Construction of a human-artificial intelligence collaborative audit teaching model for electrocardiogram based on ground truth and typical missed diagnosis/misdiagnosis cases reported by artificial intelligence

Yang Chendie; Hu Lin; Xu Qinyi; Wang Xiaoqun

doi:10.3760/cma.j.cn116021-20251217-02242

您当前的位置：

首页 >

文章列表页 >

Construction of a human-artificial intelligence collaborative audit teaching model for electrocardiogram based on ground truth and typical missed diagnosis/misdiagnosis cases reported by artificial intelligence

更新时间：2026-03-05

- Construction of a human-artificial intelligence collaborative audit teaching model for electrocardiogram based on ground truth and typical missed diagnosis/misdiagnosis cases reported by artificial intelligence
- Chinese Journal of Medical Education Research Pages: 1-7(2026)
- 作者机构：
  
  1.上海交通大学医学院附属瑞金医院心脏内科，上海 200025
  2.上海交通大学医学院附属瑞金临床医学院，上海 200025
  3.上海交通大学医学院医学技术学院，上海 200025
- 作者简介：
  
  Wang Xiaoqun, Email: wangxq@shsmu.edu.cn
- 基金信息：
  
  Shanghai Jiao Tong University School of Medicine Gaofeng Clinical Medicine Grant Support(20240801)
- DOI：10.3760/cma.j.cn116021-20251217-02242
  CLC： R-05
- Received：17 December 2025，
  
  Online First：05 March 2026，
- 稿件说明：
移动端阅览
Yang Chendie, Hu Lin, Xu Qinyi, et al. Construction of a human-artificial intelligence collaborative audit teaching model for electrocardiogram based on ground truth and typical missed diagnosis/misdiagnosis cases reported by artificial intelligence[J/OL]. Chinese Journal of Medical Education Research, 2026, 1-7.
DOI：

Yang Chendie, Hu Lin, Xu Qinyi, et al. Construction of a human-artificial intelligence collaborative audit teaching model for electrocardiogram based on ground truth and typical missed diagnosis/misdiagnosis cases reported by artificial intelligence[J/OL]. Chinese Journal of Medical Education Research, 2026, 1-7. DOI： 10.3760/cma.j.cn116021-20251217-02242.

摘要

目的

本研究旨在为住院医师规范化培训构建一种基于“人机协作审核”的新型教学模式。

方法

选取2025年在本院轮转的60名住院医师为研究对象，随机分为对照组（采用传统讲授式教学，

=30）与试验组（采用“人机协作审核”教学模式，

=30）。采用“双源策略”构建教学资源。选取临床真实的心肌梗死及复杂心律失常等心电图作为原始素材，利用ChatGPT-4o、DeepSeek-V2.5、豆包pro等主流大语言模型生成诊断报告。心内科专家团队对报告进行逻辑核验，构建了包含“正确示范”与“标准化典型错误”的混合案例库。基于此，设计“理论讲授—案例实战—对抗复盘”的三阶段教学流程，并采用“双维加权评分”体系进行过程性考核。应用SPSS 25.0软件进行数据分析，组间比较采用独立样本

检验。

结果

成功构建包含200份正误混合的心电图教学案例库，其中AI正确案例140份，典型陷阱案例60份。初步教学实践表明，应用该模式后，试验组学员心电图出科考核成绩为（86.20±4.78）分，显著优于接受传统讲授式教学的对照组的（78.53±5.15）分，差异有统计学意义（

0.01）。问卷调查显示，试验组在“激发学习兴趣”、“建立诊断信心”及“临床思维培养”等维度的评分均显著高于对照组（

0.01）。

结论

“人机协作审核”模式还原真实的临床人机交互场景，将教学重点从单纯的心电图图形识别拓展至更高阶的批判性思维训练，不仅促进了学员心电图核心诊断能力的提升，也为AI时代的住院医师规范化培训提供了一种可推广的教学范式。

Abstract

Objective

To develop a novel human-artificial intelligence (AI) collaborative audit teaching model for standardized residency training.

