Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Wednesday, July 01, 2020 at 8:00 PM to 10:00 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

Technology, innovation, and openness in medical education in the information age.

Latest Submissions Open for Peer Review

JMIR has been a leader in applying openness, participation, collaboration and other "2.0" ideas to scholarly publishing, and since December 2009 offers open peer review articles, allowing JMIR users to sign themselves up as peer reviewers for specific articles currently considered by the Journal (in addition to author- and editor-selected reviewers). Note that this is a not a complete list of submissions as authors can opt-out. The list below shows recently submitted articles where submitting authors have not opted-out of open peer-review and where the editor has not made a decision yet. (Note that this feature is for reviewing specific articles - if you just want to sign up as reviewer (and wait for the editor to contact you if articles match your interests), please sign up as reviewer using your profile).
To assign yourself to an article as reviewer, you must have a user account on this site (if you don't have one, register for a free account here) and be logged in (please verify that your email address in your profile is correct). Add yourself as a peer reviewer to any article by clicking the '+Peer-review Me!+' link under each article. Full instructions on how to complete your review will be sent to you via email shortly after. Do not sign up as peer-reviewer if you have any conflicts of interest (note that we will treat any attempts by authors to sign up as reviewer under a false identity as scientific misconduct and reserve the right to promptly reject the article and inform the host institution).
The standard turnaround time for reviews is currently 2 weeks, and the general aim is to give constructive feedback to the authors and/or to prevent publication of uninteresting or fatally flawed articles. Reviewers will be acknowledged by name if the article is published, but remain anonymous if the article is declined.
The abstracts on this page are unpublished studies - please do not cite them (yet). If you wish to cite them/wish to see them published, write your opinion in the form of a peer-review!

Tip: Include the RSS feed of the JMIR submissions on this page on your iGoogle homepage, blog, or desktop RSS reader to stay informed about current submissions!

JMIR Submissions under Open Peer Review

↑ Grab this Headline Animator


If you follow us on Twitter, we will also announce new submissions under open peer-review there.

Titles/Abstracts of Articles Currently Open for Review

Titles/Abstracts of Articles Currently Open for Review:

  • Application of Immersive VR, AR, and MR Training Systems in Acupuncture Education: A scoping review

    Date Submitted: Feb 12, 2026
    Open Peer Review Period: Feb 12, 2026 - Apr 9, 2026

    Background: Acupuncture is one of the most commonly used interventions for pain management, particularly within traditional Chinese medicine (TCM) and Korean medicine (KM). It is a therapeutic technique that involves the insertion of fine needles into anatomically defined points on the body. Accurately locating and needling these acupoints is inherently difficult due to the subtle anatomical variations and the precision required in point identification. Objective: This scoping review maps the landscape of immersive Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR) in acupuncture education, aiming to address the safety and standardization limitations of traditional training. Methods: We conducted a comprehensive search across six databases, including PubMed and CNKI, for studies applying immersive XR technologies to acupuncture training. Twelve primary studies published between 2008 and 2025 were selected for analysis. Results: VR and MR were primarily utilized for visualizing needling depth and internal anatomy, whereas AR demonstrated high utility for surface acupoint localization. Recent advancements include markerless tracking and haptic feedback integration, although challenges regarding hardware accessibility and tactile realism persist. Conclusions: Immersive XR technologies provide a safe and interactive environment for standardized acupuncture skills acquisition. Future development should focus on enhancing haptic fidelity and expanding anatomical scope to better bridge the gap between virtual training and clinical practice.

