Evaluation of three artificial intelligence chatbots for generating clinical hematology multiple choice questions for medical students.

Aim. To assess the level of awareness of students (clinical pharmacists) of advanced training in the specialization “Clinical Pharmacy” about adverse reactions of beta-lactam antibiotics and the level of cross-reactivity between these drugs. Materials and methods. We developed a questionnaire and conducted an anonymous voluntary remote survey of 34 students specializing in “Clinical Pharmacy”. The survey instrument was a questionnaire containing multiple-choice questions to assess knowledge about the adverse reactions of beta-lactam antibiotics, which prescription was a significant component of all prescriptions of antimicrobial drugs, and the conditions for their prevention. The reliability and validity of the questionnaire questions were previously assessed by experts from the group who did not participate in the survey. Results. The survey included participants working as clinical pharmacists in clinical healthcare institutions. It was determined that 62 % of respondents had up to 3 years of work experience in the specified position, 32 % had more than 10 years of work experience, and 6 % of respondents had 3 to 5 years of experience as a clinical pharmacist. It was found that clinical pharmacists were well aware of the groups of beta-lactam antibiotics, and most of them were well aware of the adverse reactions of these drugs. Most respondents identified the significance of allergic reactions and the presence of a high risk of cross-reactivity between groups of beta-lactam antibiotics, which was not consistent with current scientific data. Only 53 % of respondents correctly knew that the main factor in allergic reactions to cephalosporins was the side chains of the chemical structure of the drugs, and 59 % of respondents correctly named cephalosporins with side chains similar to penicillins, and therefore, there might be significant cross-reactivity between these agents. The majority of 85 % of respondents mistakenly believed that the frequency of cross-reactivity between penicillins and carbapenems was significant, and at the same time, the majority of 73 % participants were correctly aware of the low risk of cross-reactivity between penicillins and monobactams, while 50 % of clinical pharmacists correctly indicated ceftazidime as a drug that had cross-reactivity with aztreonam. Conclusions. The results obtained show that clinical pharmacists are well aware of the adverse reactions of beta-lactam antibiotics and the lack of accurate knowledge about the factors and the risk of cross-reactivity between these drugs, which indicates the feasibility of constant updating and improvement of knowledge already at the stage of postgraduate education.

Background: Multimodal large language models (MLLMs) integrating multiple AI systems and unimodal large language models (LLMs) represent distinct approaches to clinical decision support. Their comparative performance against human clinical experts in complex cardiovascular emergencies remains inadequately characterized. Objective: To compare the performance of a combined MLLM system (GPT-4V + Med-PaLM 2 + BioGPT), a unimodal LLM (ChatGPT-5.2), and human physicians from multiple centers (radiologists, emergency medicine specialists, cardiovascular surgeons) on aortic dissection clinical questions across diagnosis, treatment, and complication management domains. Methods: This multicenter cross-sectional study was conducted across five tertiary care centers in Turkey (Elazığ, Ankara, Antalya). A total of 25 validated multiple-choice questions were categorized into three domains: diagnosis (n = 8), treatment (n = 9), and complication management (n = 8). Questions were administered to the MLLM, ChatGPT-5.2 (Unimodal), and nine physicians from five centers: radiologists (n = 3), emergency medicine specialists (n = 3), and cardiovascular surgeons (n = 3). Statistical comparisons utilized chi-square tests. Results: Overall accuracy was 92.0% for the MLLM and 96.0% for ChatGPT-5.2 (Unimodal). Among human physicians, cardiovascular surgeons achieved 96.0%, radiologists 92.0%, and emergency medicine specialists 89.3%. The MLLM excelled in diagnosis (100%) but showed lower performance in treatment (88.9%) and complication management (87.5%). No significant differences were observed between AI models and human physician groups (all p > 0.05). Conclusions: Both the MLLM and unimodal ChatGPT-5.2 demonstrated performance within the range of human clinical experts in this controlled assessment of aortic dissection scenarios, though definitive conclusions regarding equivalence require larger-scale validation. These findings support further investigation of complementary roles for different AI architectures in clinical decision support.

