Harnessing ChatGPT for SBA Question Writing: A Novel Approach to Medical Student Learning

Harnessing ChatGPT for SBA Question Writing: A Novel Approach to Medical Student Learning | 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 Harnessing ChatGPT for SBA Question Writing: A Novel Approach to Medical Student Learning Daniel McGrane, Alice Gore, Isaac Barnes This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9361324/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract The introduction of the Medical Licensing Assessment (MLA) requires UK medical graduates to pass two 100-item single best answer (SBA) question papers. Access to high-quality question banks is often restricted by expensive paywalls, evolving guidelines, and institutional safeguarding of questions for future assessments. This limits students' ability to practice exam-style questions and develop their clinical reasoning. During this study, two medical students used ChatGPT to produce a bank of 100 SBA questions which were representative of those used in the MLA. Questions were analysed and revised by a student and a member of a university assessment team. A critical review of ChatGPT’s benefits was conducted to determine the scope for application in both self-directed learning and formal medical school assessment. ChatGPT successfully formulates clinical vignettes and accompanying SBA questions in less than 50% of the time taken to generate a comparable question, when written by a clinical educator. It is, however, not yet reliable enough to be used without expert supervision or direct correlation with current medical guidelines. Both medical students and clinical educators will benefit from ChatGPT’s ability to rapidly compose SBA questions that emulate those used in medical school exams, provided they remain conscious of the risk of false clinical recommendations. The acquisition and implementation of our question bank by a leading commercial medical education provider confirms the value of ChatGPT generated SBA questions as a resource for medical students preparing for the UKMLA. Assessment Artificial Intelligence ChatGPT Medical Education Figures Figure 1 Figure 2 Introduction With the introduction of the Medical Licensing Assessment (MLA), all medical students graduating from UK medical schools from 2024 onwards will be required to pass the Applied Knowledge Test (AKT). This test comprises two 100-item papers featuring Single Best Answer (SBA) questions and constitutes 50% of the MLA. To prepare effectively for the AKT, medical students must develop a strong understanding of SBA question formats, gain extensive practice in answering them, and acquire comprehensive knowledge of the curriculum content. SBA questions differ from traditional multiple-choice questions (MCQs), which typically test recall by presenting one correct answer among five options. In contrast, SBAs include a clinical vignette, a lead-in question, and five plausible options, all of which could be correct within the clinical context, but only one represents the best (most likely) answer. Both question formats assess a wide range of topics with high reliability and allow efficient electronic marking, reducing inconsistencies in examiner interpretation. However, SBAs offer a distinct advantage in evaluating medical students by testing higher-order thinking skills, such as application and analysis of information. To better prepare for answering SBAs, it is advantageous for students to practice using question banks. However, access to these resources may be restricted due to the challenges of developing high-quality questions, the rapid evolution of medical guidelines, institutions often reserving well-developed questions for future exams and paywalls restricting access to large question banks. A potential solution to this challenge is for medical students to create their own questions as part of their revision process. While not a novel concept, question writing by students can present challenges, particularly if they lack experience, are unfamiliar with appropriate formatting and style, or do not receive constructive feedback. Despite these challenges, engaging in question writing has been shown to provide significant benefits, including fostering a deeper understanding of the subject matter and improving performance in subsequent examinations. One potential solution to these challenges is in the use of large language models (LLMs). ChatGPT (Generative Pre-trained Transformer), an LLM developed by OpenAI, relies on transformer-based neural networks trained on over 175 billion parameters to generate responses that are both coherent and contextually relevant. This capability has enabled it to demonstrate success in tasks such as translation, question answering, and text generation. Notably, ChatGPT has shown its ability to accurately interpret complex medical information, achieving performance at or near the pass threshold in all three steps of the USMLE without specific prompting—a feat no other artificial intelligence (AI) model has matched. Furthermore, ChatGPT’s potential utility in medical education has been highlighted by its capacity to generate justifications for correct answers to SBA questions. Given the evidence that ChatGPT is capable of interpreting, answering, and justifying SBA questions designed for medical students, and Zuckerman et al have already highlighted ChatGPT’s ability to rapidly produce plausible MCQ’s , our objective is to explore its ability to generate its own SBA questions when used by students themselves. This uniformity of the SBA questions that make up the MLA underscores the potential for ChatGPT to assist in producing high-quality, exam-relevant questions. If such questions can be reliably generated with ChatGPT's help, students could create tailored revision resources, enhance their learning experience, and improve their preparation for the AKT. Methods Fifty SBA questions were independently written by two final-year medical students, designed to emulate the content and format of the AKT. The GMC’s Outcomes for Graduates and the MLA Content Map were utilized to identify a representative selection of medical subject areas and learning outcomes, from which the SBA questions were