  • Background: The integration of artificial intelligence（AI） technology and knowledge graphs（KG） in education offers novel possibilities for pedagogical innovation. This study aims to construct and evaluate the application effectiveness of an "AI+Knowledge Graph" teaching model based on the ARCS motivation model in teaching Integrated Chinese and Western Oncology, exploring its role in enhancing students' academic performance, self-directed learning ability, and learning engagement level. Objective: Construction and Evaluation of an "AI+Knowledge Graph" Teaching Model Based on the ARCS Motivation Model Methods: One hundred undergraduate medical students were randomly allocated to an experimental group (n=50) and a control group (n=50). The experimental group adopted the "AI+Knowledge Graph" teaching model based on the ARCS motivation model, while the control group adopted the traditional teaching model. Differences in educational outcomes were systematically assessed using examinations, self-directed learning scales, learning engagement scales, and satisfaction questionnaires. Results: The experimental group demonstrated significant superiority in total score, final exam score, usual performance, and all sub-dimensions (learning and thinking, collaboration and innovation, diagnosis and summary) compared to the control group (p<0.05). The experimental group also exhibited markedly higher levels of self-directed learning ability and learning engagement level than the control group (p<0.05). Students expressed overall satisfaction with the "AI+Knowledge Graph" teaching model based on the ARCS motivation model and provided positive feedback. Conclusions: This study demonstrated that the "AI+Knowledge Graph" teaching model based on the ARCS motivation model effectively enhances students' academic performance, self-directed learning ability, and learning engagement level, exhibiting significant advantages in teaching Integrated Chinese and Western Oncology. Future research may further explore the applicability of this model across different disciplines and teaching environments, while also examining its long-term educational effects and technical optimisation pathways.

  • Background: Artificial Intelligence (AI) is increasingly integrated into healthcare, with potential to enhance disease diagnosis, treatment, and patient outcomes. However, successful adoption relies on healthcare providers’ preparedness and trust. Objective: To evaluate French healthcare professionals’ and students’ use, concerns, and perceptions of AI, and to assess their interest in AI-related training. Methods: We conducted a cross-sectional national survey distributed via PulseLife between December 2023 and March 2025. The 12-item questionnaire assessed demographics, AI usage, confidence, perceived benefits, concerns, and training needs. Reliability and validity of the instrument were assessed using Cronbach α, exploratory and confirmatory factor analysis. Descriptive statistics and chi-squared test were performed using R (version 4.3.1). Results: A total of 1625 healthcare respondents participated, including 1212 professionals (52.9% physicians, 19.1% nurses) and 413 students. Only 6.6% reported prior AI training, while 78.3% expressed interest in receiving training. Physicians showed the highest confidence in AI (P = .003). Main concerns included algorithmic bias (48.2%), data transparency (40.9%), and deterioration of the doctor–patient relationship (38.6%). Anticipated benefits included improved diagnosis (47.6%), time saving (42.1%), reduced medical errors (39%). Conclusions: French healthcare providers and students remain insufficiently trained in AI, despite strong interest in acquiring such skills. Structured AI training programs and transparent regulatory frameworks are urgently needed to facilitate responsible adoption of AI in healthcare.

  • Artificial Intelligence for Clinical Competency Assessment: A Scoping Review of Methods and Applications

    Date Submitted: Feb 3, 2026
    Open Peer Review Period: Feb 6, 2026 - Apr 3, 2026

    Background: Strengthening the global health workforce is central to achieving Universal Health Coverage, yet existing approaches to measuring clinical competency remain resource-intensive, episodic, and difficult to scale, especially in low- and middle-income contexts. Recent advances in large language models (LLMs) have enabled AI-led simulated standardized patients (SSPs) that may offer scalable alternatives to traditional assessments. Objective: This study aims to systematically map and characterize the existing scope, design features, and validation approaches of AI-led SSP tools used for clinical competency assessment. Methods: We conducted a scoping review following JBI guidelines, searching MEDLINE, Embase, CINAHL, Education Source, and Web of Science from inception through June 2025. Two reviewers independently screened studies and extracted data across five domains: study characteristics and populations; frontend platform and interface features; backend AI models and architectures; user interaction and automatic feedback mechanisms; and tool evaluation methods and outcomes. Results: Between 2008 and 2025, 1,185 studies were identified and 21 studies met the inclusion criteria. Most described single-site pilot evaluations or prototype systems were developed within academic institutions in high-income countries, primarily targeting pre-licensure medical or nursing students. SSPs most commonly supported text-based, web-hosted history-taking, while simulations of physical examination, laboratory tests, diagnostic reasoning, and management planning were less common. Backend architectures relied heavily on human-authored case scripts and manually defined scoring criteria, with LLMs primarily enhancing conversational fluency rather than automating clinical reasoning or evaluation. Automated feedback and scoring were reported in approximately half of the studies and showed moderate-to-high agreement with human raters when evaluated, though validation evidence was heterogeneous and limited. Conclusions: AI-led SSPs are emerging as accessible and realistic tools for clinical competency assessment, particularly across all levels of medical education. However, current implementations remain early-stage, human-dependent, and narrowly validated, constraining their widespread use as standardized or scalable instruments for health system workforce evaluation. Advancing SSPs toward end-to-end automated assessment tools will require integrated system designs, rigorous validation, and intentional development for deployment across diverse and resource-constrained settings.

