Integrating AI in Clinical Education: Evaluating General Practice Residents’ Proficiency in Distinguishing AI-Generated Hallucinations and Its Impacting Factors

Integrating AI in Clinical Education: Evaluating General Practice Residents’ Proficiency in Distinguishing AI-Generated Hallucinations and Its Impacting Factors | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Integrating AI in Clinical Education: Evaluating General Practice Residents’ Proficiency in Distinguishing AI-Generated Hallucinations and Its Impacting Factors Jiacheng Zhou, Jintao Zhang, Rongrong Wan, Xiaochuan Cui, Qiyu Liu, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5332750/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 19 Mar, 2025 Read the published version in BMC Medical Education → Version 1 posted 4 You are reading this latest preprint version Abstract OBJECTIVE To evaluate the ability of general practice residents to detect AI-generated hallucinations and assess the influencing factors. METHODS This multi-center study involved 142 general practice residents, all of whom were undergoing standardized general practice training and volunteered to participate. The study evaluated AI’s accuracy and consistency, along with the residents’ response time, accuracy, sensitivity(d’), and standard tendencies (β). Binary regression analysis was used to explore factors affecting the residents' ability to identify AI-generated errors. RESULTS 137 participants ultimately included had an mean (SD) age 25.93 ± 2.10, with 46.72% male, 81.75% undergraduates, and 45.26% from Jiangsu. Regarding AI, 52.55% were unfamiliar with it, 35.04% had never used it. ChatGPT demonstrated 80.8% overall accuracy, including 57% in professional practice. 87 AI-generated hallucinations were identified, primarily in the level of application and evaluation. The mean (SD) accuracy was 55% ±4.3%, and the mean (SD) sensitivity (d') was 0.39 ± 0.33. The median response bias (β) was 0.74 (0.31). Regression analysis revealed that shorter response times (OR = 0.92, P = 0.02), higher self-assessed AI understanding (OR = 0.16, P = 0.04), and frequent AI use (OR = 10.43, P = 0.01) were associated with stricter error detection criteria. CONCLUSIONS The study concluded that residents struggled to identify AI errors, particularly in clinical cases, emphasizing the importance of improving AI literacy and critical thinking for effective integration into medical education. AI-Generated Hallucinations General Practice Residents Standardized General Practice Training Response Bias Figures Figure 1 Figure 2 Introduction The application of artificial intelligence (AI) in medical education is becoming increasingly widespread, particularly in postgraduate training, showing significant potential for advancing educational practices 1 , 2 .In recent years, Generative Pre-trained Transformer (GPT) models, as an emerging technology, have been integrated into clinical teaching, with applications such as conducting clinical assessments, supporting self-directed learning for medical students, and creating virtual patient scenarios 3 – 5 . Studies indicated that GPT-assisted methods were comparable to traditional approaches in terms of difficulty and discriminative ability, while significantly reducing labor costs (by 57%) 3 .Additionally, large language models (LLMs) continue to evolve, demonstrating high accuracy (over 90%) in explaining medical problems and conducting clinical assessments 6 , 7 . AI has now permeated various facets of medical education and clinical practice. However, despite its potential to enhance learning efficiency and reduce educational costs, the field of medical education deals with matters concerning patient safety and the cognitive framework of healthcare professionals, necessitating a more cautious and scientifically rigorous approach to its development. 1 , 8 . Although Large Language Models (LLMs) is trained on high-quality datasets and improved accuracy through enhanced Retrieval-Augmented Generation (RAG), they may still generate errors in clinical content analysis 9 – 11 , known as "hallucinations" 12 , these hallucinations cannot be completely eliminated. In fact, the occurrence of hallucinations tends to be more common in larger LLMs.That appear logical but are often misleading 13 . Clinicians may misinterpret AI recommendations, particularly if the reasoning behind the AI’s decision-making is opaque or difficult to understand 14 . In critical fields such as healthcare, AI hallucinations and biases can pose significant risks, so it's crucial for medical professionals to identify and address these issues. However, most researches focused solely on enhancing large language models and minimizing hallucinations, but little attention has been given to the ability of general practice residents to recognise inaccuracies during self-directed learning, as well as the factors that influence this capability. As AI becomes more prevalent in clinical settings,medical students will be confronted with clinical content and question-and-answer information spanning multiple disciplines, which they must interpret while guarding against AI biases 15 , 16 . General practice residents play a crucial role in managing patient health and diagnosing diseases at an early stage. This requires not only a solid foundation of medical knowledge and skills but also a reliable cognitive framework and the ability to discern erroneous information 17 . However, many doctors lack confidence in their ability to detect AI-generated errors 18 . Given that primary care physicians handle a wide range of clinical scenarios, understanding how they identify AI hallucinations is essential, yet research on this topic remains limited. Our study evaluated general practice residents' ability to recognize and manage AI-generated hallucinations and investigated factors influencing this ability, such as AI usage frequency, clinical experience, and AI familiarity. The research seeks to explore key factors to improve residents' abilities to identify AI-generated inaccuracies during standardised general practice training, which also aims to provide valuable insights for the development of future educational frameworks that integrate AI with clinical training. Methods A cross-sectional survey using ChatGPT (GPT-4) simulated responses for the 2024 Intermediate Attending Physician Examination in General Medicine 19 . It consisted of four sections: three with 100 single-choice questions each and one with 18 clinical cases and 86 multiple-choice questions. Two senior general practitioners reviewed and assessed the consistency and validity of ChatGPT’s responses and explanations. General practice resident doctors identified correct and incorrect answers and explanations. Ethical approval has been obtained for this study, and all participants have provided informed consent ( Figure1 ). Determining characteristics of questions Miller's pyramid, proposed in 1989, outlined four levels of medical education assessment: Knowledge (Knows), Comprehension (Knows How), Application (Shows How), and Evaluation (Does) 20 . Key assessment strategies for evaluating learners' clinical reasoning abilities were outlined, spanning from knowledge acquisition to real-world application in the clinical setting 21 .At the same time e-assessment methods could replace currently conventional methods 22 . This study categorized original questions according to Miller’s pyramid to assess specific medical competencies and analyze how different cognitive levels affect the accuracy of ChatGPT-generated responses and participants' ability to identify errors. Two senior general practitioners independently applied this taxonomy to ensure consistency, resolving discrepancies through discussion until a consensus was reached. P rompt for ChatGPT's answers and explanations Prompt engineering has highlighted the potential of large language models (LLMs) as effective tools in clinical medicine 23 . To optimize the value of the responses generated by these models, we applied the ROT (Relevance, Objectives, Tasks) framework, tailored to the context and specific goals of the task 24 . Relevance: Clearly define the domain of the query, such as "Assume the role of an experienced clinician." Objectives: Articulate the aim of the query, for example, "This is a mid-level medical examination; respond to the questions provided with your best effort." Tasks: Specify the exact task to be performed, such as "Offer detailed explanations." This framework was designed to improve the relevance and quality of responses generated by LLMs in clinical scenarios, enhancing their effectiveness as tools in medical education and practice. Participant Settings General Practice Residents were involved to assess the effectiveness of hallucination recognition. Inclusion criteria : Residents from Wuxi People's Hospital in Jiangsu and the first affiliated hospital of Jiamusi University, the second affiliated hospital of Harbin medical university and the second affiliated hospital of Qiqihar medical university in Heilongjiang undergoing standardized general practice training, who voluntarily participate in the study. Exclusion criteria : who did not complete the training as scheduled, those unwilling to participate or providing inaccurate data. Detection of Hallucinations Signal Detection Theory (SDT) was employed to analyze the experimental data 25 . The errors and correct items were numbered and randomly assigned using the random function in Excel, with the subjects categorized as follows: "Basic Knowledge," "Related Professional Knowledge," and "Professional Knowledge" as subjects 1-3, and "Professional Practice" as subject 4. Participants completed the questions using a standardized approach, and their responses were collected through an electronic survey. Each participant selected one of three options (Correct, Incorrect, or Uncertain) to assess the accuracy of each response and explanation. Sensitivity (d’): Measures the participant's ability to correctly identify true positives. It was calculated as Z(Hit Rate) - Z(False Alarm Rate). Response Bias (β): Evaluates the tendency of participants to favor one response option over others. Categorized into two groups depending on whether the β value is equal or greater than 1, or less than 1. Statistical analysis All statistical analyses were performed using IBM SPSS Statistics 26.0. For normally distributed data, use t test for two groups and one-way ANOVA for multiple comparisons, with LSD or Dunnett's test for pairwise comparisons. For categorical data, use chi-squared tests and the Bonferroni correction to adjust for multiple comparisons among group rates. Binary logistic regression analysis was performed to explore factors influencing participants' response bias. All statistical tests were two-sided, with a p-value below 0.05 considered statistically significant. Results Characteristic among General Practice Resident s After an initial screening of the 142 collected questionnaires, 5 with all correct answers were excluded. The average age of the participants was 25.93. The participants consisted of 46.72% males, 81.75% undergraduates, and 45.26% from Jiangsu. The group comprised 41 participants from the class of 2021, 43 from 2022, and 53 from 2023. Regarding the Occupational Medical Examination, 13 individuals were not passed, and 73 successfully passed. As for familiarity and usage of AI, 52.55% were unfamiliar with AI, 35.04% had never applied AI, and only 8.03% used AI frequently in academic research. The average satisfaction score for AI usage was 7.0 on a 0-10 scales. ChatGPT's H allucinations Of Different Categorization of Medical questions In our study, hallucinations were discovered by two senior general practitioners via verifying each question's answers and explanations. Out of the 386 questions answered by ChatGPT, 312 answers were correct, while 74 answers were incorrect. Out of the 312 correct answers, the explanations for 299 were accurate, therefore, there were ultimately 87 incorrect explanations ( Figure1 ). ChatGPT's overall accuracy was 80.8% before human correction, dropping slightly to 80.05%. For knowledge-level questions, accuracy fell slightly from 90.29% to 89.32%, and for comprehension and application, it dropped from 86.61% to 86.61% and from 83.21% to 77.10%, respectively. Evaluation-level questions saw a sharp decline from 43.48% to 30.43% after correction ( Table1 ). The majority of the incorrect explanations are comprised of factual inaccuracies and meaningless responses. Hallucinations are especially evident on the cognitive layers of application and evaluation, which were key elements in cultivation of clinical reasoning for general practitioners. Recognition of AI-Generated Hallucinations by General Practice Resident s Our study found the median hit rate, false alarm rate, correct rejection rate, and miss rate for general practice resident doctor in identifying generative hallucinations were 86% (15%), 74% (21%), 26% (21%), and 14% (21%), respectively. This indicates a high false alarm rate and a low correct rejection rate. Overall accuracy was only 55%, with sensitivity at 0.39 ± 0.33. Additionally, 87% of participants took a more permissive approach in evaluating hallucinations. In addition to the overall recognition results, a stratified analysis by subject showed a median hit rate of 98% for Subject 1 and 87% for Subject 4. The interquartile range (IQR) , measured between the 25th and 75th percentiles, spanned from 73% to 91% for Subjects 1-3 and from 81% to 95% for Subject 4. False alarm rates had medians of 89% for Subjects 1-3 and 34% for Subject 4, with IQRs of 54% to 77% and 73% to 92%, respectively. For correct rejection rates, medians were 34% for Subjects 1-3 and 14% for Subject 4, with IQRs of 23% to 46% and 8% to 27%. The median miss rate was 17% for Subjects 1-3 and 11% for Subject 4, with IQRs of 9% to 28% and 33% to 80%. Uncertainty rates had medians of 57% for Subject 1 and 53% for Subject 4, with narrow IQRs of 0% to 9% and 0% to 8%, respectively. Comparison of accuracy, sensitivity, and response propensity revealed that Subject 4 had lower accuracy than the overall average ( P < 0.05) and was significantly less accurate than Subjects 1-3 ( P < 0.001). As cognitive levels advance, accuracy progressively decreased and this trend was statistically significant ( P < 0.001). Subject 4 also had notably lower sensitivity compared to Subjects 1-3 ( P < 0.05). As cognitive stratification progresses, sensitivity increases progressively, with all differences being statistically significant except between the knowledge and comprehensive groups, and the application and evaluation groups ( P < 0.001). In Subject 4, there are more people choosing neutral and conservative standards compared to the overall group and Subjects 1-3(all P < 0.001). Furthermore, more people choose strict standards for evaluation cognitive layer compared to the overall and the application level groups (all P < 0.001) ( Figure2A-D ). Factors impact respond bias of identification of hallucinations This study investigated factors influencing general practice resident doctors' choice of liberal, neutral, or conservative criteria for assessing GPT-generated responses. Among the variables analyzed, only AI usage frequency showed significant differences ( P = 0.02). Factors such as age, gender, degree, institution, grade, AI familiarity, use cases, and user satisfaction had no significant impact. Participants with stricter criteria tended to use AI more frequently ( Table 2 ). Table 2 Factors impact Response Bias of identification of hallucinations among General Practice Residents Characteristic Liberal (n=119) Neutral+ Conservative (n=18) Value (t or X 2 ) P age 25.87 + 2.10 26.33 + 2.01 -0.87 0.39 Gender (male%) 54(45%) 10(55.60%) 0.65 0.42 Degree(undergraduate%) 96(80.67) 16(88.9) 0.71 0.40 Institution(Jiangsu%) 54(45.38) 8(44.44) 0.01 0.94 Grade (2021%) 37(31.09) 4(22.2) 1.18 0.55 Occupational Physician Examination (unpass%) 52(43.7) 12(66.7) 3.31 0.07 Knowledge of AI (unknown%) 63(52.94) 9(50.0) 0.05 0.81 Application of AI Utilized (no%) 38(31.93) 11(61.1) 5.80 0.02* AI use cases (Academic%) 10(8.40) 1(5.6) 6.82 0.15 User satisfaction 7.09±1.52 6.13±1.12 2.22 0.05 Response time 72.82±6.84 67.61±15.21 1.43 0.17 Binary regression analysis identified response time, AI familiarity, and AI usage frequency as key factors affecting the choice of a neutral and stricter criteria. Specifically, shorter response times, less familiarity with AI, and frequently AI usage were linked to stricter criteria ( P = 0.02; P = 0.04; P = 0.01, respectively)( Table 2 and Table 3 ). Table 3 Binary logistics regression Analysis of Factors impact neutral and Conservative Response Bias Factor B Exp(B) 95%CI P Response Time -0.08 0.92 0.86-0.99 0.02* Knowledge of AI Known -1.09 0.16 0.03-0.88 0.04* unknown 1[Reference] Application of AI Utilized 2.35 10.43 1.57-62.05 0.01* unused 1[Reference] Age 0.05 1.06 0.78-1.42 0.73 Gender female 0.54 1.72 0.54-5.44 0.36 male 1[Reference] Institution Heilongjiang 0.38 1.45 0.34-6.22 0.62 Wuxi/Jiangsu 1[Reference] Degree graduate 0.38 1.46 0.25-8.54 0.68 undergraduate Grade (2021) 2023 0.22 1.25 0.16-9.98 0.84 2022 -0.23 0.79 0.10-6.20 0.79 2021 1[Reference] Occupational Physician Examination pass 1.19 3.30 0.58-18.64 0.18 not passed 1[Reference] Discussion To evaluate the application value of ChatGPT in various clinical scenarios in medical education, we utilized Miller's Pyramid to categorize these questions, assessing different levels of medical competency. The findings indicated that GPT is relatively accurate in handling basic clinical knowledge and problems, even after reviewing by experienced general practitioners. However, its performance deteriorated significantly (44.2%-57.0%) when dealing with more complex clinical cases or scenarios. The results showed that ChatGPT demonstrated a generally high accuracy rate, but its accuracy is lower in the domain of professional practice, indicating that it was more suitable for judgment and guidance on basic knowledge and simpler clinical problems. Our research further examined the ability of general practice residents to recognize AI-generated hallucinations. The results revealed a concerning trend: residents exhibit significant difficulty in identifying AI-generated hallucinations, with an overall accuracy rate of only 55.0%±4.3%. They showed relatively high sensitivity in recognizing basic medical knowledge, understanding, and application cognitive levels but tend to adopt more lenient standard as response time increases. There were no notable differences in outcomes among general practice residents across different training years and regions. All had limited ability to detect AI-generated hallucinations, especially in complex clinical cases. Key risk factors included frequent AI use and limited understanding of how it works, which led to more stringent evaluation standards. This underscored the importance of understanding AI’s fundamentals and limitations, reducing over-reliance on it, and enhancing clinical reasoning and critical thinking 26 – 28 . These were crucial steps for improving residents' ability to identify hallucinations and essential for integrating ChatGPT effectively into medical education. Residents undergoing standardized general practice training encounter AI-generated clinical data and support information throughout chronic disease management 29 , 30 , necessitating the integration, understanding, and coordination of patient information from all healthcare professionals, environments, and timelines 31 . As advanced AI technologies and applications emerged, it was crucial to critically evaluate their impacts, especially potential pitfalls. Despite consciously applying relatively strict criteria in clinical cases, the accuracy of identifying AI-generated errors significantly declined, nearly resembling random guessing. Participants who perceived themselves as knowledgeable about AI tended to adopt lenient standards, while those who frequently used AI preferred stricter criteria, likely due to their prior experiences. Participants commonly applied AI in personal and research contexts, where AI provided relatively reliable feedback. Errors in these domains incurred minimal time and financial costs. However, when AI was applied in clinical settings, it activated participants' sense of medical responsibility and their subconscious awareness of potential impacts on patients 32 .Our study also found that as the complexity of cognitive tasks increased, sensitivity in recognizing AI hallucinations tended to rise. Our findings are consistent with earlier studies on AI in healthcare and clinical decision support systems, while also bringing to light new challenges. AI has shown substantial benefits in enhancing diagnostic precision and offering treatment recommendations grounded in evidence 33 , 34 . For single-text discharge summaries, AI could function without hallucinations 35 . In scenarios involving creative tasks, hallucinations were more likely to occur 36 .Our findings built on previous research, revealing that AI systems encounter significant challenges when processing complex, multidimensional clinical data. These hallucinations were not only common in large-scale medical AI models but also proved especially difficult for residents undergoing standardized general practice training to identify 37 – 40 , which increased the risk of misdiagnosis and clinical errors 41 . Primary care physicians tended to accept plausible yet incorrect answers without much skepticism and seldom questioned AI outputs. Although AI has demonstrated diagnostic accuracy on par with, or even surpassing, that of experienced doctors 42 , 43 , it's crucial to remain mindful of cognitive limitations, AI-driven biases, and the risk of illusory understanding 44 , 45 . Some researchers have proposed that focusing on decision-making grounded in relative uncertainty, rather than simply aiming to minimize uncertainty, offers a more insightful approach to applying AI in medicine 46 . This suggested that when using AI, we should embrace and manage uncertainty instead of placing undue trust in AI's seemingly conclusive outputs. Excessive dependence on AI risked undermining clinical intuition, potentially relegating clinicians to simply following algorithm-driven recommendations without exercising critical judgment 47 . Failure to recognize hallucinations may prolong decision times and lead to erroneous analyses 48 , 49 . As highlighted by the WHO guidelines on the ethics and governance of AI in healthcare, it was crucial to ensure data transparency, interpretability, and accountability while fostering a sense of responsibility among clinicians 50 . This is also a key concern that educators in medical education need to pay close attention to. Strengths Our study had several strengths. The study population consisted of residents undergoing standardized general practice training, a critical stage in medical education. Moreover, our research not only evaluated AI performance but also examined resident doctors' ability to identify AI hallucinations, offering a comprehensive view of AI's practical application in clinical settings. They often applied lenient recognition standards with low sensitivity to AI responses, highlighting the need for specialized training to enhance critical thinking and evaluation skills. Additionally, general practice residents were more susceptible to hallucinations compared to experienced physicians, but they might also benefit more from AI support 14 , 51 .The low recognition accuracy calls for stricter AI usage guidelines and clearer uncertainty markers in AI systems to help distinguish valid recommendations from potential errors. Limitations However, there were also several limitations to our study. Firstly, the reliance on a specific AI system (GPT) may limit the generalizability of the findings to other AI models. Secondly, the assessment, conducted through closed-ended questions, ranging from basic knowledge to complex clinical scenarios, may not fully reflect real clinical situations. Conclusion General practice residents have limited ability to recognize AI-generated hallucinations, particularly in complex clinical scenarios. As the frequency and unfamiliarity of AI use increases, residents tended to apply stringent assessment standards, reflecting an over reliance on AI. Declarations Ethical approval has been obtained for this study by Research Ethics Committee of Wuxi People’s Hospital, and all participants have provided informed consent. Acknowledgements We are very grateful to all the residents who participated in this research. Author Contributions: Yunyun Zhang and Zhiyong Zhang had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Conceive and design: Jiacheng Zhou, Xiaochuan Cui, Yunyun Zhang and Zhiyong Zhang. Acquisition, analysis, or interpretation of data: All authors. Statistical analysis: Jintao Zhang and Yunyun Zhang. Drafting of the manuscript: Jiacheng Zhou and Yunyun Zhang. Obtained funding: Yunyun Zhang and Jiacheng Zhou. Supervision: Yunyun Zhang and Zhiyong Zhang Funding: This study was funded by the "Top Talent Support Program for Young and Middle-aged people of Wuxi Health Committee(BJRC-8 and HB2023015, Dr Zhang); Taihu Light Basic Research Project, Wuxi Municipal Science and Technology Bureau(K20221023, Dr Zhang);Emerging Discipline Leader Program (2024-YZ-HBDTR-ZYY-2024, Dr Zhang); Wuxi Association for Science and Technology (KX-24-C154; Jiacheng Zhou) Data Sharing Statement: The data supporting the findings of this study are available from the corresponding author upon reasonable request. Consent for publication The authors provide their consent for the publication of this article. Conflict of Interest Disclosures : The authors declare no competing interests. References Mekki YM, Zughaier SM. Teaching artificial intelligence in medicine. Nat Rev Bioeng. 