Niko Michaeel Motaghi, Karbasi Zahra, Kazemi Maryam, Zahmatkeshan Maryam
Student Research Committee, Fasa University of Medical Sciences, Fasa, Iran.
Department of Health Information Sciences, Faculty of Management and Medical Information Sciences, Kerman University of Medical Sciences, Kerman, Iran.
Hypertens Res. 2024 May;47(5):1401-1409. doi: 10.1038/s41440-024-01624-8. Epub 2024 Mar 4.
High blood pressure is one of the major public health problems that is prevalent worldwide. Due to the rapid increase in the number of users of artificial intelligence tools such as ChatGPT and Bing, it is expected that patients will use these tools as a source of information to obtain information about high blood pressure. The purpose of this study is to check the accuracy, completeness, and reproducibility of answers provided by ChatGPT and Bing to the knowledge questionnaire of blood pressure control at home. In this study, ChatGPT and Bing's responses to the HBPM 10-question knowledge checklist on blood pressure measurement were independently reviewed by three cardiologists. The mean accuracy rating of ChatGPT was 5.96 (SD = 0.17) indicating the responses were highly accurate overall, with the vast majority receiving the top score. The mean accuracy and completeness of ChatGPT were 5.96 (SD = 0.17) and 2.93 (SD = 0.25) and in Bing were 5.31 (SD = 0.67), and 2.13 (SD = 0.53) Respectively. Due to the expansion of artificial intelligence applications, patients can use new tools such as ChatGPT and Bing to search for information and at the same time can trust the information obtained. we found that the answers obtained from ChatGPT are reliable and valuable for patients, while Bing is also considered a powerful tool, it has more limitations than ChatGPT, and the answers should be interpreted with caution.
高血压是全球普遍存在的主要公共卫生问题之一。由于ChatGPT和必应等人工智能工具的用户数量迅速增加,预计患者将使用这些工具作为获取高血压信息的来源。本研究的目的是检验ChatGPT和必应对家庭血压控制知识问卷提供答案的准确性、完整性和可重复性。在本研究中,由三位心脏病专家独立审查ChatGPT和必应对关于血压测量的家庭血压监测10题知识清单的回答。ChatGPT的平均准确率评分为5.96(标准差=0.17),表明回答总体上高度准确,绝大多数获得了最高分。ChatGPT的平均准确率和完整性分别为5.96(标准差=0.17)和2.93(标准差=0.25),在必应中分别为5.31(标准差=0.67)和2.13(标准差=0.53)。由于人工智能应用的扩展,患者可以使用ChatGPT和必应等新工具搜索信息,同时可以信任所获得的信息。我们发现,从ChatGPT获得的答案对患者来说是可靠且有价值的,而必应也被认为是一个强大的工具,但其局限性比ChatGPT更多,对答案的解读应谨慎。