  • Research has highlighted clinicians’ lack of confidence in safely diagnosing and managing dermatological disease in patients with skin of colour (SOC).The imagery and language surrounding skin colour in medical education often presents a narrow spectrum of individual experiences and the undergraduate voice appears to be lacking from these discussions. We aimed to capture the descriptors which medical students use in relation to their own skin and how this relates to their education, to better understand how we can diversify the language used, for the benefit of the patient and future clinician. An ethically-approved, digital survey was distributed to all UK medical schools between October 1st and December 31st 2024. Participants were asked to describe their skin type at baseline, when inflamed, and in relation to the Fitzpatrick scale; rate preparedness to examine and discuss differences in skin conditions in SOC; suggest how best to describe SOC when ethnicity is unknown; and describe any experiences of unacceptable SOC terminology used in the context of medical education. The survey generated 367 responses from 21 different medical schools. Self-ascribed ethnicity included: White British/Irish/White other (48%), Black/African/Caribbean/Black British (10%), Asian/Asian British (30%) and Mixed White/Black/Asian (7%) and Other (5%). The responses demonstrated that neither the pictorial nor text versions of the Fitzpatrick scale adequately represent how medical students consider their own skin, with 56% positioning themselves outside of the six images. The top descriptors of skin colour reflected that most participants self-identified as: brown (125 mentions), pale (123) and white (116). For inflamed skin, the terms red or reddish (337) were most frequently used, followed by pink (97) and brown (27). There were differing views on the use of ethnicity-related terminology where patient ethnicity is unknown. A total of 71% and 75% of participants felt between neutral and unprepared in relation to seeing patients with diverse skin tones and discussing differences in the presentation of dermatoses in SOC with patients, respectively. Our findings suggest that skin descriptors and imagery in medical education needs to encompass greater variation in skin tone. We recommend further involvement of medical students in the diversification of undergraduate curricula, and for educators to consider the language they use to improve comprehension and preparation for clinical practice.

  • A systematic review of engagement in medical education

    Date Submitted: Jan 28, 2026
    Open Peer Review Period: Jan 30, 2026 - Mar 27, 2026

    Background: Ask any educator, and they will respond that engagement is an important factor in their teaching. However, engagement is a complex, multidimensional construct comprising behavioural, cognitive, emotional, and agentic dimensions. Despite growing interest in this area, the conceptualisation and measurement of engagement in medical education remain inconsistent. Objective: This systematic review aims to examine how engagement is defined, conceptualised, and measured in studies involving medical students. Methods: A systematic literature search was conducted in February 2025 across five databases for peer-reviewed studies published within the last decade. Studies were included if they focused on medical students, collected original data, and measured engagement within the context of a medical curriculum. Data extraction and screening were performed independently by two reviewers following PRISMA guidelines. Studies were analysed for their conceptual framework, dimensions of engagement measured, data collection methods, and study design. Results: A total of 26 studies that met the eligibility criteria were included in this systematic review. Most studies measured behavioural (n=21), cognitive (n=19), and emotional engagement (n=17), while agentic engagement was least frequently measured (n=4). Most studies employed a quantitative approach, using survey instruments (n=14) and engagement metrics (n=5) to measure engagement, while a small number of studies adopted a qualitative approach, including interviews (n=4) and observations (n=4) to measure engagement. Engagement was mainly measured as a multidimensional construct, but some studies treated it as a unidimensional construct Conclusions: Engagement remains inconsistently and often poorly defined, as evidenced by the exclusion of more than half of initially screened studies for lacking rigorous measurement of engagement. The rise of technology-driven interventions has led to an increasing interest in ensuring that students are engaged in learning to achieve the desired learning outcomes successfully. Future research should systematically incorporate behavioural, cognitive, emotional, and agentic engagement dimensions to advance understanding and enhance educational practices. Clinical Trial: Not applicable