developed. These topics were evenly divided between the two students. Each SBA question consisted of a preceding clinical scenario, followed by a question stem and five answer options. To ensure consistency, guidance from Harris, Walsh, and Smith was applied. Multiple ChatGPT prompts were initially tested to determine which approach enabled the model to produce SBA questions that most closely resembled the desired structure. The following prompt was used in the development of the question bank: “Please write me a single best answer question, for a final year medical student in the United Kingdom. It should contain a clinical scenario (the stem), followed by a lead-in (the question) and use British English terms. It should have 5 options in alphabetical order. Please highlight the correct option with justification. Please link relevant guidelines where appropriate. The question should be about ____” Each question generated was then immediately reviewed by the medical students and subsequently rewritten or edited using additional ChatGPT prompts. These further prompts enabled the authors to refine the focus of the questions (e.g., shifting from diagnosis to management), increase the level of difficulty, and ensure that the questions evaluated higher-order thinking skills. Upon completion of each ChatGPT-generated SBA question, the content, answers, and explanations were reviewed against current medical guidelines and resources. The authoring medical student made specific edits to address inconsistent wording, correct inaccuracies in the answers, and refine elements of the clinical scenarios. Each student then reviewed and provided feedback on the other's set of 50 questions, resulting in a final 100-question paper. This collection was subsequently scrutinised and edited by an experienced member of the university’s assessment team, who has expertise in SBA writing. This final review ensured that the questions were standardised, realistic, and appropriate for final-year assessments, while also maintaining accurate medical content. The 100-piece mock exam was subsequently purchased by a leading medical education provider with over 300000 users for inclusion in their UKMLA question bank. Results ChatGPT produced the basis for 100 SBA questions which, after review and appropriate edits, closely mirrored the style and content used in the MLA AKT. ChatGPT was capable of generating initial question ideas, which with further prompting, addressed a broad range of learning outcomes, including: · Generating management plans for acute and chronic conditions · Understanding pathophysiology of disease · Handling complexity and uncertainty · Requesting and interpreting relevant investigations · Safe prescribing · Understanding patient capacity, consent, and confidentiality An example SBA question, randomly selected from the bank of 100 that was produced, is shown below: ‘A 65-year-old right-handed patient presents to the emergency department after an accidental injury to the left arm caused by pruning shears whilst gardening. There is a deep laceration located on the dorsal aspect of the forearm, approximately 5 cm proximal to the wrist crease. She is unable to adduct or abduct the fingers on the affected hand. Question Lead-in: Which anatomical structure is most likely injured? Options: A) Abductor pollicis longus tendon B) Radial nerve C) Extensor digitorum tendon D) Median nerve E) Ulnar nerve Correct Option: E) Ulnar nerve Justification: The inability to abduct the fingers can be attributed to the dysfunction of the ulnar nerve, which innervates the dorsal interossei muscles responsible for finger abduction. Radial nerve injury primarily affects wrist and finger extension. Median nerve injury typically leads to deficits in thumb opposition and flexion of the lateral fingers. Damage to the abductor pollicis longus tendon and extensor digitorum tendon would not directly lead to the observed presentation.’ ChatGPT frequently produced flawed questions, with common pitfalls including: · Oversimplified or overly limited question stems · Missing critical information necessary for making a diagnosis · Insufficient depth and complexity in the questions · Incorrect information derived from outdated guidelines · Bias or prejudice evident in the question stems · Blood test results presented in undesired or inconsistent units Without specific prompting, ChatGPT did not produce questions that included clinical images or results from investigations such as X-rays, blood tests, or ECGs. However, with further prompting, ChatGPT was able to generate detailed tables of blood test results appropriate for the scenario, as well as written reports of clinical findings from radiologic, hematologic, and pathologic investigations. Although ChatGPT failed to generate original medical images such as X-rays or ECGs, it can provide links to appropriate websites containing such images when prompted. For example, using the prompt, "Please include a link to a relevant website showing the clinical image/investigation that I can include in the stem," allows the writer to locate and use the image for personal purposes. Figure 1. Discussion ChatGPT can accurately and consistently produce SBA-style questions when provided with an appropriate initial prompt and supplemented with further prompts to add variety, make corrections, and increase complexity. This capability has significant implications for medical students. By engaging with ChatGPT to create SBA-style questions, students can deepen their understanding of the question-writing process and enhance their ability to answer such questions. This approach may improve comprehension and retention of the medical topics they are exploring while also facilitating the creation of a substantial bank of revision questions. The complexity and accuracy of ChatGPT generated questions, when reviewed and amended by a clinician, is comparable with existing educational resources. This is evidenced by the addition of our mock exam in the question bank of a sector leading, medical education provider. Collectively, these benefits provide a valuable resource for preparing for the AKT. As previously demonstrated, question writing promotes a deeper understanding of medical topics and