In recent years, advancements in artificial intelligence (AI) have led to the widespread integration of large language models and their chatbot applications into various fields, including dental education. This study aimed to evaluate the accuracy of ChatGPT-4 Omni (ChatGPT-4o) and Gemini Advanced in answering multiple-choice questions from the Turkish Dentistry Specialization Exams (DUS) across various disciplines. A total of 1,504 multiple-choice questions from 10 years of DUS exams were analyzed to compare the accuracy of ChatGPT-4o and Gemini Advanced. The questions were categorized into Fundamental Medical Sciences (n = 514) and Clinical Dental Sciences (n = 990). Each question was submitted to both chatbots, resulting in 3,008 responses. Accuracy was assessed using the official answer keys. Chi-square tests and Bonferroni post-hoc analyses were used to compare accuracy across disciplines and examine year-based variations. ChatGPT-4o achieved an overall accuracy rate of 84%, while Gemini Advanced achieved 81.8% (p = 0.110). For the Fundamental Medical Sciences questions, no statistically significant differences were observed across sub-disciplines, with overall accuracies of 92.6% for ChatGPT-4o and 93.4% for Gemini Advanced. For the Clinical Dental Sciences questions, ChatGPT-4o outperformed Gemini Advanced in Prosthetic Dentistry (p = 0.013) and Dentomaxillofacial Radiology (p = 0.001), whereas Gemini Advanced showed higher accuracy in Pediatric Dentistry (p = 0.008). Across all Clinical Dental Sciences questions, ChatGPT-4o achieved an accuracy of 79.5%, compared to 75.8% for Gemini Advanced, and this difference was statistically significant (p = 0.046). AI-based chatbots demonstrate strong potential in answering multiple-choice dentistry questions. However, variations in performance across disciplines were observed, indicating differences in accuracy depending on the subject area. These findings highlight the potential educational implications of integrating AI into dental curricula, particularly as supplementary tools for exam preparation and knowledge reinforcement. Nevertheless, cautious integration is required to ensure that AI supports, rather than replaces, critical thinking and professional expertise.

A roadmap for future clinical research in advanced prostate cancer: Findings from the Advanced Prostate Cancer Consensus Conference (APCCC) 2024.

Traditional teaching (TT) is lecturer-centred, while student-centred teaching, including Jigsaw (JS), fosters student interaction. However, the results of research on the effectiveness of JS on learning outcomes is inconsistent. This randomised controlled trial compared the effect of JS and TT on higher-education students' knowledge of the anatomy and physiology of the sensory apparatus and on their perceptions of the two teaching methods. Forty-eight undergraduate social nursing students were randomised to either a TT or JS-based learning activity. One JS student dropped out before completing the activity and was excluded, resulting in 47 participants (23 TT, 24 JS). Improvement in knowledge was assessed by recording the number of correct answers (points) on a 25-question multiple-choice test at three time points: before the activity (baseline), immediately after the activity (post-intervention) and at a three-month follow-up. Additionally, the students' perceptions were surveyed post-intervention using a questionnaire. Improvement in knowledge was analysed using a longitudinal model, and group differences in perceptions were analysed using Wilcoxon-Mann-Whitney tests. Both groups significantly improved knowledge from baseline to post-intervention (TT: 5.2 points; JS: 4.6 points; p < 0.001 both groups) and from baseline to the three-month follow-up (TT: 3.8 points; JS: 3.3 points; p < 0.001 for both groups). No statistically significant differences in knowledge were observed between the groups (post-intervention difference: -0.5 points, 95% CI: -2.3 to 1.2, p = 0.54; three-month difference: -0.6 points, 95% CI: -2.4 to 1.2, p = 0.52). Overall, 91% improved from baseline to post-intervention, and 83% from baseline to three months. A higher proportion of TT students than JS students preferred their assigned teaching activity over others (78% vs. 38%, p = 0.002) or believed it would improve their grades (87% vs. 42%, p = 0.004). On the other hand, a lower proportion thought that it provided useful practice in oral presentation / communication (61% vs. 88%, p = 0.02). This study found that both the Jigsaw method and traditional lectures effectively improve anatomy and physiology knowledge among students. Despite a student preference for traditional teaching, there were no significant differences in knowledge acquisition or retention between the two methods. Clinical trial number not applicable.

