期刊文献+

ChatGPT应用于糖尿病患者教育和沟通的适当性与准确性

Appropriateness and Comprehensiveness of Using ChatGPT for Diabetes Patient Education and Communication
下载PDF
导出
摘要 目的检验ChatGPT 3.5在线人工智能平台对糖尿病相关诊治、药物选择和患者管理等问题回复的准确性及适当性,探讨其在糖尿病患者教育和沟通中的潜在应用。方法参考当地三甲医院糖尿病门诊与患者在线咨询平台,制订了一个涵盖糖尿病诊断、患者管理、药物选择等方面相关10个问题的问卷。可靠性使用专家权威程度进行衡量。将每个问题询问ChatGPT 3.53次,得出3个回复。问卷中每个问题的3个回复都由来自多个中心的专家进行独立主观地评分,且专家可对问题回复提供进一步的补充意见和建议。结果专家对于糖尿病相关的临床知识及诊治的熟悉程度Cs为0.91,专家对ChatGPT回答糖尿病相关问题的准确性的判断依据Ca为0.99、权威系数Cr为0.95,7名专家的评价ICC值为0.753,大多数回复中最常见的质量等级是“好”或“非常好”(92.9%)。结论ChatGPT 3.5对糖尿病相关问题的回答质量相对准确可靠,ChatGPT有望作为帮助患者和医师改善患者教育与沟通,促进良好临床结局的辅助信息工具而发挥作用。 Objective To evaluate the accuracy and appropriateness of the ChatGPT 3.5 online artificial intelligence platform in responding to questions related to diabetes diagnosis and treatment,drug selection,and patient management,and to explore its potential application in diabetes patient education and communication.Methods Referring to the outpatient consultation platform for diabetes at a local top-tier hospital,a questionnaire covering 10 questions related to diabetes diagnosis,patient management,and drug selection was developed.The reliability was measured by the level of expert authority.Each question was posed to ChatGPT 3.5 three times,yielding three responses.The three answers to each question in the questionnaire were independently and subjectively scored by experts from multiple centers,and the experts could provide further comments and suggestions on the responses.Results The experts'familiarity level(Cs)with clinical knowledge and treatment related to diabetes was 0.91.The judgment basis(Ca)for the accuracy of ChatGPT's answers to diabetes-related questions was 0.99,the authority coefficient(Cr)was 0.95,and the ICC value of the evaluations by the 7 experts was 0.753.The most common quality rating in the majority of the responses was"good"or"very good"(92.9%).Conclusions The response quality of ChatGPT 3.5 to diabetes-related questions is relatively accurate and reliable.ChatGPT holds promise as a supplementary information tool to assist patients and doctors in improving patient education and communication,thereby promoting favorable clinical outcomes.
作者 黄玮 王存泽 方美琳 王凌 HUANG Wei;WANG Cunze;FANG Meilin;WANG Ling(Party Committee Office,Fujian Provincial Hospital,Fuzhou 350001,China;Department of Pharmacy,Fujian Provincial Hospital,Fuzhou 350001,China;School of Pharmacy,Fujian Medical University,Fuzhou 350004,China)
出处 《中国医药指南》 2024年第22期22-25,共4页 Guide of China Medicine
基金 福建省财政厅项目[闽财指(20227783)] 福建省卫生健康中青年骨干人才培养项目(2022GGA001)。
关键词 人工智能 糖尿病 患者教育 沟通 Artificial intelligence Diabetes Patient education Communication
  • 相关文献

参考文献1

二级参考文献4

共引文献12

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部