2024;2(6):450–1. 10.1038/s44222-024-00195-0 . Yan M, Cerri GG, Moraes FY. ChatGPT and medicine: how AI language models are shaping the future and health related careers. Nat Biotechnol. 2023;41(11):1657–8. 10.1038/s41587-023-02011-3 . Lam G, Shammoon Y, Coulson A, et al. Utility of large language models for creating clinical assessment items. Med Teach Published online August. 2024;26:1–5. 10.1080/0142159X.2024.2382860 . Shang L, Li R, Xue M, Guo Q, Hou Y. Evaluating the application of ChatGPT in China’s residency training education: An exploratory study. Med Teach Published online July. 2024;12:1–7. 10.1080/0142159X.2024.2377808 . Cook DA. Creating virtual patients using large language models: scalable, global, and low cost. Med Teach Published online July 11, 2024:1–3. 10.1080/0142159X.2024.2376879 Chaurasia AK, Greatbatch CJ, Hewitt AW. Diagnostic Accuracy of Artificial Intelligence in Glaucoma Screening and Clinical Practice. J Glaucoma. 2022;31(5):285. 10.1097/IJG.0000000000002015 . Chavez H, Chavez-Arias B, Contreras-Rosas S, Alvarez-Rodríguez JM, Raymundo C. Artificial neural network model to predict student performance using nonpersonal information. Front Educ. 2023;8. 10.3389/feduc.2023.1106679 . Lambert SI, Madi M, Sopka S, et al. An integrative review on the acceptance of artificial intelligence among healthcare professionals in hospitals. NPJ Digit Med. 2023;6(1):111. 10.1038/s41746-023-00852-5 . Kresevic S, Giuffrè M, Ajcevic M, Accardo A, Crocè LS, Shung DL. Optimization of hepatological clinical guidelines interpretation by large language models: a retrieval augmented generation-based framework. NPJ Digit Med. 2024;7(1):102. 10.1038/s41746-024-01091-y . Ge J, Sun S, Owens J, et al. Development of a liver disease-specific large language model chat interface using retrieval-augmented generation. Hepatol Published online March. 2024;7. 10.1097/HEP.0000000000000834 . Giuffrè M, Kresevic S, Pugliese N, You K, Shung DL. Optimizing large language models in digestive disease: strategies and challenges to improve clinical outcomes. Liver Int. 2024;44(9):2114–24. 10.1111/liv.15974 . Herrmann-Werner A, Festl-Wietek T, Holderried F, et al. Assessing ChatGPT’s Mastery of Bloom’s Taxonomy Using Psychosomatic Medicine Exam Questions: Mixed-Methods Study. J Med Internet Res. 2024;26:e52113. 10.2196/52113 . Tran CG, Chang J, Sherman SK, De Andrade JP. Performance of ChatGPT on American Board of Surgery In-Training Examination Preparation Questions. J Surg Res. 2024;299:329–35. 10.1016/j.jss.2024.04.060 . Jussupow E, Spohrer K, Heinzl A, Gawlitza J. Augmenting Medical Diagnosis Decisions? An Investigation into Physicians’ Decision-Making Process with Artificial Intelligence. Inform Syst Res. 2021;32(3):713–35. 10.1287/isre.2020.0980 . Chen D, Parsa R, Hope A, et al. Physician and Artificial Intelligence Chatbot Responses to Cancer Questions From Social Media. JAMA oncology. Published online May. 2024;16:e240836. 10.1001/jamaoncol.2024.0836 . Goodman RS, Patrinely JR, Stone CA Jr, et al. Accuracy and Reliability of Chatbot Responses to Physician Questions. JAMA Netw Open. 2023;6(10):e2336483. 10.1001/jamanetworkopen.2023.36483 . Buck C, Doctor E, Hennrich J, Jöhnk J, Eymann T. General Practitioners’ Attitudes Toward Artificial Intelligence–Enabled Systems: Interview Study. J Med Internet Res. 2022;24(1):e28916. 10.2196/28916 . Hantel A, Walsh TP, Marron JM, et al. Perspectives of Oncologists on the Ethical Implications of Using Artificial Intelligence for Cancer Care. JAMA Netw Open. 2024;7(3):e244077. 10.1001/jamanetworkopen.2024.4077 . Du QF, Wang JJ. 2024 General Medicine Practice Mock Exam. People’s Medical Publishing House; 2023: 3–39. ISBN: 9787117355421. Ten Cate O, Carraccio C, Damodaran A, et al. Entrustment Decision Making: Extending Miller’s Pyramid. Acad Med. 2021;96(2):199–204. 10.1097/ACM.0000000000003800 . Thampy H, Willert E, Ramani S. Assessing Clinical Reasoning: Targeting the Higher Levels of the Pyramid. J Gen Intern Med. 2019;34(8):1631–6. 10.1007/s11606-019-04953-4 . Hasani H, Khoshnoodifar M, Khavandegar A, et al. Comparison of electronic versus conventional assessment methods in ophthalmology residents; a learner assessment scholarship study. BMC Med Educ. 2021;21(1):342. 10.1186/s12909-021-02759-9 . Meskó B. Prompt Engineering as an Important Emerging Skill for Medical Professionals: Tutorial. J Med Internet Res. 2023;25:e50638. 10.2196/50638 . Wang L, Chen X, Deng X, et al. Prompt engineering in consistency and reliability with the evidence-based guideline for LLMs. NPJ Digit Med. 2024;7(1):41. 10.1038/s41746-024-01029-4 . Aujla H. d[Formula: see text]: Sensitivity at the optimal criterion location. Behav Res Methods. 2023;55(5):2532–58. 10.3758/s13428-022-01913-5 . Larson BZ, Moser C, Caza A, Muehlfeld K, Colombo LA. Critical Thinking in the Age of Generative AI. AMLE. 2024;23(3):373–8. 10.5465/amle.2024.0338 . Moulin TC. Learning with AI Language Models: Guidelines for the Development and Scoring of Medical Questions for Higher Education. J Med Syst. 2024;48(1):45. 10.1007/s10916-024-02069-9 . Suriano R, Plebe A, Acciai A, Fabio RA. Student interaction with ChatGPT can promote complex critical thinking skills. Learn Instruction. 2025;95:102011. 10.1016/j.learninstruc.2024.102011 . Wang S, Shi Y, Sui M, et al. Telephone follow-up based on artificial intelligence technology among hypertension patients: Reliability study. J Clin Hypertens (Greenwich). 2024;26(6):656–64. 10.1111/jch.14823 . Li J, Guan Z, Wang J, et al. Integrated image-based deep learning and language models for primary diabetes care. Nat Med Published online July. 2024;19. 10.1038/s41591-024-03139-8 . Everson J, Hendrix N, Phillips RL, Adler-Milstein J, Bazemore A, Patel V. Primary Care Physicians’ Satisfaction With Interoperable Health Information Technology. JAMA Netw Open. 2024;7(3):e243793. 10.1001/jamanetworkopen.2024.3793 . Simmenroth-Nayda A, Marx G, Lorkowski T, Himmel W. Working as simulated patient has effects on real patient life - Preliminary insights from a qualitative study. GMS J Med Educ. 2016;33(3):Doc42. 10.3205/zma001041 . Buch VH, Ahmed I, Maruthappu M. Artificial intelligence in medicine: current trends and future possibilities. Br J Gen Pract. 2018;68(668):143–4. 10.3399/bjgp18X695213 . Li J, Zhou L, Zhan Y, et al. How does the artificial intelligence-based image-assisted technique help physicians in diagnosis of pulmonary adenocarcinoma? A randomized controlled experiment of multicenter physicians in China. J Am Med Inf Assoc. 2022;29(12):2041–9. 10.1093/jamia/ocac179 . Tung JYM, Gill SR, Sng GGR, et al. Comparison of the Quality of Discharge Letters Written by Large Language Models and Junior Clinicians: Single-Blinded Study. J Med Internet Res. 2024;26:e57721. 10.2196/57721 . Zaretsky J, Kim JM, Baskharoun S, et al. Generative Artificial Intelligence to Transform Inpatient Discharge Summaries to Patient-Friendly Language and Format. JAMA Netw open. 2024;7(3):e240357. 10.1001/jamanetworkopen.2024.0357 . Mirkin S, Albensi BC. Should artificial intelligence be used in conjunction with Neuroimaging in the diagnosis of Alzheimer’s disease? Front Aging Neurosci. 2023;15. 10.3389/fnagi.2023.1094233 . Schubert MC, Lasotta M, Sahm F, Wick W, Venkataramani V. Evaluating the Multimodal Capabilities of Generative AI in Complex Clinical Diagnostics. Published online November 2, 2023:2023.11.01.23297938. 10.1101/2023.11.01.23297938 Brown C, Nazeer R, Gibbs A, et al. Breaking Bias: The Role of Artificial Intelligence in Improving Clinical Decision-Making. Cureus. 2023;15. 10.7759/cureus.36415 . Rueckel J, Huemmer C, Fieselmann A, et al. Pneumothorax detection in chest radiographs: optimizing artificial intelligence system for accuracy and confounding bias reduction using in-image annotations in algorithm training. Eur Radiol. 2021;31(10):7888–900. 10.1007/s00330-021-07833-w . Levinson AV, Goyal A, Man RHC, et al. Using Audio Data to Facilitate Depression Risk Assessment in Primary Health Care. Published online Oct. 2023;16. 10.48550/arXiv.2310.10928 . Rauschecker AM, Rudie JD, Xie L, et al. Artificial Intelligence System Approaching Neuroradiologist-level Differential Diagnosis Accuracy at Brain MRI. Radiology. 2020;295(3):626–37. 10.1148/radiol.2020190283 . Shen J, Zhang CJP, Jiang B, et al. Artificial Intelligence Versus Clinicians in Disease Diagnosis: Systematic Review. JMIR Med Inf. 2019;7(3):e10010. 10.2196/10010 . Messeri L, Crockett MJ. Artificial intelligence and illusions of understanding in scientific research. Nature. 2024;627(8002):49–58. 10.1038/s41586-024-07146-0 . Why scientists trust AI too much. - and what to do about it. Nature. 2024;627(8003):243. 10.1038/d41586-024-00639-y . Harish V, Morgado F, Stern AD, Das S. Artificial Intelligence and Clinical Decision Making: The New Nature of Medical Uncertainty. Acad Med. 2021;96(1):31–6. 10.1097/ACM.0000000000003707 . Liu T, Duan Y. Beware the self-fulfilling prophecy: enhancing clinical decision-making with AI. Crit Care. 2024;28(1):276. 10.1186/s13054-024-05062-3 . Triberti S, Durosini I, Pravettoni G. A Third Wheel Effect in Health Decision Making Involving Artificial Entities: A Psychological Perspective. Front Public Health. 2020;8. 10.3389/fpubh.2020.00117 . Jacobs M, Pradier MF, McCoy TH, Perlis RH, Doshi-Velez F, Gajos KZ. How machine-learning recommendations influence clinician treatment selections: the example of antidepressant selection. Transl Psychiatry. 2021;11(1):1–9. 10.1038/s41398-021-01224-x . Ethics and governance of artificial intelligence for health. Accessed September 14. 2024. https://www.who.int/publications/i/item/9789240029200 Wang W, Gao G (Gordon), Agarwal R, editors. Friend or Foe? Teaming Between Artificial Intelligence and Workers with Variation in Experience. Management Science. 2024;70(9):5753–5775. 