Objective: To assess the improvement in the understanding of third year medical students about medication safety after a brief cinemeducation session. Methodology: We conducted a single group pre and posttest educational intervention study at the Liaquat National Medical College in Karachi from July to September 2025. A total of 52 third-year undergraduate medical students participated in this study. The educational intervention consisted of a two-hour teaching session. The session included assigned pre readings, a short introductory lecture, several video clips demonstrating medication errors, group discussions, reflective writing based on the Gibbs reflection cycle. Multiple-choice questions were used before and after the course to assess students' knowledge. Post session feedback was obtained and reflections were thematically analyzed. Paired t-test was used to compare participants' overall knowledge scores before and after the test. P &lt; 0.05 was considered as statistically significant. We analyzed the feedback data using descriptive statistical methods. Results: A total of 52 students participated in the study and completed pre and posttests. The mean knowledge score improved from 9.6 ± 1.6 to 10.2 ± 1.7. This modest but significant improvement had a mean difference of 0.60 ± 2.1 (p = 0.025). The group of 22 students provided feedback. The majority of students said that multiple-choice questions helped them learn better (95.4%), the multimedia section was useful (90.9%), and the course helped them identify and avoid medication errors (95.5%). After analyzing the students' reflection, five main themes were identified: recognizing drug safety errors, feelings about adverse events, the personal and systemic factors at play, methods to prevent problems, and how students planned to apply what they had learned to their future work. Conclusion: Cinemeducation was effective in strengthening students’ understanding and attitudes related to medication safety. While knowledge gains were modest, the approach encouraged reflection, emotional engagement, and practical insight. This suggested the value of this tool as an engaging method for teaching patient safety and other longitudinal themes within the medical curriculum

To evaluate the current euthanasia procedures of Flemish veterinarians in companion animal veterinary practice. An online survey was distributed to veterinarians in Flanders (Belgium) through social media, veterinary websites, and email. The questionnaire included multiple-choice, Likert-scale, and open questions about the demographics, veterinarian-client communication, and euthanasia procedure. 512 veterinarians completed the survey. Participants indicated that they (almost) always explain the euthanasia procedure to the owner and discuss potential side effects of the medication used. Overall, 93.5% indicated that communication with the owner throughout the procedure is (almost) always easy, and 76.4% indicated that they can (almost) always deal easily with the owner's emotions. Medetomidine and ketamine were most often used to sedate and anesthetize the animal, followed by an IV lethal dose of pentobarbital sodium. The euthanasia procedure was mainly learned from other veterinarians after graduation. Unwanted adverse effects were generally rare, but a painful injection and vomiting were commonly reported as possible side effects. Several veterinarians reported a need for more education. A wide range of different euthanasia procedures was found, which may be linked to limited education on this subject in the veterinary curriculum. Further in-depth research into the occurrence of adverse events during pet euthanasia is warranted to ensure the well-being of all stakeholders.

