Application of AI-empowered scenario-based simulation teaching mode in cardiovascular disease education

Abstract

Background

Cardiovascular diseases present a significant challenge in clinical practice due to their sudden onset and rapid progression. The management of these conditions necessitates cardiologists to possess strong clinical reasoning and individual competencies. The internship phase is crucial for medical students to transition from theory to practical application, with an emphasis on developing clinical thinking and skills. Despite the critical need for education on cardiovascular diseases, there is a noticeable gap in research regarding the utilization of artificial intelligence in clinical simulation teaching.

Objective

This study aims to evaluate the effect and influence of AI-empowered scenario-based simulation teaching mode in the teaching of cardiovascular diseases.

Methods

The study utilized a quasi-experimental research design and mixed-methods. The control group comprised 32 students using traditional teaching mode, while the experimental group included 34 students who were instructed on cardiovascular diseases using the AI-empowered scenario-based simulation teaching mode. Data collection included post-class tests, “Mini-CEX” assessments, Clinical critical thinking scale from both groups, and satisfaction surveys from experimental group. Qualitative data were gathered through semi-structured interviews.

Results

Research shows that compared with traditional teaching models, AI-empowered scenario-based simulation teaching mode significantly improve students’ performance in many aspects. The theoretical knowledge scores(P < 0.001), clinical operation skills(P = 0.0416) and clinical critical thinking abilities of students(P < 0.001) in the experimental group were significantly improved. The satisfaction survey showed that students in the experimental group were more satisfied with the teaching scene(P = 0.008), Individual participation(P = 0.006) and teaching content(P = 0.009). There is no significant difference in course discussion, group cooperation and teaching style of teachers(P > 0.05). Additionally, the qualitative data from the interviews highlighted three themes: (1) Positive new learning experience, (2) Improved clinical critical thinking skills, and (3) Valuable suggestions and concerns for further improvement.

Conclusion

The AI-empowered scenario simulation teaching Mode plays an important role in the improvement of clinical thinking and skills of medical undergraduates. This study believes that the AI-empowered scenario simulation teaching mode is an effective and feasible teaching model, which is worthy of promotion in other courses.

Introduction

Cardiovascular diseases, including myocardial infarction and arrhythmia, frequently manifest abruptly and progress rapidly, placing individuals in critical situations. In addition to the physical distress, the substantial rates of disability and mortality linked to these conditions impose a significant burden on both families and society. Furthermore, the presence of commodities such as diabetes and chronic obstructive pulmonary disease in many cardiovascular disease patients adds further complexity to treatment strategies [1, 2]. In light of this context, the importance of internship training in cardiology is underscored [3]. In China, when medical students enter their fourth and fifth years of undergraduate study, they will be placed in hospitals for a clinical internship lasting one to two years. Internship plays a critical role in the development of medical students, facilitating the transition from theoretical knowledge to practical application and fostering the growth of clinical reasoning and skills [4]. Nevertheless, the prevailing mode of internship education primarily relies on conventional instructional approaches, which prioritize teacher-led dissemination of knowledge through lectures and demonstrations [5, 6]. Although these methods are successful in facilitating knowledge acquisition, they are inadequate in motivating students, promoting clinical reasoning, and cultivating the skills necessary to manage emergency situations, particularly when dealing with critically ill patients. As a result, it is essential to implement a shift in teaching methodologies, specifically within the realm of cardiology internship training.

In recent years, the rapid development of Artificial Intelligence (AI) technology has led to the emergence of various products profoundly impacting various aspects of people’s lives [7]. Generative AI, a type of AI based on deep learning, involves training large-scale language models to generate new text, images, or other types of data. Notably, models like OpenAI’s ChatGPT use deep learning algorithms trained on extensive datasets to generate human-like responses in conversation. In the realm of education, generative AI exhibits tremendous potential. Firstly, it can offer personalized learning experiences by tailoring learning paths based on individual student needs and proficiency levels, enhancing learning effectiveness and making education more targeted and efficient [8, 9]. Secondly, generative AI plays a crucial role in automatic assessment and feedback, providing students with immediate and constructive feedback, promoting better understanding and mastery of knowledge. Additionally, through simulated dialogues, role-playing, and other mode, generative AI can help students improve communication and problem-solving skills, offering new possibilities for flexible, intelligent teaching mode and driving innovation and progress in education [10].