improved performance in examinations for medical students. iv However, this process can be time-consuming and burdensome. ChatGPT offers the framework of an accurate clinical scenario and question, which, with further refinement, can be shaped into a high-quality SBA. The time saved in generating the initial question stem and answer options can instead be spent verifying the accuracy of the scenario, answers, and justifications. For medical students, these benefits enable the creation of robust formative assessment banks tailored to their learning needs. The use of formative assessment has been shown to improve student performance and foster a positive attitude toward learning in large cohorts of medical students. Medical students value formative feedback as a means of identifying knowledge gaps and increasing learn, especially when involved in generating their own assessment content. Further research is needed to explore how large groups of medical students can engage with ChatGPT as a tool for creating formative assessments. Figure 2 There are, however, several limitations to ChatGPT’s ability to generate SBA-style questions for student learning. Firstly, and arguably most importantly, ChatGPT is not infallible in its recall of factual information. Due to the dynamic nature of medical guidelines, we observed that it struggles to remain current with the latest updates. Additionally, ChatGPT may produce "hallucinations," where it generates factually incorrect information or cites research papers that do not exist. Writers using ChatGPT must exercise caution and verify all content for accuracy. Secondly, relying solely on the initial prompt often results in questions that lack scope, depth, or complexity. To produce SBA questions that are suitable for medical student learning, we propose the following steps: Begin with a predefined prompt that directs ChatGPT to generate the initial skeleton of a question in the desired format. Request a question about a broad topic (e.g., cardiology rather than Wolff-Parkinson-White syndrome) and then make supplementary requests to refine specific sections of the clinical scenario, adjust the difficulty, or rewrite the question as needed. Specify that both the answer and justification must be based on relevant and current guidelines. Review each question thoroughly, revising as necessary, with oversight from both the initial writer and an experienced educator to ensure accuracy, consistency, and appropriateness. In addition, we recommend that medical students experiment with the proposed prompt and adapt it to meet their individual needs. Educators should collaborate with students to assist in the creation of student-directed question banks, ensuring that questions align with learning objectives. Furthermore, educators and students could consider organising group writing sessions, where multiple students each contribute a small number of questions in one sitting. This approach could streamline the creation of formative assessment papers, allowing for the rapid development of a valuable learning resource. By following these recommendations, students can leverage ChatGPT to enhance their preparation for the AKT, the assessment required for medical registration. Declarations Conflict of Interest: The authors acknowledge that they received financial payment for the purchase of the question bank that they produced, by an unnamed medical education provider. The work was not commissioned or affected by its subsequent purchase. The purchasing organisation had no input in the direction or method of question production, nor the conclusions documented in this paper. The authors alone are responsible for the content and writing of this article. Human Ethics and Consent to Participate: No human participants were involved in the duration of this study. Data Availability: Data sharing is not applicable to this article as no datasets were generated or analysed during the current study. The question bank developed by the authors is subject to copyright held by the purchasing organisation. Access to the materials may be granted for reference purposes upon reasonable request. Funding: The authors received no funding to support the research described in this paper. Consent to participate: Full consent to participate is provided. Consent for publication: Not Applicable. Authors’ contributions: Dr McGrane and Dr Gore produced the questions described in this paper. Dr Barnes oversaw the review of the produced questions. Dr McGrane was the primary author of the manuscript with Dr Barnes and Dr Gore acting as editors. References Medical Schools Council. (2024) UK Medical Schools Applied Knowledge Test Student Handbook. ms-akt-student-handbook-24-25.pdf Sam, A. H., Westacott, R., Gurnell, M., Wilson, R., Meeran, K., & Brown, C. (2019). Comparing single-best-answer and very-short-answer questions for the assessment of applied medical knowledge in 20 UK medical schools: Cross-sectional study. BMJ open, 9(9), e032550. https://doi.org/10.1136/bmjopen-2019-032550 Abdul Rahim AF, Simok AA, Abdull Wahab SF. (2022). A guide for writing single best answer questions to assess higher-order thinking skills based on learning outcomes. Education in Medicine Journal. 14(2):111–124. https://doi.org/10.21315/eimj2022.14.2.9 Herrero, J.I., Lucena, F. & Quiroga, J. (2019). Randomized study showing the benefit of medical study writing multiple choice questions on their learning. BMC Med Educ 19, 42 https://doi.org/10.1186/s12909-019-1469-2 Brown T, Mann B, Ryder N, et al. (2020). Language models are few-shot learners. Adv Neural Inf Process Syst. 33:1877–1901. https://arxiv.org/abs/2005. 14165 Tong, L., Wang, J., Rapaka, S., & Garg, P. S. (2024). Can ChatGPT generate practice question explanations for medical students, a new faculty teaching tool? Medical Teacher, 1–5. https://doi.org/10.1080/0142159X.2024.2363486 Zuckerman M, Flood R, Tan RJB, Kelp N, Ecker DJ, Menke J, Lockspeiser T. ChatGPT for assessment writing. Med Teach. 