  • Background: Traditional Problem-Based Learning (PBL) in pediatric nursing education often uses static cases and lacks personalized, real-time feedback. The integration of generative AI like ChatGPT could address these limitations, yet its systematic application in nursing internships remains understudied. Objective: To explore the effectiveness and feasibility of a ChatGPT-assisted Problem-Based Learning (PBL) model in pediatric nursing undergraduate internship education, providing empirical evidence for artificial intelligence(AI) nursing education integration. Methods: A single-center, assessor-blinded randomized controlled pilot study was conducted. Eighty-four interns were randomly assigned to the ChatGPT-PBL group (n=42) or traditional PBL group (n=42) at a 1:1 ratio. Based on traditional PBL, the experimental group integrated ChatGPT-4 to construct a "instructor-student dual-layer" supported PBL teaching framework, including dynamic generation of personalized clinical cases, provision of real-time operational feedback, and decision-making simulation training. The traditional PBL group received standardized traditional PBL teaching. The intervention lasted for 4 weeks. The primary outcome measures included theoretical assessment scores, Objective Structured Clinical Examination (OSCE) scores, Chinese Version of Critical Thinking Disposition Inventory (CTDI-CV) scores, Holistic Clinical Assessment Tool for Nursing Undergraduates (HCAT) scores, and teaching satisfaction. Results: Post-intervention, the theoretical score of the ChatGPT-PBL group was significantly higher than that of the traditional PBL group (82.76±5.02 vs 71.88±5.88, P<0.001). The ChatGPT-PBL group also showed significant advantages over the traditional PBL group in OSCE total score (43.24±2.75 vs 36.99±3.71, P<0.001), CTDI-CV total score (60.14±5.21 vs 49.87±5.74, P<0.001), and HCAT total score (51.14±3.46 vs 41.88±4.71, P<0.001). The overall satisfaction rates of the ChatGPT-PBL group with Instructors, teaching plans, and teaching content were 90.5%-95.2%, which were significantly higher than those of the traditional PBL group (64.3%-71.4%,<0.05). Conclusions: The ChatGPT-assisted PBL teaching model significantly improves the theoretical knowledge level, specialized operational skills, critical thinking ability, and clinical nursing competence of pediatric nursing undergraduate interns, with higher teaching satisfaction. It provides a replicable practical paradigm for the in-depth integration of AI and pediatric nursing education, and holds important clinical application and promotion value. Clinical Trial: The study protocol was registered in the Chinese Clinical Trial Registry (ChiCTR2500114150) .

  • Background: While Evidence-Based Medicine (EBM) is a fundamental pillar of modern healthcare, its implementation into general practice is often hindered by time constraints, resource deficits, and the inherent complexity of primary care. This challenge is further exacerbated by a lack of consensus on EBM instruction, highlighting a critical need for standardized educational frameworks. Objective: To systematically synthesize intervention studies evaluating the effectiveness of EBM training, including EBM skills, and the impact of EBM on reactions, behavioral changes, attitudes, and practices among general practitioners and residents in family medicine. Methods: We conducted a systematic synthesis of interventional studies that used the Fresno test to assess EBM skills among residents or general practitioners after educational interventions (lectures, workshops, journal clubs, or e-learning program). A comprehensive search was performed across the Cochrane Library, Embase, and Medline databases for records published between January 1980 and July 2025. Study quality was assessed using the Modified Medical Education Research Study Quality Instrument (MMERSQI), and risk of bias was evaluaAmong the 200 records screened, eight studies involving 431 participants (residents and general practitioners) met the inclusion criteria. Study designs included one randomized controlled trial, six before–after studies, and one cross-sectional study. Mean methodological quality (MMERSQI) was 65.3 (SD 7.2). One study had a low risk of bias, five had a moderate risk, and two were rated as presenting with a high risk of bias, mainly due to confounding factors and selection into analysis. Six studies reported significant improvement in Fresno test scores after training, with mean score increases ranging from 4% to 60% (p<0.05), and two found no significant change. The greatest benefits were achieved after interactive or clinically integrated sessions combining lectures, workshops, or journal clubs. Participants reported higher confidence in applying EBM (+3.2 points on the Likert scale) and greater engagement with research (+2.5 hours of reading and 3.5 additional articles per week). ted using RoB 2 for randomized studies and ROBINS-I v2 for non-randomized studies. Owing to study heterogeneity, results were synthesized qualitatively. Results: Among the 200 records screened, eight studies involving 431 participants (residents and general practitioners) met the inclusion criteria. Study designs included one randomized controlled trial, six before–after studies, and one cross-sectional study. Mean methodological quality (MMERSQI) was 65.3 (SD 7.2). One study had a low risk of bias, five had a moderate risk, and two were rated as presenting with a high risk of bias, mainly due to confounding factors and selection into analysis. Six studies reported significant improvement in Fresno test scores after training, with mean score increases ranging from 4% to 60% (p<0.05), and two found no significant change. The greatest benefits were achieved after interactive or clinically integrated sessions combining lectures, workshops, or journal clubs. Participants reported higher confidence in applying EBM (+3.2 points on the Likert scale) and greater engagement with research (+2.5 hours of reading and 3.5 additional articles per week). Conclusions: EBM training for residents and general practitioners improves both knowledge and practical application of evidence-based skills, particularly when it is interactive or clinically integrated. Evidence remains limited regarding long-term retention and patient-related outcomes.

