PDF(Pubmed)
Abstract：
Researchers recently utilized ChatGPT as a tool for composing clinic letters, highlighting its ability to generate accurate and empathetic communications. Here we demonstrated the potential application of ChatGPT as a medical assistant in Mandarin Chinese-speaking outpatient clinics, aiming to improve patient satisfaction in high-patient volume settings. ChatGPT achieved an average score of 72.4% in the Chinese Medical Licensing Examination\'s Clinical Knowledge section, ranking within the top 20th percentile. It also demonstrated its potential for clinical communication in non-English speaking environments. Our study suggests that ChatGPT could serve as an interface between physicians and patients in Chinese-speaking outpatient settings, possibly extending to other languages. However, further optimization is required, including training on medical-specific datasets, rigorous testing, privacy compliance, integration with existing systems, user-friendly interfaces, and the development of guidelines for medical professionals. Controlled clinical trials and regulatory approval are necessary before widespread implementation. As chatbots\' integration into medical practice becomes more feasible, rigorous early investigations and pilot studies can help mitigate potential risks.
摘要：
研究人员最近利用ChatGPT作为撰写临床信件的工具，强调其产生准确和善解人意的沟通能力。在这里，我们展示了ChatGPT在讲普通话的门诊诊所中作为医疗助理的潜在应用，旨在提高高患者容量设置中的患者满意度。ChatGPT在中医执业考试的临床知识部分平均得分为72.4%，排名前20百分位。它还证明了其在非英语环境中进行临床交流的潜力。我们的研究表明，ChatGPT可以在讲中文的门诊环境中充当医生和患者之间的接口，可能延伸到其他语言。然而,需要进一步优化，包括医疗特定数据集的培训，严格的测试，隐私合规性，与现有系统集成，用户友好的界面，以及制定医疗专业人员指南。在广泛实施之前，对照临床试验和监管部门的批准是必要的。随着聊天机器人融入医疗实践变得更加可行，严格的早期调查和试点研究可以帮助减轻潜在风险。