2023 Nov;45(11):1224-1227. doi: 10.1080/0142159X.2023.2249239. Epub 2023 Oct 16. PMID: 37789636. J L Walsh, B H L Harris, P E Smith. (February 2017). Single best answer question-writing tips for clinicians. Postgraduate Medical Journal. Volume 93, Issue 1096, Pages 76–81. https://doi.org/10.1136/postgradmedj-2015-133893 Dr Govinddas G Akbari, Dr Yogesh Umraniya, Dr Prashant Kariya, Dr Kiran Thorat, Dr Jyothi Vybhavi V S, Dr Milav H Bhavsar, Dr Jitendra Patel. (2024 Apr. 10). Analysis of Impact of Regular Formative Assessment on Final Summative Assessment Among MBBS Students. Kuey [Internet]. 30(4):1013-6. https://kuey.net/index.php/kuey/article/view/1602 Lameris, A.L., Hoenderop, J.G., Bindels, R.J. et al. (2015). The impact of formative testing on study behaviour and study performance of (bio)medical students: a smartphone application intervention study. BMC Med Educ 15, 72 https://doi.org/10.1186/s12909-015-0351-0 Ma, T., Li, Y., Yuan, H. et al. (2023). Reflection on the teaching of student-centred formative assessment in medical curricula: an investigation from the perspective of medical students. BMC Med Educ 23, 141 https://doi.org/10.1186/s12909-023-04110-w Lakhtakia, R., Otaki, F., Alsuwaidi, L., & Zary, N. (2022). Assessment as Learning in Medical Education: Feasibility and Perceived Impact of Student-Generated Formative Assessments. JMIR medical education, 8(3), e35820. https://doi.org/10.2196/35820 Masters K. (2023). Medical Teacher's first ChatGPT's referencing hallucinations: Lessons for editors, reviewers, and teachers. Medical teacher, 45(7), 673–675. https://doi.org/10.1080/0142159X.2023.2208731 Additional Declarations Competing interest reported. The authors acknowledge that they received financial payment for the purchase of the question bank that they produced, by an unnamed medical education provider. The work was not commissioned or affected by its subsequent purchase. The purchasing organisation had no input in the direction or method of question production, nor the conclusions documented in this paper. The authors alone are responsible for the content and writing of this article. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 10 May, 2026 Reviews received at journal 23 Apr, 2026 Reviewers agreed at journal 21 Apr, 2026 Reviewers invited by journal 16 Apr, 2026 Editor assigned by journal 16 Apr, 2026 Editor invited by journal 16 Apr, 2026 Submission checks completed at journal 15 Apr, 2026 First submitted to journal 15 Apr, 2026 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-9361324","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":628514657,"identity":"67bb0b34-6e57-4aae-ad08-e8ba30e711b6","order_by":0,"name":"Daniel McGrane","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxklEQVRIiWNgGAWjYDCCAzAGewMbqVp4DkC08BCvRSKBSC18B7gTP37dY5fHL/n22MMfDHfk7AlpkTzAu1la5llyseTsvHRjHoZnxgRtMTjAu41Z4gBz4obbOWbSDAyHE3uI1FKfuP/mGTPJHwyH64nSwvjhwOHEDRI8ZhI8DIcTCDpM8jDQLwwHjifOOAN0GI/BYcOeAwS08B3v3fjxx4HqxP52kMMqDsuzNxCyhhmIEG4xIKQcChh/EKlwFIyCUTAKRigAANt8PvNHetQgAAAAAElFTkSuQmCC","orcid":"","institution":"Chesterfield Royal Hospital NHS Foundation Trust","correspondingAuthor":true,"prefix":"","firstName":"Daniel","middleName":"","lastName":"McGrane","suffix":""},{"id":628514659,"identity":"78c4abb1-bfaa-4cdb-8b2b-be3a85007ec8","order_by":1,"name":"Alice Gore","email":"","orcid":"","institution":"St George’s University Hospitals NHS Foundation Trust","correspondingAuthor":false,"prefix":"","firstName":"Alice","middleName":"","lastName":"Gore","suffix":""},{"id":628514660,"identity":"b76713ed-dc67-44cd-8452-6a29be36de7b","order_by":2,"name":"Isaac Barnes","email":"","orcid":"","institution":"Kings Mill Hospital","correspondingAuthor":false,"prefix":"","firstName":"Isaac","middleName":"","lastName":"Barnes","suffix":""}],"badges":[],"createdAt":"2026-04-08 22:38:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9361324/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9361324/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107750694,"identity":"01729a91-4185-472d-9706-9369bda23437","added_by":"auto","created_at":"2026-04-24 17:09:55","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":104520,"visible":true,"origin":"","legend":"\u003cp\u003erecommendations for effectively editing ChatGPT produced questions.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9361324/v1/cdae3756a0c3003645d2b73d.png"},{"id":107750702,"identity":"ad632b7e-ad40-4f97-80a5-ebd0707a6ead","added_by":"auto","created_at":"2026-04-24 17:09:57","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":28220,"visible":true,"origin":"","legend":"\u003cp\u003eA comparison between ChatGPT generated and manually written SBA questions.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9361324/v1/6fcee4d9af43b7a6f5925826.png"},{"id":107869122,"identity":"b1971b7b-f1a2-4338-9287-affe98f6f5c4","added_by":"auto","created_at":"2026-04-27 07:36:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":242704,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9361324/v1/efc7625a-ea5b-4f1a-950c-ac2e87ad13cb.pdf"}],"financialInterests":"Competing interest reported. The authors acknowledge that they received financial payment for the purchase of the question bank that they produced, by an unnamed medical education provider. The work was not commissioned or affected by its subsequent purchase. The purchasing organisation had no input in the direction or method of question production, nor the conclusions documented in this paper. The authors alone are responsible for the content and writing of this article.","formattedTitle":"Harnessing ChatGPT for SBA Question Writing: A Novel Approach to Medical Student Learning","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWith the introduction of the Medical Licensing Assessment (MLA), all medical students graduating from UK medical schools from 2024 onwards will be required to pass the Applied Knowledge Test (AKT). This test comprises two 100-item papers featuring Single Best Answer (SBA) questions and constitutes 50% of the MLA.\u003ca class=\"FNLink\" href=\"#Fn1\" id=\"#FNLinkFn1\"\u003e\u003c/a\u003e To prepare effectively for the AKT, medical students must develop a strong understanding of SBA question formats, gain extensive practice in answering them, and acquire comprehensive knowledge of the curriculum content.