摘要
目的:研究基于DeepSeek R1的临床思维培养路径,分析智能技术的教育适配性。方法:以DeepSeek R1开源大模型为研究对象,针对医学教育中智能技术适配性不足的问题,提出“技术特性-认知过程-教育价值”三螺旋分析模型。结果:通过600题临床执业医师资格考试模拟测试的对照试验(DeepSeek R1 vs.GPT-4 Turbo-2025)发现,DeepSeek R1总体准确率达93.50%,显著优于GPT-4(81.50%)(P<0.001)。该模型在临床决策逻辑缺陷(41.03%)、循证医学偏离(23.08%)及伦理情境判断缺失(10.26%)等认知边界,精准映射医学教育中跨学科整合与人文素养培养的结构性短板。结论:AI的决策偏差可作为教学资源重构临床思维培养路径,提出动态思维可视化、课程矩阵靶向迭代、多维评估范式三大教育适配策略,为智能时代医学教育范式转型提供实证依据与方法论创新。
Objective:To investigate the clinical thinking cultivation path based on DeepSeek R1 and analyze the educational adaptability of intelligent technology.Methods:Using the open-source large language model DeepSeek R1 as the research subject,and addressing the issue of insufficient adaptability of intelligent technology in medical education,a three-helix analysis model of"Technical Characteristics-Cognitive Process-Educational Value"was proposed.Results:A controlled experiment involving 600 simulated clinical practitioner qualification exam questions(DeepSeek R1 vs.GPT-4 Turbo-2025)revealed that R1 achieved an overall accuracy of 93.50%,significantly outperforming the control group(81.50%,P<0.001).The model's cognitive boundaries,manifested in clinical decision-making logic flaws(41.03%),deviations from evidence-based medicine(23.08%),and lack of ethical situational judgment(10.26%),precisely mirror the structural shortcomings in interdisciplinary integration and humanistic literacy cultivation within medical education.Conclusion:AI's decision biases can serve as teaching resources to reconstruct the clinical thinking cultivation path.Three educational adaptation strategies are proposed:dynamic thinking visualization,targeted iteration of the curriculum matrix,and a multi-dimensional assessment paradigm.This provides empirical evidence and methodological innovation for the paradigm shift in medical education in the intelligent era.
作者
苏友利
苗长城
程轶
SU Youli;MIAO Changcheng;CHENG Yi(Department of Clinical Medicine,Wanbei Health Vocational College,Suzhou 234000,China;Department of Cardiology,Suqian first Hospital,Suqian 223800,China)
出处
《延边大学医学学报》
2025年第9期128-131,共4页
Journal of Medical Science Yanbian University
基金
安徽省高等学校省级重点质量工程项目(编号:2022jyxm1677)。