This study aimed to evaluate the reliability and accuracy of four AI chatbots—ChatGPT-3.5, ChatGPT-4.0, Gemini 2.5 Flash, and DeepSeek-V3—in the field of veterinary anatomy. A total of 85 multiple-choice questions encompassing major anatomical systems were presented individually to each model under identical conditions. Responses were evaluated for accuracy, and success rates were calculated as percentages. Statistical differences among models were analyzed using the Pearson chi-square test (p

To our knowledge, no study assessed the published obesity-related nutrition educational interventions (ONEI) delivered to medical students, residents or fellows. The current scoping review undertook this task. We searched PubMed, Web of Science and EMBASEdatabases. Original articles of ONEI delivered to medical students, residents or fellows were included. The ONEIs' goals, content, teaching and learning characteristics, health professionals involved, duration of the intervention, trainees, target populations, evaluation methods, and outcomes were scrutinized. Extracted data were categorized, condensed, and mapped to identify gaps. Thirty six ONEI were included, mostly (88.8%) from North America. Physicians were the instructors in less than half the ONEI (44.4%), followed by dietitians (30.5%). Most ONEI (75%) were for medical students, 25% for residents, and none for fellows. The majority of ONEI (83.3%) were for adult patients with obesity (PWO), whereas 16.7% focused on childhood obesity. ONEIs' overarching goals were to improve nutrition knowledge, skills and confidence in assessment, counseling, intervention, and attitudes towards PWO. The teaching and learning methods included face-to-face traditional lectures, oral presentations, case-based scenarios, interactive group discussions, simulated and standardized patient interactions and role play activities. Technology-based education modalities were also utilized such asonline videos, web-based and computer-based learning. Other interventionsinvolved culinary classes to deepen trainees' understanding of nutrition and health. A range of evaluation tools wasused to appraise effectiveness including pre- and post-surveys, multiple-choice questions, observed structured clinical examinations, learning essays and narrative reflections. The ONEI resulted in significant improvements in knowledge, competencies, confidence, nutrition assessment, counseling, weight management, andpromoting healthy lifestyle. Positive improvements were also noted in communication skills, attitudes toward PWO, and in the trainees' own personal behaviors. Obesity-related nutrition education does not appear to be sufficiently prioritized across the medical education continuum. The review also identified five paradoxes: (1)a slow developing evidence base of obesity-related nutrition education in the face of a fast-paced obesity epidemic; (2)only a minority of physician instructors participate in nutrition education, despite their complementary role to dietitians and the well-recognized benefits of collaborative nutrition training; (3)a low representation of ONEI outside of North America despite the global prevalence of obesity; (4)a dearth of ONEI delivered to residents, and none identified for fellows, despite the roles of these young physicians in delivering care to PWO. In addition, (5)ONEItargeting childhood obesity remain insufficiently emphasized, despite its significant long-term health consequences. These 'mismatches' suggest the necessity for efforts to enhance the obesity-related nutrition education for young physicians-in-training.

Background Every day we are faced with different treatment claims, in the news, in social media, and by our family and friends. Some of these claims are true, but many are unsubstantiated. Without being supported by reliable evidence such guidance can lead to waste and harmful health choices. The Informed Health Choices (IHC) Network facilitates development of interventions for teaching children and adults the ability to assess treatment claims (informedhealthchoices.org). Our objective was to develop and evaluate a new assessment tool developed from the item bank for use in an upcoming trial of lower secondary school resources in Uganda, Kenya, and Rwanda. Methods A cross-sectional study evaluating a questionnaire including two item-sets was used. The first evaluated ability using multiple-choice questions (scored dichotomously) and the other evaluated intended behaviour and self-efficacy (measured using Likert scales). This study was conducted in Uganda, Kenya, and Rwanda in 2021. We recruited children (over 12 years old) and adults through schools and our networks. We entered 1,671 responses into our analysis. Summary and individual fit to the Rasch model (including Cronbach’s Alpha) were assessed using the RUMM2030 software. Results Both item-sets were found to have good fit to the Rasch model and were acceptable to our target audience. The reliability was good (Cronbach’s alpha &gt;0.7). Observations of the individual item and person fit provided us with guidance on how we could improve the design, scoring, and administration of the two item-sets. There was no local dependency in either of the item-sets, and both item-sets were found to have acceptable unidimensionality. Conclusion Overall, the two item-sets were found to have satisfactory measurement properties. Based on our analysis, we consider these instruments to be suitable for our target audiences in Uganda, Kenya and Rwanda.