10.1287/mnsc.2021.00588 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 19 Mar, 2025 Read the published version in BMC Medical Education → Version 1 posted Editorial decision: Revision requested 28 Oct, 2024 Editor assigned by journal 28 Oct, 2024 Submission checks completed at journal 28 Oct, 2024 First submitted to journal 25 Oct, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5332750","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":371529125,"identity":"eb23838b-4440-4423-a130-30826659fd29","order_by":0,"name":"Jiacheng Zhou","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jiacheng","middleName":"","lastName":"Zhou","suffix":""},{"id":371529126,"identity":"41db35b9-a119-4eed-be3c-411c3eafeaae","order_by":1,"name":"Jintao Zhang","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jintao","middleName":"","lastName":"Zhang","suffix":""},{"id":371529127,"identity":"1745ae25-b3ba-4ce0-9319-f466f2ec0325","order_by":2,"name":"Rongrong Wan","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Rongrong","middleName":"","lastName":"Wan","suffix":""},{"id":371529128,"identity":"a9689806-6f27-4977-aebe-73c31cfcb838","order_by":3,"name":"Xiaochuan Cui","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaochuan","middleName":"","lastName":"Cui","suffix":""},{"id":371529129,"identity":"f5d558c1-c4fb-4795-86d3-e134639b6deb","order_by":4,"name":"Qiyu Liu","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qiyu","middleName":"","lastName":"Liu","suffix":""},{"id":371529130,"identity":"ad70b5d0-f692-4b6e-b241-4aee85c629e7","order_by":5,"name":"Hua Guo","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hua","middleName":"","lastName":"Guo","suffix":""},{"id":371529131,"identity":"062acc90-b78c-4965-81af-0fab2934d2e3","order_by":6,"name":"Xiaofen Shi","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaofen","middleName":"","lastName":"Shi","suffix":""},{"id":371529132,"identity":"3cd0bdbf-0c70-42a5-b314-eaa060c84e04","order_by":7,"name":"Bingbing Fu","email":"","orcid":"","institution":"The First Affiliated Hospital of Jiamusi University","correspondingAuthor":false,"prefix":"","firstName":"Bingbing","middleName":"","lastName":"Fu","suffix":""},{"id":371529133,"identity":"c15d9117-eaac-459b-82e9-56a935b68493","order_by":8,"name":"Jia Meng","email":"","orcid":"","institution":"The Second Affiliated Hospital of Harbin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jia","middleName":"","lastName":"Meng","suffix":""},{"id":371529134,"identity":"2e321472-2716-458e-9c08-2296fc00b038","order_by":9,"name":"Bo Yue","email":"","orcid":"","institution":"The Second Affiliated Hospital of Qiqihar Medical University","correspondingAuthor":false,"prefix":"","firstName":"Bo","middleName":"","lastName":"Yue","suffix":""},{"id":371529135,"identity":"5d1d65a3-a1bf-4e92-9007-f49bcd24f154","order_by":10,"name":"Yunyun Zhang","email":"","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yunyun","middleName":"","lastName":"Zhang","suffix":""},{"id":371529136,"identity":"9187508d-d062-4123-9d7d-8d9686aa466f","order_by":11,"name":"Zhiyong Zhang","email":"data:image/png;base64,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","orcid":"","institution":"The Affiliated Wuxi People's Hospital of Nanjing Medical University","correspondingAuthor":true,"prefix":"","firstName":"Zhiyong","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2024-10-25 13:38:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5332750/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5332750/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12909-025-06916-2","type":"published","date":"2025-03-19T15:57:19+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":69420701,"identity":"98526d98-d72e-4da4-aabd-9aa12af26f16","added_by":"auto","created_at":"2024-11-20 07:44:47","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1249513,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExperimental Flowchart\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5332750/v1/45e1113d2eab7d585f0c6b2e.jpg"},{"id":69420699,"identity":"54763c4e-3ced-4b92-bef6-835ea145fc99","added_by":"auto","created_at":"2024-11-20 07:44:47","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":309664,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(A-D) Sensitivity, accuracy and propensity to response of hallucinations among General Practice Residents.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe analysis focused on the differences in sensitivity (Figure2A) , accuracy (Figure2B), and response tendencies (Figure2C and 2D) between the overall group and individual subjects, as well as across different cognitive levels(*P ＜0.05; **P ＜ 0.01; ***P ＜0.001).\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5332750/v1/4f5080f755b56d2c60c881cd.jpg"},{"id":79120427,"identity":"779776f1-cbf2-4432-8c74-57d0c99fc010","added_by":"auto","created_at":"2025-03-24 16:08:04","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2613464,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5332750/v1/2dd73a32-aa2f-4d28-94c4-7d91565c52ba.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integrating AI in Clinical Education: Evaluating General Practice Residents’ Proficiency in Distinguishing AI-Generated Hallucinations and Its Impacting Factors","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe application of artificial intelligence (AI) in medical education is becoming increasingly widespread, particularly in postgraduate training, showing significant potential for advancing educational practices\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e.In recent years, Generative Pre-trained Transformer (GPT) models, as an emerging technology, have been integrated into clinical teaching, with applications such as conducting clinical assessments, supporting self-directed learning for medical students, and creating virtual patient scenarios\u003csup\u003e\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eStudies indicated that GPT-assisted methods were comparable to traditional approaches in terms of difficulty and discriminative ability, while significantly reducing labor costs (by 57%)\u003csup\u003e3\u003c/sup\u003e.Additionally, large language models (LLMs) continue to evolve, demonstrating high accuracy (over 90%) in explaining medical problems and conducting clinical assessments\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. AI has now permeated various facets of medical education and clinical practice. However, despite its potential to enhance learning efficiency and reduce educational costs, the field of medical education deals with matters concerning patient safety and the cognitive framework of healthcare professionals, necessitating a more cautious and scientifically rigorous approach to its development.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAlthough Large Language Models (LLMs) is trained on high-quality datasets and improved accuracy through enhanced Retrieval-Augmented Generation (RAG), they may still generate errors in clinical content analysis\u003csup\u003e\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e, known as \"hallucinations\"\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, these hallucinations cannot be completely eliminated. In fact, the occurrence of hallucinations tends to be more common in larger LLMs.That appear logical but are often misleading\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Clinicians may misinterpret AI recommendations, particularly if the reasoning behind the AI\u0026rsquo;s decision-making is opaque or difficult to understand\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. In critical fields such as healthcare, AI hallucinations and biases can pose significant risks, so it's crucial for medical professionals to identify and address these issues. However, most researches focused solely on enhancing large language models and minimizing hallucinations, but little attention has been given to the ability of general practice residents to recognise inaccuracies during self-directed learning, as well as the factors that influence this capability.\u003c/p\u003e \u003cp\u003eAs AI becomes more prevalent in clinical settings,medical students will be confronted with clinical content and question-and-answer information spanning multiple disciplines, which they must interpret while guarding against AI biases\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. General practice residents play a crucial role in managing patient health and diagnosing diseases at an early stage. This requires not only a solid foundation of medical knowledge and skills but also a reliable cognitive framework and the ability to discern erroneous information\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. However, many doctors lack confidence in their ability to detect AI-generated errors\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Given that primary care physicians handle a wide range of clinical scenarios, understanding how they identify AI hallucinations is essential, yet research on this topic remains limited.\u003c/p\u003e \u003cp\u003eOur study evaluated general practice residents' ability to recognize and manage AI-generated hallucinations and investigated factors influencing this ability, such as AI usage frequency, clinical experience, and AI familiarity. The research seeks to explore key factors to improve residents' abilities to identify AI-generated inaccuracies during standardised general practice training, which also aims to provide valuable insights for the development of future educational frameworks that integrate AI with clinical training.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eA cross-sectional survey using ChatGPT (GPT-4) simulated responses for the 2024 Intermediate Attending Physician Examination in General Medicine\u003csup\u003e19\u003c/sup\u003e.\u0026nbsp;It consisted of four sections: three with 100 single-choice questions each and one with 18 clinical cases and 86 multiple-choice questions. Two senior general practitioners reviewed\u0026nbsp;and assessed\u0026nbsp;the consistency and validity of ChatGPT\u0026rsquo;s responses and explanations.\u0026nbsp;General practice resident doctors identified\u0026nbsp;correct and incorrect answers\u0026nbsp;and explanations.\u0026nbsp;Ethical approval has been obtained for this study, and all participants have provided informed consent\u0026nbsp;(\u003cstrong\u003eFigure1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDetermining characteristics of questions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMiller\u0026apos;s pyramid, proposed in 1989, outlined\u0026nbsp;four levels of medical education assessment: Knowledge (Knows),\u0026nbsp;Comprehension\u0026nbsp;(Knows How), Application (Shows How), and Evaluation (Does)\u003csup\u003e20\u003c/sup\u003e.\u0026nbsp;Key assessment strategies for evaluating learners\u0026apos; clinical reasoning abilities\u0026nbsp;were\u0026nbsp;outlined, spanning from knowledge acquisition to real-world application in the clinical setting\u003csup\u003e21\u003c/sup\u003e.At the same time\u0026nbsp;e-assessment methods could replace currently\u0026nbsp;conventional methods\u003csup\u003e22\u003c/sup\u003e.\u0026nbsp;This study categorized original questions according to Miller\u0026rsquo;s pyramid to assess specific medical competencies and analyze how different cognitive levels affect the accuracy of ChatGPT-generated responses and participants\u0026apos; ability to identify errors. Two senior general practitioners independently applied this taxonomy to ensure consistency, resolving discrepancies through discussion until a consensus was reached.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eP\u003c/strong\u003e\u003cstrong\u003erompt for ChatGPT\u0026apos;s answers and explanations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrompt engineering has highlighted the potential of large language models (LLMs) as effective tools in clinical medicine\u003csup\u003e23\u003c/sup\u003e. To optimize the value of the responses generated by these models, we applied\u0026nbsp;the ROT (Relevance, Objectives, Tasks) framework, tailored to the context and specific goals of the task\u003csup\u003e24\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRelevance:\u003c/strong\u003e Clearly define the domain of the query, such as \u0026quot;Assume the role of an experienced clinician.\u0026quot;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObjectives:\u003c/strong\u003e Articulate the aim of the query, for example, \u0026quot;This\u0026nbsp;is\u0026nbsp;a mid-level medical examination; respond to the questions provided with your best effort.\u0026quot;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTasks:\u003c/strong\u003e Specify the exact task to be performed, such as \u0026quot;Offer detailed explanations.\u0026quot;\u003c/p\u003e\n\u003cp\u003eThis framework\u0026nbsp;was\u0026nbsp;designed to improve the relevance and quality of responses generated by LLMs in clinical scenarios, enhancing their effectiveness as tools in medical education and practice.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eParticipant Settings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGeneral Practice Residents were involved\u0026nbsp;to assess the effectiveness of hallucination recognition. \u003cstrong\u003eInclusion criteria\u003c/strong\u003e: Residents\u0026nbsp;from Wuxi People\u0026apos;s Hospital in Jiangsu and\u0026nbsp;the\u0026nbsp;first\u0026nbsp;affiliated\u0026nbsp;hospital of Jiamusi University, the\u0026nbsp;second\u0026nbsp;affiliated\u0026nbsp;hospital of\u0026nbsp;Harbin\u0026nbsp;medical\u0026nbsp;university and the\u0026nbsp;second\u0026nbsp;affiliated\u0026nbsp;hospital of\u0026nbsp;Qiqihar\u0026nbsp;medical\u0026nbsp;university\u0026nbsp;in\u0026nbsp;Heilongjiang undergoing standardized general practice training, who voluntarily participate in the study.