  • Educational interventions that improve the accuracy of self-assessment in health professions: A Systematic Review

    Date Submitted: Jan 18, 2026
    Open Peer Review Period: Jan 22, 2026 - Mar 19, 2026

    Background: Existing research on the accuracy of self-assessment (SA) in health professions (HP) has shown poor accuracy of SA compared to external assessors. Objective: We systematically reviewed the evidence for educational interventions aimed at improving the accuracy of SA for technical (procedural) and non-technical (critical thinking, decision making and knowledge) Methods: We conducted this systematic review according to the PRISMA guidelines using Medline, Cochrane Library, Embase, CINAHL, AMED, ERIC, Education Source, Web of Science and Scopus databases. We included studies in English that reported on educational interventions aimed at improving the accuracy of SA versus external assessment across all health professions. A narrative synthesis of the extracted data was used using a convergent integrated approach, which reported both quantitative and qualitative data. We used the modified Medical Education Research Study Quality Instrument (MMERSQI) as the critical appraisal and bias tool to evaluate the methodological quality of included studies. Results: After abstract and full text screening of 7439 studies, we included 35 studies and 3127 participants, the majority of which were of good methodological quality. Twenty-four studies explored SA of non-technical competencies, while 11 studies explored SA of technical competencies. Health professions included medicine (n=16), dentistry (n=9), pharmacy (n=4), nursing (n=2), physiotherapy (n=2), midwifery (n=1) and occupational therapy (n=1). The accuracy of SA was improved with the use of self-assessment rubrics (11 out of 14 studies), video review for feedback (5 out of 12 studies), verbal feedback (2 of 2 studies), electronic portfolios (2 of 2 studies), simulation (2 of 2 studies), and coaching (1 of 1 study). The use of internet-based applications (1 of 1 study) and didactic learning (1 of 1 study) did not improve the accuracy of SA. Conclusions: The accuracy of self-assessment can be improved by using SA rubrics, video and verbal feedback, simulation, electronic portfolios and coaching. Limitations include a clear definition of self-assessment across research studies resulting in exclusion of systematic review. This information can be used by educators to improve the accuracy of SA within health professions education. Clinical Trial: PROSPERO (CRD42024586510)

  • Background: Artificial intelligence (AI) is rapidly integrating into health professions education (HPE) and clinical practice, creating significant opportunities alongside new ethical challenges. Although current international and professional guidance establishes essential values, it offers limited direction for how clinicians, educators, learners, and institutions should act in routine educational, research, and clinical contexts. The CARE-AI (Contextual, Accountable and Responsible Ethics for AI) project responds to this practice-level gap by articulating guidance that moves beyond values toward professional accountability and equity, with explicit attention to educational and clinical practice contexts. Objective: Our study objective was to develop and validate a consensus-based, actionable framework of principles to guide responsible AI use across health professions education, research, and clinical care. Methods: We conducted a three-phase modified Delphi consensus study, reported in accordance with the Accurate Consensus Reporting Document (ACCORD). Phase I involved two international professional meetings and three purposively sampled focus groups (AI/technology, HPE, ethics/professionalism) to adapt and refine draft principles using an exploratory qualitative approach. Phase II employed an online survey with a 5-point importance scale and prespecified consensus criteria (inclusion ≥70% high ratings; exclusion ≥70% low ratings). Phase III used include/exclude/undecided voting on revised principles. Quantitative thresholds determined consensus. Qualitative free-text comments informed iterative refinement. Results: Participants represented diverse communities of practice across health professions education, clinical care, ethics, and digital health, spanning multiple professional roles and training levels. Across all phases, 303 unique participants contributed to the study. Phase I focus groups (n=61) provided early insight and direction. In Phase II, Delphi survey round 1, 242 participants initiated the survey, with 120 completing it (49.6%). In Phase III, Delphi survey round 2, 103 participants were invited based on expressed interest at the end of Round 1; 78 initiated the survey and 75 completed it (96.2% of starters). In Phase II, 58 of 61 statements (95%) met inclusion, and participants submitted 1,887 comments (697 were content-rich), prompting clearer accountability language, stronger equity commitments, and more usable wording. In Phase III, all nine principles and their statements met inclusion. Participants contributed 224 comments (179 were content-rich) that informed final refinements. Endorsement was near-unanimous: 96% agreed or strongly agreed that the framework clearly defined professionalism expectations for AI to meet educational, technological, and ethical needs in the health professions. Conclusions: The Health CARE-AI Framework, with its preamble and nine principles, articulates actionable, consensus-validated guidance that moves from values to competence, into professional accountability, and toward structural commitments to equity. Paired with a companion implementation guide and toolkit, the framework is intended to support use across education, research, and clinical settings. Clinical Trial: Not applicable