\u003c/p\u003e \u003cp\u003eSBA questions differ from traditional multiple-choice questions (MCQs), which typically test recall by presenting one correct answer among five options. In contrast, SBAs include a clinical vignette, a lead-in question, and five plausible options, all of which could be correct within the clinical context, but only one represents the best (most likely) answer. \u003ca class=\"FNLink\" href=\"#Fn2\" id=\"#FNLinkFn2\"\u003e\u003c/a\u003e Both question formats assess a wide range of topics with high reliability and allow efficient electronic marking, reducing inconsistencies in examiner interpretation. However, SBAs offer a distinct advantage in evaluating medical students by testing higher-order thinking skills, such as application and analysis of information.\u003ca class=\"FNLink\" href=\"#Fn3\" id=\"#FNLinkFn3\"\u003e\u003c/a\u003e\u003c/p\u003e \u003cp\u003eTo better prepare for answering SBAs, it is advantageous for students to practice using question banks. However, access to these resources may be restricted due to the challenges of developing high-quality questions, the rapid evolution of medical guidelines, institutions often reserving well-developed questions for future exams and paywalls restricting access to large question banks. A potential solution to this challenge is for medical students to create their own questions as part of their revision process. While not a novel concept, question writing by students can present challenges, particularly if they lack experience, are unfamiliar with appropriate formatting and style, or do not receive constructive feedback. Despite these challenges, engaging in question writing has been shown to provide significant benefits, including fostering a deeper understanding of the subject matter and improving performance in subsequent examinations.\u003ca class=\"FNLink\" href=\"#Fn4\" id=\"#FNLinkFn4\"\u003e\u003c/a\u003e\u003c/p\u003e \u003cp\u003eOne potential solution to these challenges is in the use of large language models (LLMs). ChatGPT (Generative Pre-trained Transformer), an LLM developed by OpenAI, relies on transformer-based neural networks trained on over 175\u0026nbsp;billion parameters to generate responses that are both coherent and contextually relevant. This capability has enabled it to demonstrate success in tasks such as translation, question answering, and text generation.\u003ca class=\"FNLink\" href=\"#Fn5\" id=\"#FNLinkFn5\"\u003e\u003c/a\u003e Notably, ChatGPT has shown its ability to accurately interpret complex medical information, achieving performance at or near the pass threshold in all three steps of the USMLE without specific prompting\u0026mdash;a feat no other artificial intelligence (AI) model has matched. Furthermore, ChatGPT\u0026rsquo;s potential utility in medical education has been highlighted by its capacity to generate justifications for correct answers to SBA questions.\u003ca class=\"FNLink\" href=\"#Fn6\" id=\"#FNLinkFn6\"\u003e\u003c/a\u003e\u003c/p\u003e \u003cp\u003eGiven the evidence that ChatGPT is capable of interpreting, answering, and justifying SBA questions designed for medical students, and Zuckerman et al have already highlighted ChatGPT\u0026rsquo;s ability to rapidly produce plausible MCQ\u0026rsquo;s\u003ca class=\"FNLink\" href=\"#Fn7\" id=\"#FNLinkFn7\"\u003e\u003c/a\u003e, our objective is to explore its ability to generate its own SBA questions when used by students themselves. This uniformity of the SBA questions that make up the MLA underscores the potential for ChatGPT to assist in producing high-quality, exam-relevant questions. If such questions can be reliably generated with ChatGPT's help, students could create tailored revision resources, enhance their learning experience, and improve their preparation for the AKT.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eFifty SBA questions were independently written by two final-year medical students, designed to emulate the content and format of the AKT. The GMC\u0026rsquo;s \u003cem\u003eOutcomes for Graduates\u003c/em\u003e and the \u003cem\u003eMLA Content Map\u003c/em\u003e were utilized to identify a representative selection of medical subject areas and learning outcomes, from which the SBA questions were developed. These topics were evenly divided between the two students. Each SBA question consisted of a preceding clinical scenario, followed by a question stem and five answer options. To ensure consistency, guidance from Harris, Walsh, and Smith\u003ca class=\"FNLink\" href=\"#Fn8\" id=\"#FNLinkFn8\"\u003e\u003c/a\u003e was applied. Multiple ChatGPT prompts were initially tested to determine which approach enabled the model to produce SBA questions that most closely resembled the desired structure. The following prompt was used in the development of the question bank:\u003c/p\u003e \u003cp\u003e\u0026ldquo;Please write me a single best answer question, for a final year medical student in the United Kingdom. It should contain a clinical scenario (the stem), followed by a lead-in (the question) and use British English terms. It should have 5 options in alphabetical order. Please highlight the correct option with justification. Please link relevant guidelines where appropriate. The question should be about ____\u0026rdquo;\u003c/p\u003e \u003cp\u003eEach question generated was then immediately reviewed by the medical students and subsequently rewritten or edited using additional ChatGPT prompts. These further prompts enabled the authors to refine the focus of the questions (e.g., shifting from diagnosis to management), increase the level of difficulty, and ensure that the questions evaluated higher-order thinking skills.\u003c/p\u003e \u003cp\u003eUpon completion of each ChatGPT-generated SBA question, the content, answers, and explanations were reviewed against current medical guidelines and resources. The authoring medical student made specific edits to address inconsistent wording, correct inaccuracies in the answers, and refine elements of the clinical scenarios. Each student then reviewed and provided feedback on the other's set of 50 questions, resulting in a final 100-question paper. This collection was subsequently scrutinised and edited by an experienced member of the university\u0026rsquo;s assessment team, who has expertise in SBA writing. This final review ensured that the questions were standardised, realistic, and appropriate for final-year assessments, while also maintaining accurate medical content.