Methods

Sixty residents rotating in 2025 were selected and randomly divided into a control group (receiving traditional lecture-based learning

=30) and an experimental group (receiving the " Human-AI Collaborative Audit" teaching model

=30). A dual-source strategy was adopted to construct teaching resources. Real clinical electrocardiograms (ECGs) of myocardial infarction and complex arrhythmias were selected as raw materials. Diagnostic reports were generated using mainstream large language models

including ChatGPT-4o

DeepSeek-V2.5

and Doubao Pro. A team of cardiology experts performed logical verification of the reports to construct a mixed case library containing both correct demonstrations and standardized typical errors. Based on this

a three-stage teaching process including theoretical lecture

case practice

and adversarial debriefing was designed

and a dual-dimension weighted scoring system was employed for process assessment. Data were analyzed using SPSS 25.0 software

and independent sample

-tests were performed for group comparisons.

Results

An ECG teaching case library containing 200 cases was successfully constructed

including 140 correct cases reported by AI and 60 typical trap cases. Preliminary teaching practice showed that after applying this mo

del

the ECG examination score of the experimental group was significantly higher than that of the control group receiving traditional lecture-based teaching (86.20±4.78 vs. 78.53±5.15

0.01). Questionnaire surveys indicated that the experimental group scored significantly higher than the control group in the dimensions of stimulating learning interest

building diagnostic confidence

and clinical thinking cultivation (all

0.01).

Conclusions

By replicating authentic clinical human-AI interaction scenarios

the human-AI collaborative audit model expands the teaching focus from simple ECG pattern recognition to higher-order critical thinking training. This approach not only promotes the core ECG diagnostic abilities of residents but also provides a scalable teaching paradigm for standardized residency training in the AI era.

关键词

Keywords

references

Hannun AY , Rajpurkar P , Haghpanahi M , et al . Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network [J ] . Nat Med , 2019 , 25 ( 1 ): 65 - 69 . DOI: 10.1038/s41591-018-0268-3 http://dx.doi.org/10.1038/s41591-018-0268-3 .

Siontis KC , Noseworthy PA , Attia ZI , et al . Artificial intelligence-enhanced electrocardiography in cardiovascular disease management [J ] . Nat Rev Cardiol , 2021 , 18 ( 7 ): 465 - 478 . DOI: 10.1038/s41569-020-00503-2 http://dx.doi.org/10.1038/s41569-020-00503-2 .

余梦思，宁聪华，孟存仁，等 . 人工智能阅片平台在医学检验专业实习生血细胞形态教学中的应用研究 [J ] . 诊断学理论与实践， 2025 ， 24 （ 6 ）： 668 - 672 . DOI: 10.16150/j.1671-2870.2025.06.015 http://dx.doi.org/10.16150/j.1671-2870.2025.06.015 .

李瑞利，苏壮志，张苗，等 . 临床实践阶段影像医学生及岗前住培学员AI神经影像教育平台的构建 [J ] . 中华医学教育探索杂志， 2025 ， 24 （ 2 ）： 150 - 154 . DOI： 10.3760/cma.j.cn116021-20230927-01964 http://dx.doi.org/10.3760/cma.j.cn116021-20230927-01964 .

Attia ZI , Harmon DM , Behr ER , et al . Application of artificial intelligence to the electrocardiogram [J ] . Eur Heart J , 2021 , 42 ( 46 ): ehab649 . DOI: 10.1093/eurheartj/ehab649 http://dx.doi.org/10.1093/eurheartj/ehab649 .

Gaube S , Suresh H , Raue M , et al . Do as AI say: susceptibility in deployment of clinical decision-aids [J ] . NPJ Digit Med , 2021 , 4 ( 1 ): 31 . DOI: 10.1038/s41746-021-00385-9 http://dx.doi.org/10.1038/s41746-021-00385-9 .

杨文洁，朱晓雷，严福华 . 放射基地住院医师规范化培训中推行分层递进理念的思考 [J ] . 诊断学理论与实践， 2024 ， 23 （ 1 ）： 96 - 99 . DOI: 10.16150/j.1671-2870.2024.01.013 http://dx.doi.org/10.16150/j.1671-2870.2024.01.013 .

张君，郭烽，文涛，等 . 改良CBL联合分层递进教学在核医学住院医师规范化培训中的应用研究 [J ] . 中华医学教育探索杂志， 2025 ， 24 （ 10 ）： 1357 - 1364 . DOI: 10.3760/cma.j.cn116021-20240601-02136 http://dx.doi.org/10.3760/cma.j.cn116021-20240601-02136 .