  • Large Language Models And Examination Performance In Healthcare Education: A Bibliometric Analysis

    Date Submitted: Jan 16, 2026
    Open Peer Review Period: Jan 16, 2026 - Mar 13, 2026

    Background: Large language models (LLMs) are increasingly used and evaluated in health professions education, including studies assessing model performance on healthcare examination questions. The rapid growth and heterogeneity of this literature make it difficult to track research concentration, collaboration patterns, and emerging themes. Objective: To map publication trends, key contributors, collaboration networks, and thematic hotspots in research on LLM-supported exam solving in healthcare education. Methods: We conducted a bibliometric analysis of publications from 2023–2025. Searches were performed in PubMed, Scopus, CINAHL Ultimate (EBSCOhost), and Web of Science using structured terms for AI/LLMs (eg, ChatGPT, generative AI, large language models) combined with healthcare education and training concepts. Eligible studies addressed AI-based technologies within healthcare education or training contexts; studies focused solely on clinical practice or non-educational applications were excluded. Bibliographic metadata from PubMed (TXT) and Scopus (BIB) were merged and analyzed using bibliometrix/Biblioshiny (R) and VOSviewer to quantify productivity, collaboration (including international co-authorship), and keyword co-occurrence patterns. Results: The dataset comprised 262 documents from 158 sources, with an annual publication growth rate of 36.58% and a mean document age of 1.83 years. A total of 1,351 authors contributed (mean 5.97 co-authors per document); international co-authored publications accounted for 13.36%. Most records were journal articles (253/262), followed by letters (8/262) and one conference paper. Annual output rose from 52 (2023) to 113 (2024; +117.3%), then decreased to 97 (2025; −14.2% vs 2024) while remaining above 2023 levels. JMIR Medical Education published the most articles on this topic (34/262), followed by Scientific Reports (9/262) and BMC Medical Education (7/262). Frequent keywords included “humans” (n=144), “artificial intelligence” (n=82), “generative AI” (n=30), and “large language models” (n=20); education-focused terms such as “educational measurement/methods” were also prominent (n=76). Conclusions: Research on LLMs and exam performance in healthcare education expanded rapidly from 2023–2025, with publication activity concentrated in a limited set of journals and relatively low international collaboration. Thematic patterns emphasize assessment-related outcomes and LLM/ChatGPT performance, supporting the need for more comparable, transparent reporting (eg, prompts and model versions) and education-centered outcomes beyond accuracy in future studies. Clinical Trial: /