Nutritional medicine is an important area of gastroenterology. It requires trained medical personnel who have qualified through additional training in nutritional medicine.In fall 2023, a survey was conducted among young gastroenterologists (JUGA) of the DGVS. Demographic data, the institutional and personal importance of nutritional medicine, and additional training were surveyed using Likert scales (rating scale 1 to 10), single-choice, and multiple-choice questions. The influence of gender, age, position, and institution was also examined.The personal importance of nutritional medicine among the 225 participants was 8 (7-10) on average, while its importance in the institution where they worked was 5 (3-7). This was rated significantly (p=0.012) higher in university hospitals than in non-university institutions. Obstacles cited include an unclear process for additional training, availability of case seminars and courses, and a lack of support at the management level.The importance of nutritional medicine at clinical institutions is lower than the personal interest in nutritional medicine and additional training, but this is currently associated with significant obstacles. Structural improvements and greater recognition are needed to keep additional training attractive and to meet the growing demand gap.

BackgroundThe rapidly changing landscape of healthcare in the era of digital transformation brings an abundance of opportunities to improve quality of patient care; however, in parallel, it poses unique challenges to healthcare professionals, which mandate constant efforts to develop both digital competencies and a crucial set of transversal skills. In this context, to address existing unmet needs, the consortium of the H-PASS (Health Professionals and the DigitAl team SkillS advancement) European project has developed a blended learning-based educational programme.ObjectivesThe primary objective of the hereby presented preliminary analysis is to evaluate the degree of improvement in participants’ knowledge on key aspects of technologies currently incorporated in healthcare, after completion of the training’s first module.MethodsThe training courses are currently being delivered in 6 EU member stated, and are divided into 5 distinct educational modules; four primary and one secondary module, focused on virtual reality (VR) technology. A detailed description of the pilot training course conducted in Greece is provided in Figure 1.Before initiation and after completion of each educational module’s activities, all participants are invited to answer the same set of module-specific multiple choice questions. For the purposes of the present analysis, we comparatively evaluated the percentages of participants that answered Module 1-specific questions correctly, prior to- and after completion of all educational activities of the course’s first module.ResultsAs of 10/6/2025, thirty nine of the 75 in total healthcare professionals that have already been enrolled in the Greek pilot training, have completed Module 1 and answered both the pre- and the post-module six-question knowledge evaluation survey.Questions included, as well as percentages of trainees having provided the correct answers to each question prior to- and after completion of the first module, are outlined in Figure 2.For all 6 questions, a significant increase of the percentage of trainees that provided the correct answer was observed (9.7%-31,6%, mean value=18.6%).ConclusionsWhile the delivery of the H-PASS pilot training courses is still ongoing, our preliminary results suggest that appropriately designed, blended learning-based educational programs could be valuable in promoting key digital skills, which are sine qua non for healthcare professionals in the era of digital transformationFigure 1.Figure 2