Scenario-based simulation teaching is an instructional method that involves simulating real-world situations for teaching purposes, commonly used in clinical education. In this approach, students are placed in virtual or real scenarios where they face specific problems, challenges, or tasks, engaging in practical activities and decision-making to proficiently apply knowledge [11]. This teaching method emphasizes practicality and interactivity, allowing students not only to apply theoretical knowledge in simulated situations but also to actively participate in discussions, collaborate on problem-solving, and enhance their practical application and teamwork skills [12]. Research indicates that scenario-based simulation teaching stimulates student interest, increases motivation, and fosters critical thinking and innovation by integrating theoretical knowledge into practice [13].

Nowadays, with the rapid development of science, new technologies such as Virtual Reality and Augmented Reality have brought significant changes to clinical medicine. For example, clinical scenario simulation surgery allows doctors to create a virtual surgical training platform. This allows them to practice complex surgical skills in a safe, repeatable practice environment [14–17]. While studies have demonstrated the effectiveness of scenario-based simulation teaching in clinical courses [11–13, 18], there is currently no research on the application of generative AI in simulating clinical scenarios related to cardiovascular diseases. In this study, we aim to investigate the effectiveness of the AI-empowered scenario-based simulation teaching mode in cardiovascular disease education. Our goal is to explore the impact of this innovative teaching model on clinical interns, focusing on their basic knowledge, clinical operation ability and clinical critical thinking ability.

Methods

Experimental design

A combination of quasi-experimental research design and descriptive qualitative research methods was employed to form both a control group and an experimental group. Our study integrated Kolb’s experiential learning model into the experimental group’s teaching methods to enhance the learning process [19, 20]. Kolb’s experiential learning model involves providing learners with real or simulated situations and activities. Under the guidance of teachers, learners participate in these activities to gain personal experience. They then reflect on and summarize their observations, developing theories or conclusions, which are ultimately applied in practice (Fig. 1).

Fig. 1.

Kolb’s experiential learning model

Study participants

A total of 66 first-year students from two classes in the clinical major at Nantong University were selected as the study participants. Inclusion criteria comprised: (1) absence of current physical or mental abnormalities; (2) full-time undergraduate students in medical majors; (3) no prior experience using the AI platform for medical course learning before the experiment; (4) voluntary participation in the study with the signing of an informed consent form. The control group consisted of 32 students, following a traditional teaching model, while the experimental group comprised 34 students undergoing scenario-based simulation teaching mode empowered by AI.

All students entered university directly through the national college entrance examination (gaokao) after completing 12 years of education. After inclusion, an assessment of the characteristics of the two student groups, including age, gender in pre-professional courses, revealed comparable learning abilities between the two groups (P > 0.05). Both groups received instruction in internal medicine. The students in both groups used the ninth edition of the textbook “Internal Medicine,” edited by Ge Junbo and others and published by People’s Medical Publishing House, and were taught by the same instructor.

Teaching interventions

Teaching mode of control group

The control group adopted the traditional teaching model, and the course arrangement was divided into two parts: theoretical classes and practical classes. In weekly theoretical classes, teachers use PPT to impart knowledge according to the teaching objectives and syllabus. The contents of these theoretical courses include basic knowledge of cardiovascular diseases, pathophysiology, diagnostic methods and treatment principles. Teachers help students understand complex medical concepts through detailed explanations and illustrations, and answer students’ questions in class to ensure they master the necessary theoretical knowledge.

In the practical class, the teacher led the students to conduct practical training based on the teaching content of the previous theoretical class. Practical classes were usually conducted in simulated wards or clinical skills laboratories. Teachers first demonstrated the operations on a standardized patient(SP), including specific operating steps such as cardiac examination, auscultation, and electrocardiogram interpretation. Teachers explained in detail the key points and precautions of each operation link, and demonstrated on-site how to communicate with patients to improve students’ clinical operation skills and doctor-patient communication abilities.

