of cardiovascular disease","volume":"130","author":"Powell-Wiley","year":"2022","journal-title":"Circ Res"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2025.1710594\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,9]],"date-time":"2025-12-09T06:30:38Z","timestamp":1765261838000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2025.1710594\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,9]]},"references-count":29,"alternative-id":["10.3389\/fdgth.2025.1710594"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2025.1710594","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,9]]},"article-number":"1710594"}}