  • Scientific writing is a core competency in medical education and academic medicine, yet it remains a major barrier for early-career clinicians and researchers, particularly in resource-limited settings. Common challenges include limited formal training in scientific writing, heavy clinical workloads, restricted access to journals and editorial support, and difficulties writing in English as a non-native language. Recent advances in artificial intelligence (AI) have generated widespread interest as potential tools to support academic writing. However, most available guidance focuses on proprietary platforms or presents overly generic advice generated by large language models, offering limited practical value for trainees and educators working under real-world constraints. In this Viewpoint, we present a practice-informed, tool-agnostic workflow illustrating how freely accessible or freemium AI tools may be used to support scientific writing in medical research and education. Rather than claiming empirical validation or comparative performance, we offer a scholarly perspective grounded in the lived experience of medical educators and researchers who routinely supervise early-career authors. We argue that the educational value of AI lies not in content generation, but in supporting core academic skills such as literature navigation, structured reading, drafting clarity, and iterative revision. We outline key functional categories of free AI tools relevant to scientific writing, including literature discovery, reference management, PDF-based summarization, drafting and editing support, and table or figure preparation. We also address important limitations, including learning curves, internet connectivity requirements, data privacy concerns, disciplinary variability, and the risk of over-reliance on AI at the expense of critical thinking. Ethical considerations and transparency in AI use are emphasized in line with current editorial guidance. We conclude that, when used deliberately and ethically, free AI tools may help to lower barriers to scientific writing for medical trainees and early-career researchers. Their greatest educational value lies in complementing—not replacing—foundational research skills, thereby supporting more equitable

  • Knowledge, attitudes, and use of artificial intelligence by medical students: a mixed-method study

    Date Submitted: Jan 13, 2026
    Open Peer Review Period: Jan 14, 2026 - Mar 11, 2026

    Background: Artificial intelligence (AI) is transforming medicine by enhancing care and reducing administrative tasks, and facilitating research. AI also raises many concerns, including a lack of clinical context awareness, data dependence, and the absence of ethical judgment. As future practitioners, medical students must be prepared for these changes. Most studies assessing students' attitudes and knowledge were conducted before artificial intelligence became accessible and tailored to the needs of the population. Therefore, how medical students actually use AI remains largely unexplored. Objective: This study explores French medical students' knowledge and attitudes toward AI. Methods: A mixed-methods study was conducted in 2025 among French medical students in their 4th to 6th year of school, corresponding to the clerkship year. An online survey adapted from Ten et al. 2025 included open-ended questions about AI definition and feelings toward AI, a Likert scale item to assess specific attitudes, and multiple-choice questions about the characteristics of the student. Quantitative analysis was performed using non-parametric tests (Kruskal-Wallis) to compare attitudes by AI knowledge level, academic years, career aspirations, and ranking within the class. Qualitative analysis was performed inductively. Results: Of 1,377 responses received, 1,342 were included. Students had a mean age of 23.1 years and were predominantly in their 5th year. Only 6% provided a correct definition of AI, while 51% gave incorrect responses. Attitudes toward AI were generally positive, with a mean score of 6.85, with significant differences by correct response to the definition (p <0.01; Unknown: 6.12, Incorrect: 6.84, Partially correct: 6.94, Correct: 6.88) and by career goals (p<0.01; clinical: 6.58; research: 6.83; private practice: 7.19). Regarding learning, 49% of students think that AI learning should be outside the curriculum, compared to 44%. Most of the students suggested AI training through multiple workshops Qualitative analysis revealed five themes: Representation, Nuanced Optimism, Critical Consideration, Replacement, and AI Use. Students represent AI as a robot, as an improved search engine, or as an unlimited data source. Their nuanced optimism blends enthusiasm for efficient patient care and provides an opportunity to focus more on the patient relationship, with fears of dehumanization, energy costs, and skill regression. Critical consideration underscores distrust in ethical dilemmas and data security risks. Replacement concerns arise over shifting professional roles, though many believe human empathy remains irreplaceable. For AI use, students highlight administrative aid, personalized training, and clinical support. Conclusions: There is growing interest in AI among medical students, accompanied by new ecological concerns and fears of skill loss. Students seem to have learned to use AI on their own for learning. These results highlight the need to adapt training programs to include the responsible use of these technologies and how to use AI to its fullest potential.

  • Gender Differences in Medical Students’ Self-Assessment: A Linear Mixed-Effects Model

    Date Submitted: Jan 12, 2026
    Open Peer Review Period: Jan 12, 2026 - Mar 9, 2026