\u003c/p\u003e \u003cp\u003eThe 100-piece mock exam was subsequently purchased by a leading medical education provider with over 300000 users for inclusion in their UKMLA question bank.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eChatGPT produced the basis for 100 SBA questions which, after review and appropriate edits, closely mirrored the style and content used in the MLA AKT. ChatGPT was capable of generating initial question ideas, which with further prompting, addressed a broad range of learning outcomes, including:\u003c/p\u003e\n\u003cp\u003e· Generating management plans for acute and chronic conditions\u003c/p\u003e\n\u003cp\u003e· Understanding pathophysiology of disease\u003c/p\u003e\n\u003cp\u003e· Handling complexity and uncertainty\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e· Requesting and interpreting relevant investigations\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e· Safe prescribing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e· Understanding patient capacity, consent, and confidentiality\u003c/p\u003e\n\u003cp\u003eAn example SBA question, randomly selected from the bank of 100 that was produced, is shown below:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e‘A 65-year-old right-handed patient presents to the emergency department after an accidental injury to the left arm caused by pruning shears whilst gardening. There is a deep laceration located on the dorsal aspect of the forearm, approximately 5 cm proximal to the wrist crease. She is unable to adduct or abduct the fingers on the affected hand.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eQuestion Lead-in:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWhich anatomical structure is most likely injured?\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eOptions:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eA)\u0026nbsp; \u0026nbsp;Abductor pollicis longus tendon\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eB)\u0026nbsp; \u0026nbsp;Radial nerve\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eC)\u0026nbsp; \u0026nbsp;Extensor digitorum tendon\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eD)\u0026nbsp; \u0026nbsp;Median nerve\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eE)\u0026nbsp; \u0026nbsp;Ulnar nerve\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCorrect Option:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eE)\u0026nbsp; \u0026nbsp;Ulnar nerve\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eJustification:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe inability to abduct the fingers can be attributed to the dysfunction of the ulnar nerve, which innervates the dorsal interossei muscles responsible for finger abduction. Radial nerve injury primarily affects wrist and finger extension. Median nerve injury typically leads to deficits in thumb opposition and flexion of the lateral fingers. Damage to the abductor pollicis longus tendon and extensor digitorum tendon would not directly lead to the observed presentation.’\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eChatGPT frequently\u0026nbsp;produced flawed questions, with common pitfalls including:\u003c/p\u003e\n\u003cp\u003e· Oversimplified or overly limited question stems\u003c/p\u003e\n\u003cp\u003e· Missing critical information necessary for making a diagnosis\u003c/p\u003e\n\u003cp\u003e· Insufficient depth and complexity in the questions\u003c/p\u003e\n\u003cp\u003e· Incorrect information derived from outdated guidelines\u003c/p\u003e\n\u003cp\u003e· Bias or prejudice evident in the question stems\u003c/p\u003e\n\u003cp\u003e· Blood test results presented in undesired or inconsistent units\u003c/p\u003e\n\u003cp\u003eWithout specific prompting, ChatGPT did not produce questions that included clinical images or results from investigations such as X-rays, blood tests, or ECGs. However, with further prompting, ChatGPT was able to generate detailed tables of blood test results appropriate for the scenario, as well as written reports of clinical findings from radiologic, hematologic, and pathologic investigations. Although ChatGPT failed to generate original medical images such as X-rays or ECGs, it can provide links to appropriate websites containing such images when prompted. For example, using the prompt, \u003cem\u003e\"Please include a link to a relevant website showing the clinical image/investigation that I can include in the stem,\"\u003c/em\u003e allows the writer to locate and use the image for personal purposes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 1.\u003c/strong\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eChatGPT can accurately and consistently produce SBA-style questions when provided with an appropriate initial prompt and supplemented with further prompts to add variety, make corrections, and increase complexity. This capability has significant implications for medical students. By engaging with ChatGPT to create SBA-style questions, students can deepen their understanding of the question-writing process and enhance their ability to answer such questions. This approach may improve comprehension and retention of the medical topics they are exploring while also facilitating the creation of a substantial bank of revision questions. The complexity and accuracy of ChatGPT generated questions, when reviewed and amended by a clinician, is comparable with existing educational resources. This is evidenced by the addition of our mock exam in the question bank of a sector leading, medical education provider. Collectively, these benefits provide a valuable resource for preparing for the AKT.\u003c/p\u003e \u003cp\u003eAs previously demonstrated, question writing promotes a deeper understanding of medical topics and improved performance in examinations for medical students.\u003csup\u003eiv\u003c/sup\u003e However, this process can be time-consuming and burdensome. ChatGPT offers the framework of an accurate clinical scenario and question, which, with further refinement, can be shaped into a high-quality SBA. The time saved in generating the initial question stem and answer options can instead be spent verifying the accuracy of the scenario, answers, and justifications.