\u0026nbsp;\u003cstrong\u003eExclusion criteria\u003c/strong\u003e:\u0026nbsp;who did not\u0026nbsp;complete the training as scheduled,\u0026nbsp;those unwilling to participate or providing inaccurate data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDetection of Hallucinations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSignal Detection Theory (SDT)\u0026nbsp;was\u0026nbsp;employed to analyze the experimental data\u003csup\u003e25\u003c/sup\u003e. The errors and correct items\u0026nbsp;were\u0026nbsp;numbered and randomly assigned using the random function in Excel, with the subjects categorized as follows: \u0026quot;Basic Knowledge,\u0026quot; \u0026quot;Related Professional Knowledge,\u0026quot; and \u0026quot;Professional Knowledge\u0026quot; as subjects 1-3, and \u0026quot;Professional Practice\u0026quot; as subject 4. Participants completed\u0026nbsp;the questions using a standardized approach, and their responses\u0026nbsp;were\u0026nbsp;collected through an electronic survey. Each participant selected\u0026nbsp;one of three options (Correct, Incorrect, or Uncertain) to assess the accuracy of each response and explanation.\u0026nbsp;\u003cstrong\u003eSensitivity (d\u0026rsquo;):\u003c/strong\u003e Measures the participant\u0026apos;s ability to correctly identify true positives. It\u0026nbsp;was\u0026nbsp;calculated as Z(Hit Rate) - Z(False Alarm Rate).\u0026nbsp;\u003cstrong\u003eResponse Bias (\u0026beta;):\u003c/strong\u003e Evaluates the tendency of participants to favor one response option over others.\u0026nbsp;Categorized into two groups depending on whether the \u0026beta; value is equal or greater than 1, or less than 1.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll statistical analyses were performed using IBM SPSS Statistics 26.0. For normally distributed data, use t test for two groups and one-way ANOVA for multiple comparisons, with LSD or Dunnett\u0026apos;s test for pairwise comparisons. For categorical data, use chi-squared tests and the Bonferroni correction to adjust for multiple comparisons among group rates. Binary logistic regression analysis was performed to explore factors influencing participants\u0026apos; response bias. All statistical tests were two-sided, with a p-value below 0.05 considered statistically significant.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eCharacteristic among General Practice Resident\u003c/strong\u003e\u003cstrong\u003es\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAfter an initial screening of the 142 collected questionnaires, 5 with all correct answers were excluded. The average age of the participants was 25.93. The participants consisted of 46.72% males, 81.75% undergraduates, and 45.26% from Jiangsu. The group comprised 41 participants from the class of 2021, 43 from 2022, and 53 from 2023. Regarding the Occupational Medical Examination,\u0026nbsp;13 individuals\u0026nbsp;were not passed, and 73 successfully passed. As for familiarity and usage of AI, 52.55% were unfamiliar with AI, 35.04% had never applied AI, and\u0026nbsp;only\u0026nbsp;8.03% used AI frequently in academic research. The average satisfaction score for AI usage was 7.0 on a 0-10 scales.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChatGPT\u0026apos;s H\u003c/strong\u003e\u003cstrong\u003eallucinations\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;Of Different Categorization of Medical questions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn our study, hallucinations were discovered by two senior general practitioners via verifying each question\u0026apos;s answers and explanations. Out of the 386 questions answered by ChatGPT, 312 answers were correct, while 74 answers were incorrect. Out of the 312 correct answers, the explanations for 299 were accurate, therefore, there were ultimately 87 incorrect explanations (\u003cstrong\u003eFigure1\u003c/strong\u003e). ChatGPT\u0026apos;s overall accuracy was 80.8% before human correction, dropping slightly to 80.05%. For knowledge-level questions, accuracy fell slightly from 90.29% to 89.32%, and for comprehension and application, it dropped from 86.61% to 86.61% and from 83.21% to 77.10%, respectively. Evaluation-level questions saw a sharp decline from 43.48% to 30.43% after correction (\u003cstrong\u003eTable1\u003c/strong\u003e). The majority of the incorrect explanations are comprised of factual inaccuracies and meaningless responses. Hallucinations are especially evident on the cognitive layers of application and evaluation, which were key elements in cultivation of clinical reasoning for general practitioners.\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" style=\"width: 723px;\" width=\"723\" height=\"814\"\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRecognition of AI-Generated Hallucinations by General Practice Resident\u003c/strong\u003e\u003cstrong\u003es\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur study found the median hit rate, false alarm rate, correct rejection rate, and miss rate for\u0026nbsp;general\u0026nbsp;practice\u0026nbsp;resident\u0026nbsp;doctor in identifying generative hallucinations were 86% (15%), 74% (21%), 26% (21%), and 14% (21%), respectively. This indicates a high false alarm rate and a low correct rejection rate. Overall accuracy was only 55%, with sensitivity at 0.39 \u0026plusmn; 0.33. Additionally, 87% of participants took a more permissive approach in evaluating hallucinations.\u003c/p\u003e\n\u003cp\u003eIn addition to the overall recognition results, a stratified analysis by subject showed a median hit rate of 98% for Subject 1 and 87% for Subject 4. The interquartile range (IQR) , measured between the 25th and 75th percentiles, spanned from 73% to 91% for Subjects 1-3 and from 81% to 95% for Subject 4. False alarm rates had medians of 89% for Subjects 1-3 and 34% for Subject 4, with IQRs of 54% to 77% and 73% to 92%, respectively. For correct rejection rates, medians were 34% for Subjects 1-3 and 14% for Subject 4, with IQRs of 23% to 46% and 8% to 27%. The median miss rate was 17% for Subjects 1-3 and 11% for Subject 4, with IQRs of 9% to 28% and 33% to 80%. Uncertainty rates had medians of 57% for Subject 1 and 53% for Subject 4, with narrow IQRs of 0% to 9% and 0% to 8%, respectively.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eComparison of accuracy, sensitivity, and response propensity revealed that Subject 4 had lower accuracy than the overall average (\u003cem\u003eP\u003c/em\u003e \u0026lt;\u0026nbsp;0.05) and was significantly less accurate than Subjects 1-3 (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001). As cognitive levels advance, accuracy progressively decreased\u0026nbsp;and this trend\u0026nbsp;was statistically significant\u0026nbsp;(\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001).\u0026nbsp;Subject 4 also had notably lower sensitivity compared to Subjects 1-3 (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05). As cognitive stratification progresses, sensitivity increases progressively, with all differences being statistically significant except between the knowledge and comprehensive groups, and the application and evaluation groups\u0026nbsp;(\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001).\u0026nbsp;In Subject 4, there are more people choosing neutral and\u0026nbsp;conservative\u0026nbsp;standards compared to the overall group and Subjects 1-3(all\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001). Furthermore, more people choose strict standards for evaluation\u0026nbsp;cognitive layer\u0026nbsp;compared to the overall\u0026nbsp;and\u0026nbsp;the application level\u0026nbsp;groups\u0026nbsp;(all\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001)\u0026nbsp;(\u003cstrong\u003eFigure2A-D\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFactors impact respond bias of identification of hallucinations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study investigated factors influencing general practice resident doctors\u0026apos; choice of liberal, neutral, or conservative criteria for assessing GPT-generated responses. Among the variables analyzed, only AI usage frequency showed significant differences\u0026nbsp;(\u003cem\u003eP\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e=\u0026nbsp;0.02). Factors such as age, gender, degree, institution, grade, AI familiarity, use cases, and user satisfaction had no significant impact. Participants with stricter criteria tended to use AI more frequently\u0026nbsp;(\u003cstrong\u003eTable\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e2\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2 Factors impact Response Bias of identification of hallucinations among General Practice Residents\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003eLiberal\u0026nbsp;(n=119)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003eNeutral+\u003c/p\u003e\n \u003cp\u003eConservative\u003c/p\u003e\n \u003cp\u003e(n=18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003eValue (t\u003cem\u003e\u0026nbsp;or X\u003csup\u003e2\u0026nbsp;\u003c/sup\u003e)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e25.87\u003cu\u003e+\u003c/u\u003e 2.10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e26.33\u003cu\u003e+\u003c/u\u003e 2.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e-0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.39\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eGender (male%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e54(45%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e10(55.60%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e0.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.42\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eDegree(undergraduate%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e96(80.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e16(88.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e0.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.40\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eInstitution(Jiangsu%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e54(45.38)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e8(44.44)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.94\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eGrade (2021%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e37(31.09)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e4(22.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e1.18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.55\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eOccupational Physician Examination (unpass%) \u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e52(43.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e12(66.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e3.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.07\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eKnowledge of AI\u003c/p\u003e\n \u003cp\u003e(unknown%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e63(52.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e9(50.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.81\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eApplication of AI Utilized (no%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e38(31.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e11(61.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e5.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.02*\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eAI use cases\u003c/p\u003e\n \u003cp\u003e(Academic%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e10(8.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e1(5.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e6.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.15\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eUser satisfaction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e7.09\u0026plusmn;1.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e6.13\u0026plusmn;1.12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e2.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.05\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 31.4947%;\"\u003e\n \u003cp\u003eResponse time\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.1495%;\"\u003e\n \u003cp\u003e72.82\u0026plusmn;6.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.0391%;\"\u003e\n \u003cp\u003e67.61\u0026plusmn;15.