    Background: Self-assessment is a key requirement for lifelong learning in medicine. Evidence from gender-related research indicates that important moderators affecting self-assessment are influenced by gender. Therefore, systematic gender differences in the accuracy of self-assessment may be assumed. Objective: The present study aims to examine gender differences in medical students’ self-assessment. Specifically, this study addresses two research questions: (1) Are there systematic gender differences in medical students' self-assessment accuracy? (2) What is the magnitude of these gender differences when accounting for academic progress and knowledge? Methods: Medical students from 3 cohorts at the Medical School OWL were surveyed in 3 waves between April 2023 and April 2024 during the Progress Test Medicine (PTM). Prior to answering the test, students were asked to indicate the percentage of the PTM questions they expected to answer correctly in five knowledge areas. Self-assessment accuracy was calculated as the difference between the subjective self-assessment and the objective test score. Linear mixed models (LMMs) were used to analyze the influence of gender on students’ self-assessment accuracy while accounting for academic progress and knowledge. Results: A total of 165 students participated in this study (66.58% women, 33.42% men; age: M=21.96 years, SD=3.61). Across all models, female students rated themselves significantly less accurately than their male peers. The observed gender effect ranged from -3.74 to -6.08 percentage points. Conclusions: The results indicated systematic gender differences in medical students’ self-assessment, in favor of male students, with a magnitude comparable to the average knowledge acquired in an entire semester of study. In view of the potentially negative consequences of inaccurate self-assessment, targeted support for developing realistic self-assessment during medical studies may be particularly beneficial for female students.

  • Background: Medical graduate education increasingly uses blended and online delivery, although students' academic self-regulation may be shaped by different motivational and cognitive processes across learning contexts, with emotional factors potentially playing a complementary role. Understanding how these mechanisms operate and whether their structural relationships differ between online/blended and face-to-face formats can inform targeted educational supports. Objective: The present investigation developed and tested a comparative causal model of academic self-regulation among medical graduate students in online/blended versus face-to-face programs. We examined how key motivational constructs (eg, academic self-efficacy, task value, future orientation, perfectionism, and academic help-seeking), positive achievement emotions, and cognitive factors (cognitive academic engagement and need for closure) relate to academic self-regulation, and whether these relationships differ by learning context. Methods: The design was cross-sectional, comparative causal modeling. Participants were master’s-level students at Shahid Beheshti University of Medical Sciences enrolled in either face-to-face (population n=1554; sample n=310) or blended/online (population n=449; sample n=205) programs selected using cluster sampling. Data were collected using validated instruments measuring academic self-regulation (Bouffard scale), academic self-efficacy (Midgley et al), academic engagement (Schaufeli & Bakker), multidimensional perfectionism (Frost), academic help-seeking (Ryan & Pintrich), task value (Pintrich), future orientation (Seginer), need for closure (DeBacker & Crowson), and achievement emotions (AEQ; Pekrun et al). Data were analyzed using path analysis/structural equation modeling. Model fit was evaluated using χ²/df, CFI, GFI, AGFI, and RMSEA. Direct, indirect, and total effects were estimated for each group, and comparative interpretation focused on effect patterns and explained variance. Results: The hypothesized causal model reached an acceptable fit in both face-to-face and blended/online groups (χ²/df approximately <3; CFI/GFI/AGFI in the acceptable range; RMSEA approximately 0.02–0.05). In both groups, most of the specified direct effects reached statistical significance, while the indirect effects of exogenous variables on academic self-regulation through intermediate constructs were supported overall. Cognitive academic engagement and academic self-efficacy were important proximal predictors of academic self-regulation. The need for closure had a negative direct effect with regard to academic self-regulation. However, a previously specified direct effect from need for closure to self-regulated learning strategies could not be retained in the final revised model. In both cohorts, the indirect pathway from positive achievement emotions to academic self-regulation via cognitive engagement was not supported, indicating that positive emotions alone were insufficient to increase self-regulation through cognitive engagement. The model explained a substantial proportion of variance in academic self-regulation in both groups—being approximately 0.44 in face-to-face and 0.46 in blended/online students—indicating comparable overall explanatory power across learning contexts. Conclusions: A comparative causal model integrating motivational, emotional, and cognitive pathways provided an adequate explanation of academic self-regulation among medical graduate students in both face-to-face and blended/online formats. Findings highlight the central role of cognitive engagement and academic self-efficacy as proximal levers for supporting self-regulation across contexts. The lack of a supported indirect effect from positive emotions to self-regulation via cognitive engagement suggests that emotional experiences may not be enough unless they are accompanied by cognitively engaged learning behaviors. Considering motivational and cognitive mechanisms that together shape self-regulation within different delivery modes, educational interventions in medical graduate programs should focus on strengthening self-efficacy beliefs and cognitively engaged learning practices.