After the demonstration, students were divided into groups for operational exercises, with teachers guiding them, correcting mistakes in a timely manner and providing feedback. In this way, students not only consolidated theoretical knowledge, but also enhanced practical operational abilities and developed clinical thinking and decision-making abilities. In addition, practical courses also emphasized teamwork and communication skills. Students simulated real clinical environments through group discussions and role-playing to improve their overall quality and professional abilities.

Formation of teaching research team

The team of this study was composed of 2 chief physicians, 3 attending physicians, 2 resident physicians, 5 teaching assistants, and 4 graduate students. This team consisted of teachers with more than five years of teaching experience. Before the lectures, they all underwent training in scenario simulation teaching mode and were proficient in using ChatGPT.

Implementation plan for educational reform

The teaching model of the experimental group innovatively incorporated generative artificial intelligence technology, providing students with a brand new scene simulation teaching experience. In this teaching model, teachers first provided an in-depth explanation of theoretical knowledge to ensure that students could master the core points of the course, such as the characteristics of different types of arrhythmias in electrocardiograms. These points are the basis for understanding the complexity of cardiovascular disease and are the knowledge that students must skillfully apply in subsequent simulation practices.

Students then watched a video simulating scenarios related to cardiovascular disease. These videos not only vividly reproduced clinical scenes, but also contained rich medical information and situational challenges, which greatly stimulated students’ interest in learning and enthusiasm for participation. While watching the video, students were encouraged to play the role of doctors and use the theoretical knowledge they had learned to conduct detailed analysis and inferences on the signs, symptoms, and pathogenesis shown in the video.

Students needed to use critical thinking to identify the occurrence and development of the disease from the patient’s clinical manifestations and, at the same time, master the key points of diagnosis and the basic principles of treatment. This process not only exercises the students’ clinical thinking skills but also deepens their understanding of the disease diagnosis and treatment process.

After the scenario simulation, students participated in group discussions to share their observations and analyses, complementing each other and improving their understanding of the disease. This interactive learning method promoted the exchange of knowledge and the collision of ideas, helped students examine problems from different angles, and improved their problem-solving abilities.

Finally, students would complete thinking questions related to the course content, consolidate the knowledge they have learned, and test the learning effect. Students could ask ChatGPT questions at any time, and when they had more questions, they could get help from their teachers. Except for learning theoretical knowledge, all clinical practice processes were consistent with those of the control group.

Establishment of experimental group

Reasonable grouping is an important prerequisite for team learning. To enhance group learning and achieve optimal learning outcomes, each group had a maximum of 6 students. Therefore, before class, teachers determined the groups based on students’ average GPAs to ensure that each group had similar overall learning abilities. Eventually, the students in the experimental group were divided into 6 groups. Based on feedback from teachers on student performance, adjustments to group members were made in the first week. In each group, one student was selected as the group leader, responsible for organizing group activities. Clear division of team roles ensured the participation of each member and promoted cooperation within the group.

Preparation of scenario simulation videos

Writing scenario simulation scripts

The cardiovascular teaching research group wrote script stories based on teaching objectives and typical cardiovascular cases, enriching the background and character features of the plot to make it as close to real clinical situations as possible.

Breaking down script scenes

In the production of clinical case scenario simulation videos, the breakdown script played a crucial role, providing guidance and basis for AI drawing for each scene. By inputting the case directly into ChatGPT and instructing, “How many scenes can this script be broken down into for animation video creation?” ChatGPT would then offer a breakdown of scenes as an example, subject to review by the teachers for alignment with educational goals and accuracy.

Animation drawing

By inputting the prompt “I need you to act as the Midjourney command optimization master, generating scene descriptions for the above scenes separately, I want Midjourney to draw them, please provide concise descriptions in both Chinese and English,” specific instructions would be obtained. This prompt asks ChatGPT to generate a concise description for each scenario. These descriptions should include necessary details to help Midjourney draw the scene accurately. Each scene description was reviewed, and then each English description was input into Midjourney to generate animation materials. These materials were imported into editing software to complete the production of video content, with subtitles automatically generated and added to the video.

Question bank compilation

In the process of compiling a question bank for cardiovascular teaching, ChatGPT generates questions based on the plot content of the script when prompted with the instruction, “This is a case in cardiovascular teaching, what questions can be given to students?” ChatGPT would write questions based on the relevant plot content of the script. The teacher could continue to instruct to change the format and description of the questions and could also request answers and scoring criteria for the corresponding questions.