BackgroundFive percent of the Infectious Disease board exam focuses on Tropical Medicine and Travelers’ Health and exposure to these topics in fellowship varies greatly. This often leaves Infectious Disease practitioners unprepared to manage global infections. The University of Minnesota started an asynchronous online course in 2015 focusing on Travel and Tropical Medicine specifically for Infectious Disease specialists to address this gap. The course is accredited for >35 CME hours and includes individual sections focusing on parasites, bacteria, viruses, fungi, and traveler’s health. Modules are available online for participants that currently practice in Infectious Diseases or are training to do so.Figure 1Map including states and countries participants reported residing in. States highlighted in red had at least one participant.Figure 2Graphical representation of responses to the question: “As a result of participating in this activity, will you make any measurable changes in your role?” Yes (23, 59.0%); Possibly. . . I am contemplating change (10; 25.6%); No, but it affirms my practice (6; 15.4%); No, I am unable to implement the suggested changes (0; 0%).MethodsThis project involved the analysis of course assessments for the various sessions in the course, including pre/post knowledge, skills and attitude tools. Participant evaluations consisted of multiple choice questions, Likert scales, and short answer responses.Figure 3For those that answered “yes” to making changes in clinical practice, the short-answer examples provided were coded and thematically analyzed. The themes of “improved Travel Medicine skills” and “improved clinical Infectious Disease skills” were recurrent.Results39 participants completed the course - 27 MDs, 9 advanced practice providers, 3 other. Registrants reported residency in 17 states and 7 countries (Fig. 1). Participants ranked individual sections highly, and 59% reported that they would make changes in their current practice based on course material (Fig. 2). Individuals indicating they would make a practice change provided short answer examples that were analyzed and thematically coded. The themes of improving infectious disease clinical care and improving travel advice were most frequently observed (Fig. 3).ConclusionThe University of Minnesota’s Travel & Tropical Medicine for the Infectious Disease Specialist provides a unique and flexible supplemental education that is not currently covered by ID fellowships. Current trainees or practicing providers looking to expand their knowledge base and improve clinical care could consider enrollment.DisclosuresAll Authors: No reported disclosures

BackgroundThe University of Kansas School of Medicine adult Infectious Diseases (ID) 4th year elective focused education on cases encountered on the wards, which can lead to gaps in content2. A comprehensive ID elective is important to medical student knowledge and can be a tool for recruitment into the field. The aim of this quality improvement project was to assess the impact of a structured curriculum on student’s ID knowledge and overall satisfaction with the rotation.Average Scored Survey Results by Academic Year (AY)Student rotational survey scores for questions 1-5 for pre (23/24) versus post (24/25) interventionPre/Post Test Average ScoresStudent pre and post-test scores on ID-related knowledge multiple-choice questions for 24/25 AYMethodsThis single-center quasi-experimental study used pre/post surveys and tests to compare the impact of a revised curriculum. During the 2024-2025 academic year (AY), curriculum changes included an orientation, dedicated time to complete online modules, and coaching sessions. Satisfaction surveys and pre/post tests were administered to assess the curriculum changes on student perception of experience and ID knowledge. Faculty and fellows were evaluated for satisfaction & student preparedness.Scored Faculty Survey ResultsPre (23-24) and Post (24-25) Faculty/Fellow Satisfaction Survey ResultsResultsAll student survey questions 23-24 and 24-25 AY, responses were favorable, and there was a non-significant trend towards higher scores in the intervention group (Figure 1). On faculty/fellow surveys, there were statistically significantly higher scores on post intervention survey for student preparedness, optimal elective structure and adequate time to complete online modules (Figure 2).Pre/Post Test Results: In AY 24-25 the average pretest score was 9.0 (+/- 2.0) and average post-test score was 14.4 (+/- 2.3). Using a student t-test there was a statistically significant difference in the pre versus post-test total scores (P value < 0.001, Figure 3). There was no available comparison data from 23-24 AY to compare.ConclusionPre-post test data did show a significant increase in average student ID knowledge, but it is unclear if this is a result of the curriculum changes. However, students, faculty, and fellows did report high satisfaction with the rotation and the. intervention was easily implemented with minimal cost. We are hopeful that the curriculum changes will interest more students in the elective, increase faculty excitement in teaching, and promote future recruitment.DisclosuresAll Authors: No reported disclosures