\u003c/p\u003e \u003cp\u003eFor medical students, these benefits enable the creation of robust formative assessment banks tailored to their learning needs. The use of formative assessment has been shown to improve student performance\u003ca class=\"FNLink\" href=\"#Fn9\" id=\"#FNLinkFn9\"\u003e\u003c/a\u003e and foster a positive attitude toward learning in large cohorts of medical students.\u003ca class=\"FNLink\" href=\"#Fn10\" id=\"#FNLinkFn10\"\u003e\u003c/a\u003e Medical students value formative feedback as a means of identifying knowledge gaps and increasing learn,\u003ca class=\"FNLink\" href=\"#Fn11\" id=\"#FNLinkFn11\"\u003e\u003c/a\u003e especially when involved in generating their own assessment content.\u003ca class=\"FNLink\" href=\"#Fn12\" id=\"#FNLinkFn12\"\u003e\u003c/a\u003e Further research is needed to explore how large groups of medical students can engage with ChatGPT as a tool for creating formative assessments.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThere are, however, several limitations to ChatGPT\u0026rsquo;s ability to generate SBA-style questions for student learning. Firstly, and arguably most importantly, ChatGPT is not infallible in its recall of factual information. Due to the dynamic nature of medical guidelines, we observed that it struggles to remain current with the latest updates. Additionally, ChatGPT may produce \"hallucinations,\" where it generates factually incorrect information or cites research papers that do not exist.\u003ca class=\"FNLink\" href=\"#Fn13\" id=\"#FNLinkFn13\"\u003e\u003c/a\u003e Writers using ChatGPT must exercise caution and verify all content for accuracy. Secondly, relying solely on the initial prompt often results in questions that lack scope, depth, or complexity.\u003c/p\u003e \u003cp\u003eTo produce SBA questions that are suitable for medical student learning, we propose the following steps:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eBegin with a predefined prompt that directs ChatGPT to generate the initial skeleton of a question in the desired format.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eRequest a question about a broad topic (e.g., cardiology rather than Wolff-Parkinson-White syndrome) and then make supplementary requests to refine specific sections of the clinical scenario, adjust the difficulty, or rewrite the question as needed.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eSpecify that both the answer and justification must be based on relevant and current guidelines.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eReview each question thoroughly, revising as necessary, with oversight from both the initial writer and an experienced educator to ensure accuracy, consistency, and appropriateness.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eIn addition, we recommend that medical students experiment with the proposed prompt and adapt it to meet their individual needs. Educators should collaborate with students to assist in the creation of student-directed question banks, ensuring that questions align with learning objectives. Furthermore, educators and students could consider organising group writing sessions, where multiple students each contribute a small number of questions in one sitting. This approach could streamline the creation of formative assessment papers, allowing for the rapid development of a valuable learning resource. By following these recommendations, students can leverage ChatGPT to enhance their preparation for the AKT, the assessment required for medical registration.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eConflict of Interest:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors acknowledge that they received financial payment for the purchase of the question bank that they produced, by an unnamed medical education provider. The work was not commissioned or affected by its subsequent purchase. The purchasing organisation had no input in the direction or method of question production, nor the conclusions documented in this paper. The authors alone are responsible for the content and writing of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHuman Ethics and Consent to Participate:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo human participants were involved in the duration of this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData sharing is not applicable to this article as no datasets were generated or analysed during the current study.\u0026nbsp;The question bank developed by the authors is subject to copyright held by the purchasing organisation. Access to the materials may be granted for reference purposes upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors received no funding to support the research described in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFull consent to participate is provided.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot Applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors’ contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDr McGrane and Dr Gore produced the questions described in this paper.\u003c/p\u003e\n\u003cp\u003eDr Barnes oversaw the review of the produced questions.\u003c/p\u003e\n\u003cp\u003eDr McGrane was the primary author of the manuscript with Dr Barnes and Dr Gore acting as editors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eMedical Schools Council. (2024) UK Medical Schools Applied Knowledge Test Student Handbook. ms-akt-student-handbook-24-25.pdf\u003c/li\u003e\n \u003cli\u003eSam, A. H., Westacott, R., Gurnell, M., Wilson, R., Meeran, K., \u0026amp; Brown, C. (2019). Comparing single-best-answer and very-short-answer questions for the assessment of applied medical knowledge in 20 UK medical schools: Cross-sectional study. BMJ open, 9(9), e032550. https://doi.org/10.1136/bmjopen-2019-032550\u003c/li\u003e\n \u003cli\u003eAbdul Rahim AF, Simok AA, Abdull Wahab SF. (2022). A guide for writing single best answer questions to assess higher-order thinking skills based on learning outcomes. Education in Medicine Journal. 