薛猛，吴佳 . 分层分类教学在消化内科住院医师规范化培训中的应用 [J ] . 中华医学教育探索杂志， 2022 ， 21 （ 2 ）： 199 - 203 . DOI: 10.3760/cma.j.cn116021-20200522-00765 http://dx.doi.org/10.3760/cma.j.cn116021-20200522-00765 .

Jay R , Davenport C , Patel R . Clinical reasoning-the essentials for teaching medical students, trainees and non-medical healthcare professionals [J ] . Br J Hosp Med , 2024 , 85 ( 7 ): 1 - 8 . DOI: 10.12968/hmed.2024.0052 http://dx.doi.org/10.12968/hmed.2024.0052 .

Kahneman D . A perspective on judgment and choice: mapping bounded rationality [J ] . Am Psychol , 2003 , 58 ( 9 ): 697 - 720 . DOI: 10.1037/0003-066X.58.9.697 http://dx.doi.org/10.1037/0003-066X.58.9.697 .

Armoundas AA , Narayan SM , Arnett DK , et al . Use of artificialintelligence in improving outcomes in heart disease: a scientific statement from the American Heart Association [J ] . Circulation , 2024 , 149 ( 14 ): e1028 - e1050 . DOI: 10.1161/CIR.0000000000001201 http://dx.doi.org/10.1161/CIR.0000000000001201 .

Bharatha A , Ojeh N , Fazle Rabbi AM , et al . Comparing the performance of ChatGPT-4 and medical students on MCQs at varied levels of bloom's taxonomy [J ] . Adv Med Educ Pract , 2024 , 15 : 393 - 400 . DOI: 10.2147/AMEP.S457408 http://dx.doi.org/10.2147/AMEP.S457408 .

Al-Zaiti SS , Martin-Gill C , Zègre-Hemsey JK , et al . Machine learning for ECG diagnosis and risk stratification of occlusion myocardial infarction [J ] . Nat Med , 2023 , 29 ( 7 ): 1804 - 1813 . DOI: 10.1038/s41591-023-02396-3 http://dx.doi.org/10.1038/s41591-023-02396-3 .

Cabral S , Restrepo D , Kanjee Z , et al . Clinical reasoning of a generative artificial intelligence model compared with physicians [J ] . JAMA Intern Med , 2024 , 184 ( 5 ): 581 - 583 . DOI: 10.1001/jamainternmed.2024.0295 http://dx.doi.org/10.1001/jamainternmed.2024.0295 .

Schwartzstein RM . Clinical reasoning and artificial intelligence: can AI really think [J ] . Trans Am Clin Climatol Assoc , 2024 , 134 : 133 - 145 .

Ommering BWC , van Blankenstein FM , van Diepen M , et al . Academic success experiences: promoting research motivation and self-efficacy beliefs among medical students [J ] . Teach Learn Med , 2021 , 33 ( 4 ): 423 - 433 . DOI: 10.1080/10401334.2021.1877713 http://dx.doi.org/10.1080/10401334.2021.1877713 .

Mehta N , Mehta S , Rubenstein A , et al . Not replaced, but reinvented: AI education pathways to prepare future physicians to lead healthcare transformation [J ] . Perspect Med Educ , 2025 , 14 ( 1 ): 849 - 859 . DOI: 10.5334/pme.2233 http://dx.doi.org/10.5334/pme.2233 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Application of artificial intelligence assistant system combined with CBL teaching method in the teaching of strabismus specialty of ophthalmic residents

Construction of a core competency evaluation indicator system for hepatobiliary surgery residents based on the stratification-feedback-iteration model

Research on optimization of the scholarship of teaching and learning in clinical medical colleges using artificial intelligence

Exploration and practice of a DeepSeek-enhanced, scenario-based teaching model integrated with a multidisciplinary team approach in neurology residency training

Application of ADDIE model-based situational simulation teaching curricula in pediatric residency training for infection control

Related Author

郭丽莉

鲍永珍

王乐今

Huang Simin

Wang Qiwen

Yang Weibang

Wei Yuxuan

He Longguang

Related Institution

北京大学人民医院眼科

眼视光中心

眼病与视光医学研究所

视网膜脉络膜疾病诊治研究北京市重点实验室

北京大学医学部眼视光学院

⁰