Synthesis of scenario simulation teaching videos and classroom teaching

The assessment of question and answer accuracy and scientific validity, the adjustment of question difficulty in alignment with teaching objectives, and the precise placement of questions within the video were carried out to finalize the production of cardiovascular scenario simulation teaching videos. Subsequently, these videos were integrated into the class app for classroom instruction. Feedback from both students and teachers was solicited to enhance the content and quality of the scenario simulation teaching videos(Fig. 2).

Fig. 2.

Flow chart of research on teaching reform programmes

Data collection

Post-class test

Students in both the experimental group and the control group took the post-class test, and the test content and grading criteria were exactly the same. The theoretical knowledge level and practical operational ability were each scored out of 100 points, with higher scores indicating more vital student abilities. The theoretical knowledge assessment used exam questions prepared by the teaching team, while practical operational ability used a “Mini-CEX” scoring sheet customized for cardiovascular medicine. The Mini-CEX evaluation form was adapted by the teaching and research team from a scale for assessing clinical skills written by John J Norcini et al. [21]. It is designed according to the characteristics of cardiovascular medicine. It mainly evaluates clinical history recording, electrocardiogram interpretation, humanistic care, Clinical diagnosis, communication skills and overall competency. There were five parts in total; each part had four questions, and each question adopted Likert’s five-point scoring system. The Cronbach’s alpha of the scale was 0.90, and the Cronbach’s alpha of each dimension was 0.753–0.772.

Clinical critical thinking scale

Based on Robert Ennis’s critical thinking framework and related theories, relevant questions were adapted according to the experimental purpose and subjects [22]. The final clinical critical thinking scale consisted of four dimensions, including logical reasoning, central argument, argumentation evidence and organizational structure, with a total of 5 questions in each dimension and 5 points in each question, for a total of 100 points.

Overall teaching satisfaction survey

The teaching and research team developed a teaching satisfaction questionnaire. Students completed the Teaching Satisfaction questionnaire on the WJX.cn at the end of the final exam. The questionnaire included six aspects: teaching scene satisfaction (Q1 ∼ Q4), course discussion satisfaction (Q5 ∼ Q8), group cooperation satisfaction (Q9 ∼ Q11), individual participation (Q12 ∼ Q14), teaching content satisfaction (Q15 ∼ Q18), and teaching teacher satisfaction (Q19 ∼ Q20). Each question was set on a scale of 1 to 5 (strongly disagree to strongly agree on five scales). Final satisfaction (%) is score/total score (100 points) *100%. After analyzing the preliminary collected data, Cronbach’s alpha coefficient was 0.85, indicating high internal consistency and reliability.

Qualitative assessment - semi-structured interviews

At the end of the course, we conducted a semi-structured interview to survey students in the experimental group and teachers on their evaluation of the use of AI in teaching cardiovascular disease. In selecting interviewees, we considered the gender and age and then conducted purposive sampling among the experimental group to ensure a diversity of opinions.

In order to fully understand the teaching effect and the real experience of teachers and students with the application of AI teaching mode, the research team first conducted preliminary interviews with two students and determined the final outline of the interview: (1) How do you feel about the learning of this teaching mode? (2) Do you think your learning/teaching style has changed since before? (3) What are your suggestions for the future development of this teaching mode?

A researcher who was well-versed in interviewing techniques was assigned to conduct the interviews independently. The interviews were conducted during the week following the course in a quiet and relaxing session to avoid errors as much as possible. Based on their final test results, they were divided into three grades, with three boys and three girls randomly selected from six groups from three different levels. Each interview lasted approximately 20 min. The students’ conversations were recorded using a voice recorder, and the research team pledged to keep them confidential. Recordings of the interviews were transcribed verbatim within 24 h of the end of the conversation.

Data analysis

Data entry and analysis were performed using Rstudio software (version 4.3.1). The following R packages were utilized: “stats”, “car”, “doBy”, and “ggplot2”.

For quantitative data, independent sample t-tests were employed to analyze differences between groups. For qualitative data, the chi-square test was utilized. A significance level of P < 0.05 was considered statistically significant, indicating differences between groups.

