ChatGPT在常见胃肠疾病中的准确性:对患者和提供者的影响
本研究评估基于GPT-4的ChatGPT回答患者关于肠易激综合征、炎症性肠病、结肠镜和结直肠癌筛查常见问题,以及从医生角度回答结直肠癌筛查问题和生成参考文献的准确性。此前研究主要关注GPT-3模型,尚无研究探讨ChatGPT生成支持性参考文献的能力或其作为医生教育工具的适用性。
要点
- 评估基于GPT-4的ChatGPT回答胃肠疾病常见问题的准确性。
- 涵盖肠易激综合征、炎症性肠病、结肠镜和结直肠癌筛查。
- 同时评估从医生角度提问及生成参考文献的适用性。
核心结论
- 研究评估了基于GPT-4的ChatGPT在常见胃肠疾病中的准确性及其对患者和提供者的影响。
研究设计与样本
- 研究评估基于GPT-4的ChatGPT回答患者关于肠易激综合征、炎症性肠病、结肠镜和结直肠癌筛查的常见问题,以及从医生角度回答结直肠癌筛查问题和生成参考文献的准确性。
局限性
- 原文未报告局限性
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
Since its release in 2022, Chat Generative Pre-Trained Transformer (ChatGPT) became the most rapidly expanding consumer software application in history,1 and its role in medicine is underscored by its potential to enhance patient education and physician-patient communication. Previous studies in gastroenterology and hepatology have focused primarily on the earlier Generative Pre-Trained Transformer 3 (GPT-3) model, with none investigating ChatGPT's ability to generate supportive references for its responses, or its applicability as a physician educational tool.2-6 Our study evaluated the accuracy of the more recent ChatGPT, powered by GPT-4, in addressing frequently asked questions by patients on irritable bowel syndrome (IBS), inflammatory bowel disease (IBD), colonoscopy and colorectal cancer (CRC) screening, questions on CRC screening from a physician perspective, and reference generation and suitability.
引用
- Kerbage A, Kassab J, El Dahdah J, et al. Accuracy of ChatGPT in Common Gastrointestinal Diseases: Impact for Patients and Providers. Clin Gastroenterol Hepatol. 2023. PMID: 37984563. doi:10.1016/j.cgh.2023.11.008
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。