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9715%;\"\u003e\n \u003cp\u003e1.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.3452%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.17\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eBinary regression analysis identified response time, AI familiarity, and AI usage frequency as key factors affecting the choice of a neutral and stricter criteria. Specifically, shorter response times, less familiarity with AI, and frequently AI usage were linked to stricter criteria (\u003cem\u003eP\u003c/em\u003e = 0.02; \u003cem\u003eP\u003c/em\u003e = 0.04; \u003cem\u003eP\u003c/em\u003e = 0.01, respectively)(\u003cstrong\u003eTable\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e2\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eand Table\u003c/strong\u003e\u003cstrong\u003e3\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3 Binary logistics regression Analysis of Factors impact neutral and Conservative Response Bias\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eFactor\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003eB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003eExp(B)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e95%CI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eResponse Time\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e-0.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.86-0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.02*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eKnowledge of AI\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eKnown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e-1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.03-0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.04*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eunknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1[Reference]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eApplication of AI\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eUtilized\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e2.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e10.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e1.57-62.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.01*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eunused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1[Reference]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.78-1.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.73\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003efemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.54-5.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e\u0026nbsp; male\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1[Reference]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eInstitution\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e\u0026nbsp; Heilongjiang\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.34-6.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e\u0026nbsp; Wuxi/Jiangsu\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1[Reference]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eDegree\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e\u0026nbsp; graduate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.25-8.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e\u0026nbsp; undergraduate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eGrade\u0026nbsp;(2021)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.16-9.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e2022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e-0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e0.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.10-6.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.79\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003e2021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1[Reference]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003eOccupational Physician Examination\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003epass\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e1.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e3.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e0.58-18.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e0.18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 32.0988%;\"\u003e\n \u003cp\u003enot passed\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.5203%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.9894%;\"\u003e\n \u003cp\u003e1[Reference]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.8131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5785%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Discussion","content":"\u003cp\u003eTo evaluate the application value of ChatGPT in various clinical scenarios in medical education, we utilized Miller's Pyramid to categorize these questions, assessing different levels of medical competency. The findings indicated that GPT is relatively accurate in handling basic clinical knowledge and problems, even after reviewing by experienced general practitioners. However, its performance deteriorated significantly (44.2%-57.0%) when dealing with more complex clinical cases or scenarios. The results showed that ChatGPT demonstrated a generally high accuracy rate, but its accuracy is lower in the domain of professional practice, indicating that it was more suitable for judgment and guidance on basic knowledge and simpler clinical problems.\u003c/p\u003e \u003cp\u003eOur research further examined the ability of general practice residents to recognize AI-generated hallucinations. The results revealed a concerning trend: residents exhibit significant difficulty in identifying AI-generated hallucinations, with an overall accuracy rate of only 55.0%\u0026plusmn;4.3%. They showed relatively high sensitivity in recognizing basic medical knowledge, understanding, and application cognitive levels but tend to adopt more lenient standard as response time increases. There were no notable differences in outcomes among general practice residents across different training years and regions. All had limited ability to detect AI-generated hallucinations, especially in complex clinical cases. Key risk factors included frequent AI use and limited understanding of how it works, which led to more stringent evaluation standards. This underscored the importance of understanding AI\u0026rsquo;s fundamentals and limitations, reducing over-reliance on it, and enhancing clinical reasoning and critical thinking\u003csup\u003e\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. These were crucial steps for improving residents' ability to identify hallucinations and essential for integrating ChatGPT effectively into medical education.\u003c/p\u003e \u003cp\u003eResidents undergoing standardized general practice training encounter AI-generated clinical data and support information throughout chronic disease management\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e, necessitating the integration, understanding, and coordination of patient information from all healthcare professionals, environments, and timelines\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. As advanced AI technologies and applications emerged, it was crucial to critically evaluate their impacts, especially potential pitfalls. Despite consciously applying relatively strict criteria in clinical cases, the accuracy of identifying AI-generated errors significantly declined, nearly resembling random guessing. Participants who perceived themselves as knowledgeable about AI tended to adopt lenient standards, while those who frequently used AI preferred stricter criteria, likely due to their prior experiences. Participants commonly applied AI in personal and research contexts, where AI provided relatively reliable feedback. Errors in these domains incurred minimal time and financial costs. However, when AI was applied in clinical settings, it activated participants' sense of medical responsibility and their subconscious awareness of potential impacts on patients\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e.Our study also found that as the complexity of cognitive tasks increased, sensitivity in recognizing AI hallucinations tended to rise.\u003c/p\u003e \u003cp\u003eOur findings are consistent with earlier studies on AI in healthcare and clinical decision support systems, while also bringing to light new challenges. AI has shown substantial benefits in enhancing diagnostic precision and offering treatment recommendations grounded in evidence\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. For single-text discharge summaries, AI could function without hallucinations\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. In scenarios involving creative tasks, hallucinations were more likely to occur\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e.Our findings built on previous research, revealing that AI systems encounter significant challenges when processing complex, multidimensional clinical data. These hallucinations were not only common in large-scale medical AI models but also proved especially difficult for residents undergoing standardized general practice training to identify\u003csup\u003e\u003cspan additionalcitationids=\"CR38 CR39\" citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e, which increased the risk of misdiagnosis and clinical errors\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Primary care physicians tended to accept plausible yet incorrect answers without much skepticism and seldom questioned AI outputs. Although AI has demonstrated diagnostic accuracy on par with, or even surpassing, that of experienced doctors\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e, it's crucial to remain mindful of cognitive limitations, AI-driven biases, and the risk of illusory understanding\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. Some researchers have proposed that focusing on decision-making grounded in relative uncertainty, rather than simply aiming to minimize uncertainty, offers a more insightful approach to applying AI in medicine\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. This suggested that when using AI, we should embrace and manage uncertainty instead of placing undue trust in AI's seemingly conclusive outputs. Excessive dependence on AI risked undermining clinical intuition, potentially relegating clinicians to simply following algorithm-driven recommendations without exercising critical judgment\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. Failure to recognize hallucinations may prolong decision times and lead to erroneous analyses\u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e,\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. As highlighted by the WHO guidelines on the ethics and governance of AI in healthcare, it was crucial to ensure data transparency, interpretability, and accountability while fostering a sense of responsibility among clinicians\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. This is also a key concern that educators in medical education need to pay close attention to.\u003c/p\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eStrengths\u003c/h2\u003e \u003cp\u003eOur study had several strengths. The study population consisted of residents undergoing standardized general practice training, a critical stage in medical education. Moreover, our research not only evaluated AI performance but also examined resident doctors' ability to identify AI hallucinations, offering a comprehensive view of AI's practical application in clinical settings. They often applied lenient recognition standards with low sensitivity to AI responses, highlighting the need for specialized training to enhance critical thinking and evaluation skills. Additionally, general practice residents were more susceptible to hallucinations compared to experienced physicians, but they might also benefit more from AI support\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e.The low recognition accuracy calls for stricter AI usage guidelines and clearer uncertainty markers in AI systems to help distinguish valid recommendations from potential errors.