中国全科医学 ›› 2024, Vol. 27 ›› Issue (34): 4315-4321.DOI: 10.12114/j.issn.1007-9572.2023.0544

• 论著·全科医学教育研究 • 上一篇    下一篇

基于人工智能的医患沟通共情语言教学与评价系统的开发与应用研究

邵建文1, 刘欢2, 张玥1, 郑爱明1, 陈松宇3, 王锦帆1,*()   

  1. 1.211166 江苏省南京市,南京医科大学医患沟通研究中心
    2.100730 北京市,北京协和医学院人文和社会科学学院
    3.210014 江苏省南京市,江苏布洛氪链数据科技有限公司
  • 收稿日期:2024-04-10 修回日期:2024-08-12 出版日期:2024-12-05 发布日期:2024-09-13
  • 通讯作者: 王锦帆

  • 作者贡献:

    邵建文负责软件系统运行管理、教学数据收集与整理、论文撰写;刘欢、张玥、郑爱明参与课题各环节的研究与实施;陈松宇负责软件系统的技术设计、实施、运行维护及数据提供;王锦帆为课题主持人,全面负责实施系统构建、教学实验和论文撰写及审定,对文章整体负责。

  • 基金资助:
    南京医科大学2021年一流本科课程虚拟仿真实验教学类立项课题

Development and Application of an AI-based Empathic Language Teaching and Evaluation System for Doctor-patient Communication

SHAO Jianwen1, LIU Huan2, ZHANG Yue1, ZHENG Aiming1, CHEN Songyu3, WANG Jinfan1,*()   

  1. 1. Research Center for Doctor Patient Communication, Nanjing Medical University, Nanjing 211166, China
    2. School of Humanities, Peking Union Medical College, Beijing 100730, China
    3. Jiangsu Blokrypton Chain Data Technology Co., LTD, Nanjing 210014, China
  • Received:2024-04-10 Revised:2024-08-12 Published:2024-12-05 Online:2024-09-13
  • Contact: WANG Jinfan

摘要: 背景 新医科建设背景下,教育部鼓励信息技术与医学教育深度融合,培养一流医学人才服务健康中国建设,目前医患沟通共情能力教学多以模拟沟通、小组讨论形式为主,依托人工智能技术开展教学较少。 目的 探索构建可以用于课程教学的医患沟通共情语言教学与评价系统,为今后医患沟通共情教学提供新的方向;开展教学应用,提高医学生、医生沟通共情语言表达能力,并收集反馈用于系统的优化改善。 方法 2021年9月—2022年2月课题组基于讯飞语音识别技术和共情语义识别算法,应用课题组研制的10个医患沟通典型案例、示范共情语言、共情语言语义库、共情语言技能及其整体评分标准等进行系统"医患沟通共情语言虚拟仿真教学与评价系统"(以下简称系统)的开发;2022年3—5月选取南京医科大学参与医患沟通学课程或培训的950名学生(包括515名本科生、102名医学博士生和333名临床医生)为研究对象,基于本系统开展南京医科大学《医患沟通学》技能课(2学时)的教学试验。采用自拟调查问卷收集研究对象的共情语言内涵掌握程度、共情语言能力增强程度、系统便捷性认知程度、系统融入教学合理性认知程度等,并采用NVivo软件对研究对象的反馈意见和建议进行词频分析。 结果 应用系统后,本科生、临床医生、医学博士生的共情语言内涵掌握程度、共情语言能力增强程度、系统便捷性认知程度、系统融入教学合理性认知程度比较,差异有统计学意义(P<0.05),其中76.1%(723/950)的研究对象的共情语言内涵掌握程度评价结果为"完全掌握"或"掌握程度较高";93.8%(891/950)的研究对象表明本系统可"显著增强"或"有些增强"共情语言能力,89.5%(850/950)的研究对象对该系统便捷性认知程度的评价为"非常便捷"或"比较便捷";95.1%(903/950)的研究对象对该系统融入教学合理性认知程度的评价为"非常合理"或"比较合理"。反馈意见和建议中词频位于前五的关键词依次为:沟通、语音、教学、程序、标准。 结论 该系统能够提高医学生和医生在医患沟通中由个案到一般的共情能力,自主型系统的使用解放了师生教学的时空局限,其规范的课程教学方式也得到了研究对象的正向合理性反馈,具有广泛的应用前景,但目前处于初级探索阶段,仍需要不断完善。

关键词: 人工智能, 医患沟通, 共情语言, 教学评价

Abstract:

Background

Under the background of new medical science, the deep integration of information technology and medical education is encouraged to train first-class medical talents to serve the construction of healthy China.Currently, empathy training in doctor-patient communication mainly consists of simulated communication and group discussion, with less reliance on artificial intelligence technology for learning.

Objective

To develop a system for teaching and evaluating doctor-patient communication empathy language. This system will be used in course teaching to pave the way for future doctor-patient communication empathy teaching methods. Carry out teaching applications to enhance the communication and empathy language expression skills of medical students and doctors, and gather feedback to optimize and improve the system.

Methods

Between September 2021 and February 2022, the research group focus on utilizing iFlytek speech recognition technology and the empathy semantic recognition algorithm. A system called the "Doctor-patient Communication Virtual Simulation Teaching and Evaluation System of empathic language" was developed using 10 typical cases of doctor-patient communication, demonstrations of empathic language, a semantic database of empathic language, empathic language skills, and an overall scoring standard.A total of 950 students from Nanjing Medical University, including 515 undergraduates, 102 medical doctoral students, and 333 clinicians participating in doctor-patient communication courses or training, were selected as the research subjects from March to May 2022. Based on this system, the Doctor-patient Communication Skills Course (2 class hours) teaching experiment was conducted at Nanjing Medical University. A self-designed questionnaire was used to gather information on the subjects' understanding of empathetic language connotations, their improved empathetic language skills, their perception of system ease of use, and their perception of how the system integrates into the rationality of teaching. NVivo software was used to analyze the subjects' feedback, comments, and suggestions.

Results

Following the implementation of the system, there were statistically significant differences in the mastery of empathic language connotation, the degree of enhancement of empathic language ability, the degree of convenience of the system, and the degree of integration of the system into teaching rationality among undergraduate students, clinicians, and medical doctoral students (P<0.05). 76.1% (723/950) of the participants evaluated that they had "fully mastered" or "highly mastered" the connotation of empathic language. 93.8% (891/950) of the study subjects indicated that the system could "significantly enhance" or "somewhat enhance" the empathic language ability, and 89.5% (850/950) of the study subjects rated the convenience of the system as "very convenient" or "relatively convenient". 95.1% (903/950) of the study subjects rated the degree of cognition of the rationality of integrating the system into teaching as "very reasonable" or "relatively reasonable". The top five words mentioned in the feedback and suggestions are communication, pronunciation, teaching, program, and standard.

Conclusion

This system can help improve medical student and doctors'ability to empathize in doctor-patient communication by learning from individual cases and applying those lessons more broadly. Additionally, the use of an autonomous teaching evaluation system frees up the constraints of time and space in teacher-student interactions. The system's standardized teaching method has received positive and rational feedback from participants, indicating its potential for a wide range of applications. However, the system is still in the early stages of exploration and requires further refinement.

Key words: Artificial intelligence, Doctor-patient communication, Empathy, Teaching evaluation