Results

Baseline comparison between two groups

The experimental group consisted of 34 students aged 22–24 years (mean age 23.03 ± 0.626). The Control group comprised 32 students from clinical professional classes, with ages ranging from 21 to 25 years (mean age 23.14 ± 0.976). Before the class, we assessed the basic clinical knowledge of the two groups of students, and the results showed that there was no significant difference in the demographic characteristics of the two groups (P > 0.05), and we found that there was no significant difference between them, which was comparable (Table 1).

Table 1.

Comparison of demographic characteristics between two groups

Characteristics	Experimental group (n = 34)	Control group (n = 32)	Statistics	P-value
Age	23.03 ± 0.626	23.14 ± 0.976	t = 0.477	0.635b
Gender
male	18	17	χ2 = 0.243	0.622a
female	16	15
Test before class	81.50 ± 5.48	82.72 ± 6.45	t = 0.830	0.410b

Note: a: Pearson’s chi-squared test; b: Independent-samples t-test

Final scores between two groups

Statistical analysis of examination scores for two groups revealed that students in the experimental group had an average score of 83.26 on the theoretical final exam, whereas students in the control group had an average score of 79.56. The scores of the control group were significantly lower than those of the experimental group (p < 0.05). Regarding Mini-CEX examination scores, students within the experimental group attained an average score of 76.24, which was notably greater than the average score of 70.19 achieved by students in the control group (p < 0.001). Furthermore, the clinical critical thinking proficiency of the experimental group surpassed that of the control group, as indicated by statistical significance (p < 0.001) (Table 2).

Table 2.

Comparison of final scores between two groups (± s)

Grouping	Theoretical Exam	Mini-CEX	Clinical critical thinking score
Experimental group (n = 34)	82.91 ± 5.054	76.24 ± 12.302	92.941 ± 2.131
Control group (n = 32)	72.16 ± 7.595	70.19 ± 11.255	89.312 ± 2.533
t	6.811	2.08	6.311
P-value	< 0.001	< 0.05	< 0.001

Satisfaction survey

After investigation and recovery, a total of 66 students completed the satisfaction questionnaire, and 66 valid questionnaires were recovered, with a total completion rate of 100%. As shown in the questionnaire results (Table 3), it can be seen that the overall satisfaction of experimental group in teaching scene, individual participation and teaching content is higher than that of control group, and the difference between the two groups is statistically significant (P < 0.05). There were no significant differences in other aspects (P > 0.05).

Table 3.

Comparison of teaching satisfaction between two groups (± s)

Group	Experimental group (n = 34)	Control group (n = 32)	t	P-value
Item
Teaching scene	18.029 ± 1.507	18.938 ± 1.134	2.752	0.008
Course discussion	18.059 ± 1.278	17.719 ± 1.171	1.125	0.265
Group cooperation	13.765 ± 1.130	13.469 ± 1.502	0.908	0.367
Individual participation	13.647 ± 1.454	12.531 ± 1.741	2.832	0.006
Teaching content	18.912 ± 0.900	18.094 ± 1.510	2.691	0.009
Teaching style of teachers	9.235 ± 0.781	9.281 ± 0.729	0.247	0.806
Final satisfaction(%)	91.647 ± 2.673	90.031 ± 2.776	2.409	0.019

Qualitative data analysis

In summarizing the interview findings, three primary themes emerge for analysis: (1) A new learning (teaching) experience; (2) Enhancement of clinical critical thinking ability; (3) Suggestions for improvement.

Theme 1: a new learning (teaching) experience

“In the past, we have always learned knowledge from books. Some things are very complicated and not easy to understand. With the help of AI, I think a lot of complicated knowledge has suddenly become simple and clear.”(S1).

“It is a very unimaginable experience. Through the scenario simulation course, I can intuitively see the physiological changes of the heart and blood vessels, and many theoretical knowledge are easier to understand.”(S2).

“The scenario simulation course enables us to visually see the electrophysiology and pathophysiological changes of the heart and blood vessels. Seeing the complete process makes it easier to remember and understand.”(S3).

“I’ve seen a lot of animations during the learning process, and through this method, I have a better understanding of clinical analysis and judgment.”(S4).