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eHowever, there were also several limitations to our study. Firstly, the reliance on a specific AI system (GPT) may limit the generalizability of the findings to other AI models. Secondly, the assessment, conducted through closed-ended questions, ranging from basic knowledge to complex clinical scenarios, may not fully reflect real clinical situations.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eGeneral practice residents have limited ability to recognize AI-generated hallucinations, particularly in complex clinical scenarios. As the frequency and unfamiliarity of AI use increases, residents tended to apply stringent assessment standards, reflecting an over reliance on AI.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eEthical approval has been obtained for this study by Research Ethics Committee of Wuxi People\u0026rsquo;s Hospital, and all participants have provided informed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe are very grateful to all the residents who participated in this research.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eYunyun Zhang and Zhiyong Zhang had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.\u003c/p\u003e\n\u003cp\u003eConceive and design: Jiacheng Zhou, Xiaochuan Cui, Yunyun Zhang and Zhiyong Zhang.\u003c/p\u003e\n\u003cp\u003eAcquisition, analysis, or interpretation of data: All authors.\u003c/p\u003e\n\u003cp\u003eStatistical analysis: Jintao Zhang and Yunyun Zhang.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDrafting of the manuscript: Jiacheng Zhou and Yunyun Zhang.\u003c/p\u003e\n\u003cp\u003eObtained funding: Yunyun Zhang and Jiacheng Zhou.\u003c/p\u003e\n\u003cp\u003eSupervision: Yunyun Zhang and Zhiyong Zhang\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e This study was funded by the \u0026quot;Top Talent Support Program for Young and Middle-aged people of Wuxi Health Committee(BJRC-8 and HB2023015, Dr Zhang); Taihu Light Basic Research Project, Wuxi Municipal Science and Technology Bureau(K20221023, Dr Zhang);Emerging Discipline Leader Program (2024-YZ-HBDTR-ZYY-2024, Dr Zhang); Wuxi Association for Science and Technology (KX-24-C154; Jiacheng Zhou)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Sharing Statement:\u0026nbsp;\u003c/strong\u003eThe data supporting the findings of this study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors provide their consent for the publication of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest Disclosures\u003c/strong\u003e: The authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMekki YM, Zughaier SM. Teaching artificial intelligence in medicine. Nat Rev Bioeng. 2024;2(6):450\u0026ndash;1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s44222-024-00195-0\u003c/span\u003e\u003cspan address=\"10.1038/s44222-024-00195-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan M, Cerri GG, Moraes FY. ChatGPT and medicine: how AI language models are shaping the future and health related careers. Nat Biotechnol. 2023;41(11):1657\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41587-023-02011-3\u003c/span\u003e\u003cspan address=\"10.1038/s41587-023-02011-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLam G, Shammoon Y, Coulson A, et al. Utility of large language models for creating clinical assessment items. Med Teach Published online August. 2024;26:1\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/0142159X.2024.2382860\u003c/span\u003e\u003cspan address=\"10.1080/0142159X.2024.2382860\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShang L, Li R, Xue M, Guo Q, Hou Y. Evaluating the application of ChatGPT in China\u0026rsquo;s residency training education: An exploratory study. Med Teach Published online July. 2024;12:1\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/0142159X.2024.2377808\u003c/span\u003e\u003cspan address=\"10.1080/0142159X.2024.2377808\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCook DA. Creating virtual patients using large language models: scalable, global, and low cost. Med Teach Published online July 11, 2024:1\u0026ndash;3. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/0142159X.2024.2376879\u003c/span\u003e\u003cspan address=\"10.1080/0142159X.2024.2376879\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChaurasia AK, Greatbatch CJ, Hewitt AW. Diagnostic Accuracy of Artificial Intelligence in Glaucoma Screening and Clinical Practice. J Glaucoma. 2022;31(5):285. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/IJG.0000000000002015\u003c/span\u003e\u003cspan address=\"10.1097/IJG.0000000000002015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChavez H, Chavez-Arias B, Contreras-Rosas S, Alvarez-Rodr\u0026iacute;guez JM, Raymundo C. Artificial neural network model to predict student performance using nonpersonal information. Front Educ. 2023;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/feduc.2023.1106679\u003c/span\u003e\u003cspan address=\"10.3389/feduc.2023.1106679\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLambert SI, Madi M, Sopka S, et al. An integrative review on the acceptance of artificial intelligence among healthcare professionals in hospitals. NPJ Digit Med. 2023;6(1):111. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41746-023-00852-5\u003c/span\u003e\u003cspan address=\"10.1038/s41746-023-00852-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKresevic S, Giuffr\u0026egrave; M, Ajcevic M, Accardo A, Croc\u0026egrave; LS, Shung DL. Optimization of hepatological clinical guidelines interpretation by large language models: a retrieval augmented generation-based framework. NPJ Digit Med. 2024;7(1):102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41746-024-01091-y\u003c/span\u003e\u003cspan address=\"10.1038/s41746-024-01091-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGe J, Sun S, Owens J, et al. Development of a liver disease-specific large language model chat interface using retrieval-augmented generation. Hepatol Published online March. 2024;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/HEP.0000000000000834\u003c/span\u003e\u003cspan address=\"10.1097/HEP.0000000000000834\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiuffr\u0026egrave; M, Kresevic S, Pugliese N, You K, Shung DL. Optimizing large language models in digestive disease: strategies and challenges to improve clinical outcomes. Liver Int. 2024;44(9):2114\u0026ndash;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/liv.15974\u003c/span\u003e\u003cspan address=\"10.1111/liv.15974\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerrmann-Werner A, Festl-Wietek T, Holderried F, et al. Assessing ChatGPT\u0026rsquo;s Mastery of Bloom\u0026rsquo;s Taxonomy Using Psychosomatic Medicine Exam Questions: Mixed-Methods Study. J Med Internet Res. 2024;26:e52113. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2196/52113\u003c/span\u003e\u003cspan address=\"10.2196/52113\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTran CG, Chang J, Sherman SK, De Andrade JP. Performance of ChatGPT on American Board of Surgery In-Training Examination Preparation Questions. J Surg Res. 2024;299:329\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jss.2024.04.060\u003c/span\u003e\u003cspan address=\"10.1016/j.jss.2024.04.060\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJussupow E, Spohrer K, Heinzl A, Gawlitza J. Augmenting Medical Diagnosis Decisions? An Investigation into Physicians\u0026rsquo; Decision-Making Process with Artificial Intelligence. Inform Syst Res. 2021;32(3):713\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1287/isre.2020.0980\u003c/span\u003e\u003cspan address=\"10.1287/isre.2020.0980\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen D, Parsa R, Hope A, et al. Physician and Artificial Intelligence Chatbot Responses to Cancer Questions From Social Media. JAMA oncology. Published online May. 2024;16:e240836. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamaoncol.2024.0836\u003c/span\u003e\u003cspan address=\"10.1001/jamaoncol.2024.0836\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoodman RS, Patrinely JR, Stone CA Jr, et al. Accuracy and Reliability of Chatbot Responses to Physician Questions. JAMA Netw Open. 2023;6(10):e2336483. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2023.36483\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2023.36483\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBuck C, Doctor E, Hennrich J, J\u0026ouml;hnk J, Eymann T. General Practitioners\u0026rsquo; Attitudes Toward Artificial Intelligence\u0026ndash;Enabled Systems: Interview Study. J Med Internet Res. 2022;24(1):e28916. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2196/28916\u003c/span\u003e\u003cspan address=\"10.2196/28916\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHantel A, Walsh TP, Marron JM, et al. Perspectives of Oncologists on the Ethical Implications of Using Artificial Intelligence for Cancer Care. JAMA Netw Open. 2024;7(3):e244077. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2024.4077\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2024.4077\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDu QF, Wang JJ. 2024 General Medicine Practice Mock Exam. People\u0026rsquo;s Medical Publishing House; 2023: 3\u0026ndash;39. ISBN: 9787117355421.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTen Cate O, Carraccio C, Damodaran A, et al. Entrustment Decision Making: Extending Miller\u0026rsquo;s Pyramid. Acad Med. 2021;96(2):199\u0026ndash;204. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/ACM.0000000000003800\u003c/span\u003e\u003cspan address=\"10.1097/ACM.0000000000003800\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThampy H, Willert E, Ramani S. Assessing Clinical Reasoning: Targeting the Higher Levels of the Pyramid. J Gen Intern Med. 2019;34(8):1631\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11606-019-04953-4\u003c/span\u003e\u003cspan address=\"10.1007/s11606-019-04953-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHasani H, Khoshnoodifar M, Khavandegar A, et al. Comparison of electronic versus conventional assessment methods in ophthalmology residents; a learner assessment scholarship study. BMC Med Educ. 2021;21(1):342. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12909-021-02759-9\u003c/span\u003e\u003cspan address=\"10.1186/s12909-021-02759-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMesk\u0026oacute; B. Prompt Engineering as an Important Emerging Skill for Medical Professionals: Tutorial. J Med Internet Res. 2023;25:e50638. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2196/50638\u003c/span\u003e\u003cspan address=\"10.2196/50638\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang L, Chen X, Deng X, et al. Prompt engineering in consistency and reliability with the evidence-based guideline for LLMs. NPJ Digit Med. 2024;7(1):41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41746-024-01029-4\u003c/span\u003e\u003cspan address=\"10.1038/s41746-024-01029-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAujla H. d[Formula: see text]: Sensitivity at the optimal criterion location. Behav Res Methods. 2023;55(5):2532\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3758/s13428-022-01913-5\u003c/span\u003e\u003cspan address=\"10.3758/s13428-022-01913-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLarson BZ, Moser C, Caza A, Muehlfeld K, Colombo LA. Critical Thinking in the Age of Generative AI. AMLE. 2024;23(3):373\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5465/amle.2024.0338\u003c/span\u003e\u003cspan address=\"10.5465/amle.2024.0338\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoulin TC. Learning with AI Language Models: Guidelines for the Development and Scoring of Medical Questions for Higher Education. J Med Syst. 2024;48(1):45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10916-024-02069-9\u003c/span\u003e\u003cspan address=\"10.1007/s10916-024-02069-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSuriano R, Plebe A, Acciai A, Fabio RA. Student interaction with ChatGPT can promote complex critical thinking skills. Learn Instruction. 2025;95:102011. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.learninstruc.2024.102011\u003c/span\u003e\u003cspan address=\"10.1016/j.learninstruc.2024.102011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang S, Shi Y, Sui M, et al. Telephone follow-up based on artificial intelligence technology among hypertension patients: Reliability study. J Clin Hypertens (Greenwich). 