BackgroundPrEP is vital for HIV prevention, yet few residents receive training, revealing a learning gap. We aimed to improve PrEP knowledge and prescribing comfort among internal medicine residents.Figure 1.Changes in medical knowledge and level of comfort in prescribing HIV PrEP before and after intervention Figure 2.Changes in HIV PrEP knowledge assessment and level of comfort before and after intervention by post-graduate year MethodsIn this quasi-experimental study, residents completed a 10-question multiple-choice test before and after the intervention, reported their post graduate year (PGY), their comfort prescribing PrEP (1–5 Likert scale), and described challenges in free-text responses, which were categorized into themes: sexual history-taking, PrEP indications, medication options and side effects, counseling, and cost.From 2/2025-4/2025, residents attended PrEP focused didactic sessions, including a simulated case with feedback. Supervising clinicians received a separate lecture.Ranked and standard linear models evaluated changes in test scores and comfort levels (p < 0.05), with data analyzed using R (v4.1.2).Figure 3.Change in medical knowledge and level of comfort in prescribing HIV PrEP before and after intervention by post-graduate yearTable 1.Resident reported challenges in prescribing HIV PrEP ResultsOf 46 eligible residents, 34 (74%) completed pre-training test and 30 (65%) residents completed the post-training test. Overall test scores improved by 26% (p< 0.001). Comfort level improved by 1.8 points (p< 0.001). (Figure 1). Test scores and comfort levels improved across all PGY (Figure 2).Higher test scores correlated with higher confidence among all PGY levels. (Figure 3)Reported challenges were grouped and quantified as shown in Table 1. The most reported challenges were counselling (n =12) and medication options (n = 10) despite training.ConclusionOur targeted intervention significantly improved PrEP knowledge and comfort, aligning with prior studies, though few assessed practical skills.Post-training, residents felt more confident identifying PrEP candidates, but challenges with counseling and medication selection persisted, suggesting these need continued support. Some residents with better scores still reported low comfort, pointing to non-cognitive barriers like personal discomfort.Limitations include a small sample size and no assessment of clinical outcomes like prescribing rates.This intervention effectively enhanced resident knowledge and confidence and can be integrated into residency training to support HIV prevention.DisclosuresAll Authors: No reported disclosures

A comprehensive understanding of birth mechanics is essential for safe and competent practice in obstetrics and midwifery. These dynamic, three-dimensional processes occur internally and are not directly observable, making them challenging to teach with traditional methods. Virtual Reality (VR) offers unique potential to visualize such complex, invisible mechanisms. This nationwide survey assessed the perceived need, curricular contexts, and preferred content for a high-fidelity VR application to teach birth mechanics in midwifery and medical education. An anonymous, digital survey was distributed to all study programs in midwifery science and medicine in Germany via program coordinators, department heads, and teaching leads in obstetrics and gynecology. Separate questionnaires for students and teaching professionals included 17 items on demographics, prior VR experience, attitudes toward VR, curricular timing, relevant learning objectives, and prioritized birth mechanics deviations. Responses were collected using Likert scales and multiple-choice questions, with non-parametric group comparisons (Mann-Whitney U, Chi-square test) and Bonferroni correction. A total of 1,249 complete responses were analyzed with 22.4% (n = 280) teaching professionals and 77.6% (n = 969) students; 65.7% (n = 821) medicine, 34.3% (n = 428) midwifery. Overall, 76.3% reported a high or very high need for VR, and 90.2% held a positive or rather positive attitude toward its use. Midwifery teaching professionals reported more prior VR experience (56.7%) than medical ones (12.7%). Midwifery students favored early integration, while medical students preferred later, clinically oriented phases. The most frequently prioritized deviations in birth mechanics were occiput posterior position (73.3%), breech presentation (51.4%), and direct occiput position with high station (44.0%), with notable differences between disciplines and educational roles. This study reveals a strong perceived need for a VR application uniting anatomical accuracy with the dynamic processes of childbirth. Birth mechanics deviations were considered particularly suitable for simulation when they represent significant clinical challenges or are rarely demonstrated in current teaching. Tailored, interprofessional VR tools are perceived as potentially valuable for addressing these needs and may enhance obstetric and midwifery education. As part of the V.T.O.B.S. project (Virtual training for obstetric birth simulations), the next step will be the development and evaluation of such an application. The study protocol was preregistered in the German Clinical Trials Register (DRKS00035186 27.09.2024).