14(2):111\u0026ndash;124. https://doi.org/10.21315/eimj2022.14.2.9\u003c/li\u003e\n \u003cli\u003eHerrero, J.I., Lucena, F. \u0026amp; Quiroga, J. (2019). Randomized study showing the benefit of medical study writing multiple choice questions on their learning. BMC Med Educ 19, 42 https://doi.org/10.1186/s12909-019-1469-2\u003c/li\u003e\n \u003cli\u003eBrown T, Mann B, Ryder N, et al. (2020). Language models are few-shot learners. Adv Neural Inf Process Syst. 33:1877\u0026ndash;1901. https://arxiv.org/abs/2005. 14165\u003c/li\u003e\n \u003cli\u003eTong, L., Wang, J., Rapaka, S., \u0026amp; Garg, P. S. (2024). Can ChatGPT generate practice question explanations for medical students, a new faculty teaching tool? Medical Teacher, 1\u0026ndash;5. https://doi.org/10.1080/0142159X.2024.2363486\u003c/li\u003e\n \u003cli\u003eZuckerman M, Flood R, Tan RJB, Kelp N, Ecker DJ, Menke J, Lockspeiser T. ChatGPT for assessment writing. Med Teach. 2023 Nov;45(11):1224-1227. doi: 10.1080/0142159X.2023.2249239. Epub 2023 Oct 16. PMID: 37789636.\u003c/li\u003e\n \u003cli\u003eJ L Walsh, B H L Harris, P E Smith. (February 2017). Single best answer question-writing tips for clinicians. Postgraduate Medical Journal. Volume 93, Issue 1096, Pages 76\u0026ndash;81. https://doi.org/10.1136/postgradmedj-2015-133893\u003c/li\u003e\n \u003cli\u003eDr Govinddas G Akbari, Dr Yogesh Umraniya, Dr Prashant Kariya, Dr Kiran Thorat, Dr Jyothi Vybhavi V S, Dr Milav H Bhavsar, Dr Jitendra Patel. (2024 Apr. 10). Analysis of Impact of Regular Formative Assessment on Final Summative Assessment Among MBBS Students. Kuey [Internet]. 30(4):1013-6. https://kuey.net/index.php/kuey/article/view/1602\u003c/li\u003e\n \u003cli\u003eLameris, A.L., Hoenderop, J.G., Bindels, R.J. et al. (2015). The impact of formative testing on study behaviour and study performance of (bio)medical students: a smartphone application intervention study. BMC Med Educ 15, 72 https://doi.org/10.1186/s12909-015-0351-0\u003c/li\u003e\n \u003cli\u003eMa, T., Li, Y., Yuan, H. et al. (2023). Reflection on the teaching of student-centred formative assessment in medical curricula: an investigation from the perspective of medical students. BMC Med Educ 23, 141 https://doi.org/10.1186/s12909-023-04110-w\u003c/li\u003e\n \u003cli\u003eLakhtakia, R., Otaki, F., Alsuwaidi, L., \u0026amp; Zary, N. (2022). Assessment as Learning in Medical Education: Feasibility and Perceived Impact of Student-Generated Formative Assessments. JMIR medical education, 8(3), e35820. https://doi.org/10.2196/35820\u003c/li\u003e\n \u003cli\u003eMasters K. (2023). Medical Teacher\u0026apos;s first ChatGPT\u0026apos;s referencing hallucinations: Lessons for editors, reviewers, and teachers. Medical teacher, 45(7), 673\u0026ndash;675. https://doi.org/10.1080/0142159X.2023.2208731\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"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":"Assessment, Artificial Intelligence, ChatGPT, Medical Education","lastPublishedDoi":"10.21203/rs.3.rs-9361324/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9361324/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe introduction of the Medical Licensing Assessment (MLA) requires UK medical graduates to pass two 100-item single best answer (SBA) question papers. Access to high-quality question banks is often restricted by expensive paywalls, evolving guidelines, and institutional safeguarding of questions for future assessments. This limits students' ability to practice exam-style questions and develop their clinical reasoning.\u003c/p\u003e \u003cp\u003eDuring this study, two medical students used ChatGPT to produce a bank of 100 SBA questions which were representative of those used in the MLA. Questions were analysed and revised by a student and a member of a university assessment team. A critical review of ChatGPT\u0026rsquo;s benefits was conducted to determine the scope for application in both self-directed learning and formal medical school assessment.\u003c/p\u003e \u003cp\u003eChatGPT successfully formulates clinical vignettes and accompanying SBA questions in less than 50% of the time taken to generate a comparable question, when written by a clinical educator. It is, however, not yet reliable enough to be used without expert supervision or direct correlation with current medical guidelines.\u003c/p\u003e \u003cp\u003eBoth medical students and clinical educators will benefit from ChatGPT\u0026rsquo;s ability to rapidly compose SBA questions that emulate those used in medical school exams, provided they remain conscious of the risk of false clinical recommendations.\u003c/p\u003e \u003cp\u003eThe acquisition and implementation of our question bank by a leading commercial medical education provider confirms the value of ChatGPT generated SBA questions as a resource for medical students preparing for the UKMLA.\u003c/p\u003e","manuscriptTitle":"Harnessing ChatGPT for SBA Question Writing: A Novel Approach to Medical Student Learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-24 17:09:38","doi":"10.21203/rs.3.rs-9361324/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"210298446995072731101617878609128451527","date":"2026-05-11T01:02:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-23T12:10:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"35335487503677579067614746897742878581","date":"2026-04-21T21:36:13+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-16T20:58:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-16T20:55:29+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-16T12:14:02+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-15T17:45:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Education","date":"2026-04-15T17:41:15+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":"4d176fea-a13b-44c7-9ba2-6ad817eabde1","owner":[],"postedDate":"April 24th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"210298446995072731101617878609128451527","date":"2026-05-11T01:02:24+00:00","index":56,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-24T17:09:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-24 17:09:38","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9361324","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9361324","identity":"rs-9361324","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}