“I think the course preparation process is very easy, with the help of ChatGPT, many educational resources can be found quickly, and I am even more incredible that it can produce a complete clinical simulation video! I believe I will be able to perform better in the field of clinical teaching in the future!”(T1).

Theme 2: enhancement of clinical critical thinking ability

Leveraging AI in medical and educational fields, students can utilize AI interactive platforms to simulate disease processes, enhancing their understanding of cardiovascular diseases and developing critical thinking and problem-solving skills.(S1).

“With AI assistance, my knowledge becomes more systematic and detailed. For example, when learning about acute myocardial infarction, I saw numerous relevant images such as anatomical slices of coronary arteries, their distribution, and corresponding myocardial perfusion areas, which enhances our analytical and judgment abilities.”(S2).

“During leisure time, I can use AI interactive platforms for learning and engage in question-and-answer conversations with AI, which makes self-directed learning more effective and motivating.”(S5).

“I could see the students’ progress in their learning from the exercise tests at the end of the lesson and the final Mini-CEX exam. Through the communication and discussion with them after the lesson, I found that they became more logical in their thinking about the problems, and their ability to analyse the conditions during the Mini-CEX exam was greatly improved.”(T2).

Theme 3: concerns for improvement

Regarding the application of AI in cardiovascular medicine education, students and teachers actively provided some suggestions.

“This teaching format and content are vivid and illustrative. However, I feel that some content, when interacting with AI, cannot answer my questions well.”(S3).

“With this mode of teaching, I feel that I have a higher level of mastery of this course than any other subject and am more interested and motivated to learn. I have been very willing to use ChatGPT in other courses to assist me in my studies, but I felt slightly uncomfortable communicating with the AI as opposed to the teacher.”(S4).

“This way of preparing teaching materials and the mode of lectures is indeed very innovative, with the help of ChatGPT, my pre-course preparation process will be relatively easier, and the use of it in the classroom has also greatly improved the motivation of students. However, I am concerned that the drawbacks of AI, such as academic honesty and accuracy of answers, will also have an impact on the final teaching results, so we teachers should be cautious about AI.”(T2).

Discussion

With the rapid development of technology and AI, the form of medical education is undergoing continuous changes [23, 24]. Traditional teaching mode, characterized by inefficiency and dull content, no longer meet the needs of modern medical education. This is particularly evident in the teaching of cardiovascular system diseases [25], where the content is complex and difficult to remember, often leading to a lack of student engagement and understanding during clinical practice, thereby impacting the cultivation of clinical thinking skills [26]. Currently, AI is widely applied across various fields, and research shows that it plays a crucial role in education [27, 28], including personalized learning, intelligent tutoring, instructional design, and student assessment, greatly enhancing learning outcomes and promoting educational innovation. Moreover, studies have also shown the widespread promotion and application of scenario-based teaching models in clinical practice teaching [29–31].

In this study, the scenario-based teaching model is implemented based on ChatGPT 3.5. We believe that the scenario-based teaching model based on generative AI is an important mode and development direction for educational practice reform. ChatGPT, with its outstanding adaptability, versatility, efficiency, intelligence, and comprehensive coverage, has become a favored choice for many developers and is widely used in the education sector [32, 33]. Through clever integration with the scenario-based teaching model, a new teaching experience is created.

For teachers, ChatGPT provides powerful support, significantly improving lesson preparation efficiency. Teachers can use ChatGPT’s intelligently generated dialogue scenarios to present abstract and difficult-to-understand concepts in vivid and interesting scenarios, making it easier for students to understand and remember. Additionally, teachers can adjust the generated dialogues according to students’ learning situations, personalize teaching, and improve teaching effectiveness. For students, in the scenario-based teaching model, they feel as if they are in a vivid teaching theater. They take on detective roles, cultivating clinical thinking and case analysis skills as they solve problems. ChatGPT’s intelligent dialogue can also customize learning plans based on students’ learning styles and progress, improve memory efficiency through mnemonic devices, and stimulate their interest in learning and self-directed learning motivation.