Large language models (LLMs), such as ChatGPT and Grok, have rapidly advanced in natural language understanding and are increasingly being applied to specialized fields, including medicine. In this study, we evaluated the domain-specific knowledge of LLMs in radiotherapy by assessing their performance on three certification examinations in Japan: the Japanese Medical Physicist Examination, the Japanese Board Examination for Radiologists and the Japanese Board Examination for Radiation Oncologists. We assessed five LLMs-ChatGPT-5, ChatGPT-5 Pro, Grok 4, Grok 4 heavy and Gemini 2.5 Pro-by inputting all multiple-choice questions from these exams into each model and recording their responses. The AI-generated answers were compared with reference answers determined by experienced medical physicists and radiation oncologists. The results demonstrated average accuracies of 84.7 ± 2.0% (ChatGPT-5), 94.7 ± 2.1% (ChatGPT-5 Pro), 78.4 ± 1.2% (Grok 4), 81.6 ± 2.2% (Grok 4 heavy) and 88.9 ± 1.2% (Gemini 2.5 Pro). All models achieved over 75% accuracy, with ChatGPT-5 Pro consistently outperforming others, attaining an average accuracy exceeding 90% across all examinations. These findings highlight the strong potential of advanced LLMs, particularly ChatGPT-5 Pro, for future integration into radiotherapy-related applications such as automated contouring and treatment planning support.

Artificial intelligence is among the most rapidly developing branches of technology. It has proven to be a helpful tool in various fields, including medicine. Significant advances in the development of new language models prompt an evaluation of their effectiveness across various areas of medicine, including anatomy. This study aimed to assess the effectiveness of artificial intelligence in solving theoretical anatomy exams designed for medical students. The study utilized 555 multiple-choice questions (150 in Polish and 405 in English) sourced from past anatomy exams for the medical program. The models tested included: ChatGPT-4o mini, ChatGPT-4o, DeepSeek, Copilot, Gemini, and two Polish models: Bielik and PLLum. Each question was asked only once. For analysis purposes, the questions were categorized by type and by the anatomical structure they addressed. Out of 555 questions, ChatGPT-4o mini answered 394 correctly (71%), ChatGPT-4o - 461 (83.1%), DeepSeek - 427 (76.9%), Copilot - 442 (79.6%), Gemini - 439 (78.8%), Bielik - 166 (29.9%), and PLLum - 222 (40.0%). The language models performed poorest on multiple-answer questions (37.6%) and best on questions concerning the function of a given organ (75%). Most of the tested language models are capable of independently passing the exam, which should serve as a warning to teaching staff supervising students during exams and assessments. Properly formulated questions can currently hinder students relying on artificial intelligence from passing, but ongoing AI advancements may result in even higher pass rates in the future.

This study aims to analyze students’ mastery of the verb be in various sentence forms and to evaluate the role of digital-based exercises in enhancing students’ critical thinking and digital literacy skills. The study employed a quantitative descriptive approach, with data collected through Google Forms from 12 EFL students. The research instrument consisted of multiple-choice questions, with 10 items each for positive sentences, negative sentences, and interrogative sentences. The results indicate that students are more proficient in forming positive sentences, while negative and interrogative sentences show higher error rates, particularly regarding subject-verb inversion and the use of not. Activities requiring students to explain their answer choices fostered the development of critical thinking skills, while the use of an online platform strengthened their digital literacy. The limitations of this study include a relatively small sample size and a focus on simple sentences. Future research is suggested to expand the sample, use more complex sentences, and explore the relationship between grammar mastery, critical thinking, and digital literacy in more authentic learning contexts.