2024;26(6):656\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/jch.14823\u003c/span\u003e\u003cspan address=\"10.1111/jch.14823\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Guan Z, Wang J, et al. Integrated image-based deep learning and language models for primary diabetes care. Nat Med Published online July. 2024;19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41591-024-03139-8\u003c/span\u003e\u003cspan address=\"10.1038/s41591-024-03139-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEverson J, Hendrix N, Phillips RL, Adler-Milstein J, Bazemore A, Patel V. Primary Care Physicians\u0026rsquo; Satisfaction With Interoperable Health Information Technology. JAMA Netw Open. 2024;7(3):e243793. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2024.3793\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2024.3793\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimmenroth-Nayda A, Marx G, Lorkowski T, Himmel W. Working as simulated patient has effects on real patient life - Preliminary insights from a qualitative study. GMS J Med Educ. 2016;33(3):Doc42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3205/zma001041\u003c/span\u003e\u003cspan address=\"10.3205/zma001041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBuch VH, Ahmed I, Maruthappu M. Artificial intelligence in medicine: current trends and future possibilities. Br J Gen Pract. 2018;68(668):143\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3399/bjgp18X695213\u003c/span\u003e\u003cspan address=\"10.3399/bjgp18X695213\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Zhou L, Zhan Y, et al. How does the artificial intelligence-based image-assisted technique help physicians in diagnosis of pulmonary adenocarcinoma? A randomized controlled experiment of multicenter physicians in China. J Am Med Inf Assoc. 2022;29(12):2041\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/jamia/ocac179\u003c/span\u003e\u003cspan address=\"10.1093/jamia/ocac179\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTung JYM, Gill SR, Sng GGR, et al. Comparison of the Quality of Discharge Letters Written by Large Language Models and Junior Clinicians: Single-Blinded Study. J Med Internet Res. 2024;26:e57721. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2196/57721\u003c/span\u003e\u003cspan address=\"10.2196/57721\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZaretsky J, Kim JM, Baskharoun S, et al. Generative Artificial Intelligence to Transform Inpatient Discharge Summaries to Patient-Friendly Language and Format. JAMA Netw open. 2024;7(3):e240357. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2024.0357\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2024.0357\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMirkin S, Albensi BC. Should artificial intelligence be used in conjunction with Neuroimaging in the diagnosis of Alzheimer\u0026rsquo;s disease? Front Aging Neurosci. 2023;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fnagi.2023.1094233\u003c/span\u003e\u003cspan address=\"10.3389/fnagi.2023.1094233\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchubert MC, Lasotta M, Sahm F, Wick W, Venkataramani V. Evaluating the Multimodal Capabilities of Generative AI in Complex Clinical Diagnostics. Published online November 2, 2023:2023.11.01.23297938. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/2023.11.01.23297938\u003c/span\u003e\u003cspan address=\"10.1101/2023.11.01.23297938\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrown C, Nazeer R, Gibbs A, et al. Breaking Bias: The Role of Artificial Intelligence in Improving Clinical Decision-Making. Cureus. 2023;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7759/cureus.36415\u003c/span\u003e\u003cspan address=\"10.7759/cureus.36415\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRueckel J, Huemmer C, Fieselmann A, et al. Pneumothorax detection in chest radiographs: optimizing artificial intelligence system for accuracy and confounding bias reduction using in-image annotations in algorithm training. Eur Radiol. 2021;31(10):7888\u0026ndash;900. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00330-021-07833-w\u003c/span\u003e\u003cspan address=\"10.1007/s00330-021-07833-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLevinson AV, Goyal A, Man RHC, et al. Using Audio Data to Facilitate Depression Risk Assessment in Primary Health Care. Published online Oct. 2023;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.48550/arXiv.2310.10928\u003c/span\u003e\u003cspan address=\"10.48550/arXiv.2310.10928\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRauschecker AM, Rudie JD, Xie L, et al. Artificial Intelligence System Approaching Neuroradiologist-level Differential Diagnosis Accuracy at Brain MRI. Radiology. 2020;295(3):626\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1148/radiol.2020190283\u003c/span\u003e\u003cspan address=\"10.1148/radiol.2020190283\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShen J, Zhang CJP, Jiang B, et al. Artificial Intelligence Versus Clinicians in Disease Diagnosis: Systematic Review. JMIR Med Inf. 2019;7(3):e10010. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2196/10010\u003c/span\u003e\u003cspan address=\"10.2196/10010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMesseri L, Crockett MJ. Artificial intelligence and illusions of understanding in scientific research. Nature. 2024;627(8002):49\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-024-07146-0\u003c/span\u003e\u003cspan address=\"10.1038/s41586-024-07146-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhy scientists trust AI too much. - and what to do about it. Nature. 2024;627(8003):243. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/d41586-024-00639-y\u003c/span\u003e\u003cspan address=\"10.1038/d41586-024-00639-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarish V, Morgado F, Stern AD, Das S. Artificial Intelligence and Clinical Decision Making: The New Nature of Medical Uncertainty. Acad Med. 2021;96(1):31\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/ACM.0000000000003707\u003c/span\u003e\u003cspan address=\"10.1097/ACM.0000000000003707\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu T, Duan Y. Beware the self-fulfilling prophecy: enhancing clinical decision-making with AI. Crit Care. 2024;28(1):276. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13054-024-05062-3\u003c/span\u003e\u003cspan address=\"10.1186/s13054-024-05062-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTriberti S, Durosini I, Pravettoni G. A Third Wheel Effect in Health Decision Making Involving Artificial Entities: A Psychological Perspective. Front Public Health. 2020;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fpubh.2020.00117\u003c/span\u003e\u003cspan address=\"10.3389/fpubh.2020.00117\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJacobs M, Pradier MF, McCoy TH, Perlis RH, Doshi-Velez F, Gajos KZ. How machine-learning recommendations influence clinician treatment selections: the example of antidepressant selection. Transl Psychiatry. 2021;11(1):1\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41398-021-01224-x\u003c/span\u003e\u003cspan address=\"10.1038/s41398-021-01224-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEthics and governance of artificial intelligence for health. Accessed September 14. 2024. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/publications/i/item/9789240029200\u003c/span\u003e\u003cspan address=\"https://www.who.int/publications/i/item/9789240029200\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang W, Gao G (Gordon), Agarwal R, editors. Friend or Foe? Teaming Between Artificial Intelligence and Workers with Variation in Experience. Management Science. 2024;70(9):5753\u0026ndash;5775. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1287/mnsc.2021.00588\u003c/span\u003e\u003cspan address=\"10.1287/mnsc.2021.00588\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-education","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"meed","sideBox":"Learn more about [BMC Medical Education](http://bmcmededuc.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/meed/default.aspx","title":"BMC Medical Education","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"AI-Generated Hallucinations, General Practice Residents, Standardized General Practice Training, Response Bias","lastPublishedDoi":"10.21203/rs.3.rs-5332750/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5332750/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eOBJECTIVE\u003c/b\u003e\u003c/p\u003e \u003cp\u003eTo evaluate the ability of general practice residents to detect AI-generated hallucinations and assess the influencing factors.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMETHODS\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThis multi-center study involved 142 general practice residents, all of whom were undergoing standardized general practice training and volunteered to participate. The study evaluated AI\u0026rsquo;s accuracy and consistency, along with the residents\u0026rsquo; response time, accuracy, sensitivity(d\u0026rsquo;), and standard tendencies (β). Binary regression analysis was used to explore factors affecting the residents' ability to identify AI-generated errors.\u003c/p\u003e\u003cp\u003e\u003cb\u003eRESULTS\u003c/b\u003e\u003c/p\u003e \u003cp\u003e137 participants ultimately included had an mean (SD) age 25.93\u0026thinsp;\u0026plusmn;\u0026thinsp;2.10, with 46.72% male, 81.75% undergraduates, and 45.26% from Jiangsu. Regarding AI, 52.55% were unfamiliar with it, 35.04% had never used it. ChatGPT demonstrated 80.8% overall accuracy, including 57% in professional practice. 87 AI-generated hallucinations were identified, primarily in the level of application and evaluation. The mean (SD) accuracy was 55% \u0026plusmn;4.3%, and the mean (SD) sensitivity (d') was 0.39\u0026thinsp;\u0026plusmn;\u0026thinsp;0.33. The median response bias (β) was 0.74 (0.31). Regression analysis revealed that shorter response times (OR\u0026thinsp;=\u0026thinsp;0.92, P\u0026thinsp;=\u0026thinsp;0.02), higher self-assessed AI understanding (OR\u0026thinsp;=\u0026thinsp;0.16, P\u0026thinsp;=\u0026thinsp;0.04), and frequent AI use (OR\u0026thinsp;=\u0026thinsp;10.43, P\u0026thinsp;=\u0026thinsp;0.01) were associated with stricter error detection criteria.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCONCLUSIONS\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThe study concluded that residents struggled to identify AI errors, particularly in clinical cases, emphasizing the importance of improving AI literacy and critical thinking for effective integration into medical education.\u003c/p\u003e","manuscriptTitle":"Integrating AI in Clinical Education: Evaluating General Practice Residents’ Proficiency in Distinguishing AI-Generated Hallucinations and Its Impacting Factors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-20 07:44:43","doi":"10.21203/rs.3.rs-5332750/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-29T03:54:10+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-10-29T02:36:30+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-10-29T01:56:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Education","date":"2024-10-25T13:27:15+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-education","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"meed","sideBox":"Learn more about [BMC Medical Education](http://bmcmededuc.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/meed/default.aspx","title":"BMC Medical Education","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"00e1282d-8cd7-4793-a839-54e9a14160b1","owner":[],"postedDate":"November 20th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-03-24T16:00:42+00:00","versionOfRecord":{"articleIdentity":"rs-5332750","link":"https://doi.org/10.1186/s12909-025-06916-2","journal":{"identity":"bmc-medical-education","isVorOnly":false,"title":"BMC Medical Education"},"publishedOn":"2025-03-19 15:57:19","publishedOnDateReadable":"March 19th, 2025"},"versionCreatedAt":"2024-11-20 07:44:43","video":"","vorDoi":"10.1186/s12909-025-06916-2","vorDoiUrl":"https://doi.org/10.1186/s12909-025-06916-2","workflowStages":[]},"version":"v1","identity":"rs-5332750","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5332750","identity":"rs-5332750","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}