The findings of the study indicate that students enrolled in AI-assisted teaching programs exhibit higher scores in theoretical knowledge, Mini-CEX examination performance, and clinical critical thinking skills compared to their counterparts in traditional teaching settings. These results suggest that a hybrid teaching approach may enhance students’ comprehension of knowledge and proficiency in clinical procedures, this is consistent with the findings of Yujiwang et al [34–37]. The possible reason is that in the interaction of scenario simulation, students can independently explore the process of illness and take the initiative to find and solve problems. According to Kolb’s experiential learning model [19, 20], experience to reflection to abstract concepts to practice, and finally to experience, interlocking and progressive, prompting students to understand knowledge from scenario simulation, then apply it to practice, and then find problems, which not only improves their independent learning ability, but also improves their critical thinking ability. Additionally, student interviews revealed that the new teaching method facilitates their exploration and identification of clinical issues, thereby preparing them effectively for future clinical practice.

Through an analysis of students’ teaching satisfaction questionnaires, it was found that the experimental group exhibited significantly higher levels of satisfaction in teaching scene, individual participation, and teaching content compared to the control group. These results suggest that the mixed teaching mode utilizing the AI platform may be more feasible and suitable for practical teaching in cardiovascular internal medicine. Although we found no statistically significant differences in course discussion, teamwork, and instructor teaching style, this may be due to the following reasons. First, the small sample size and short duration of this study limited the power to detect significant differences. Future research could improve this by increasing the sample size and extending the duration of the study. In addition, traditional teaching methods are already relatively mature in these aspects, and student satisfaction in these three aspects is already at a high level, and may not show significant advantages in the short term. At the same time, teaching satisfaction is affected by many factors, and a single change is not enough to significantly improve overall satisfaction. Therefore, we will continue to optimize the new AI-powered teaching model and strengthen its integration with course discussions and teamwork. We look forward to seeing more significant effects in future research.

Moreover, students say that the teaching method of scenario simulation not only helps them systematically understand and master the content of the course, but also stimulates their interest in independent learning and improves their ability to discover and solve problems. The vast majority of students hold a positive attitude towards the AI empowered scenario-based simulation teaching mode, and some students also put forward their own views on this teaching mode, mainly focusing on the accuracy and understanding of AI. This also provides us with valuable suggestions for the improvement of further study.

Limitation

This study also has the following limitations: (1) The number of participants in the survey is relatively small, resulting in insufficient data and interview views collected; (2) In this study, we used version 3.5 of generative AI ChatGPT. However, it is worth noting that a more advanced version 4.0 of ChatGPT is already available on the market. Therefore, the version we use does not fully represent the highest computing power of AI technology.

Conclusion

In comparison to the conventional teaching methodology, the novel teaching mode demonstrates clear benefits. Findings from examinations, assessments, satisfaction surveys, and interviews suggest that this innovative teaching method offers a more efficient means for interns to gain contemporary professional knowledge and enhance their clinical practice proficiency. Additionally, the cultivation of clinical critical thinking and problem-solving skills through this approach is expected to greatly support their long-term career viability. The utilization of an AI-empowered scenario-based simulation teaching mode has the potential to enhance students’ engagement and motivation, as well as improve their problem-solving skills in clinical settings. Consequently, the implementation and dissemination of our AI-empowered scenario-based simulation teaching mode in cardiovascular medicine practice teaching is recommended.

Author contributions

KLZ and HXZ designed the trial. KLZ prepared the clinical cases. HXZ collected the data. HXZ and ZYS analyzed the data. HXZ, ZHC and CC wrote the manuscript. All authors have read and approved the final manuscript.

Funding

Innovation and Entrepreneurship Training Program for College Students in Jiangsu Province (202313993027Y). Teaching Reform Research Project of Nantong University (2023B10).

Data availability

Our research encompasses sensitive personal identity information of students. Due to the potential risk of breaching individual privacy, the datasets analyzed in this study cannot be made publicly accessible. We emphasize that the data remains confidential and is not open to the public. However, if you have a compelling need for access, please reach out to the corresponding author at zhuhuixia@ntu.edu.cn to request the data.

Declarations

Ethics approval and consent to participate

This study has obtained ethical approval from the Ethics Committee of the Second Affiliated Hospital of Nantong University. All methods were conducted in accordance with relevant guidelines and regulations(approval number 2024KT045). All participation was voluntary and signed informed consent forms were obtained from each participant.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.