CoChat
Sign Up
Cited by 1Artificial Intelligence in Healthcare and EducationAnatomy and Medical TechnologyIntelligent Tutoring Systems and Adaptive Learning

Effectiveness of a generative AI-powered digital tutor integrated with a knowledge graph in anatomy education for nursing students: a randomized controlled trial

Open access
Zhao, Can, Zhu, Jianzhong, Liu, Jianhui, Zhao, Wentao, Pang, Yin·2026·BMC Medical Education
DOI 10.1186/s12909-026-09469-0PMC PMC13295220

Published version of a preprint

Open this paper in CoChat — free account

Abstract

BACKGROUND: Human Anatomy and Histology & Embryology are foundational courses in nursing education but are often challenging for students due to their complex spatial structures, dense knowledge points, and high cognitive load. Recent advances in generative artificial intelligence (AI) and knowledge graph technologies provide new opportunities for enhancing medical education. This study aimed to evaluate the effectiveness of a blended teaching model integrating a generative AI-powered digital tutor with a knowledge graph in improving learning outcomes among nursing students. METHODS: A three-arm parallel randomized controlled trial was conducted among 362 first-year nursing students at Cangzhou Medical College, of whom 301 completed the study. Participants were randomly assigned to an AI-Enhanced Group, a Blended Teaching Group, or a Traditional Teaching Group. All groups followed the same 16-week course structure based on Gagné's Nine Events of Instruction, differing only in the teaching tools and tutoring strategies used. Learning outcomes were evaluated using module examination scores, final comprehensive scores, knowledge graph comprehension scores, knowledge retention rates, case-based inference accuracy, and multidimensional questionnaires. Statistical analyses were conducted using one-way ANOVA with post-hoc comparisons. RESULTS: Significant differences were observed among the three groups across multiple learning outcomes. The AI-Enhanced Group achieved higher module scores (82.5 ± 7.3) compared with the Blended Group (77.6 ± 8.1) and Traditional Teaching Group (74.1 ± 8.5) (F = 28.64, p < 0.001). Final comprehensive scores were also significantly higher in the AI-Enhanced Group (84.2 ± 6.8) (F = 32.16, p < 0.001). Knowledge graph comprehension scores showed a large effect size (F = 103.72, p < 0.001). Furthermore, the AI-Enhanced Group demonstrated higher knowledge retention rates (82.3%) and case-based inference accuracy (91.2%) than the comparison groups. Questionnaire results indicated stronger recognition of digital tutors, higher acceptance of blended learning, and improved autonomous learning ability. CONCLUSIONS: The blended teaching model integrating a generative AI-powered digital tutor with a knowledge graph significantly improved nursing students' academic performance, knowledge retention, and clinical reasoning ability. This AI-enhanced instructional model may provide an effective approach for reforming foundational medical education and promoting structured knowledge learning.

AI snapshot

Written from the paper’s full text.

TL;DR

This three-arm randomized controlled trial among first-year nursing students found that a generative AI-powered digital tutor integrated with a knowledge graph improved academic performance, knowledge graph comprehension, knowledge retention, and case-based inference compared with blended and traditional teaching. The model may offer an effective approach for reforming foundational anatomy education, though the authors note single-center and single-semester limitations.

Key findings

  • The AI-Enhanced Group achieved higher module scores (82.5 ± 7.3) than the Blended Group (77.6 ± 8.1) and Traditional Teaching Group (74.1 ± 8.5) (F = 28.64, p < 0.001, η² = 0.16), and higher final comprehensive scores (84.2 ± 6.8 vs 78.9 ± 7.5 vs 75.2 ± 7.9; F = 32.16, p < 0.001, η² = 0.18).
  • Knowledge graph comprehension was highest in the AI-Enhanced Group (4.54 ± 0.50 vs 3.81 ± 0.70 vs 2.96 ± 0.88; F = 103.72, p < 0.001, η² = 0.41), as were knowledge retention rates (82.3% ± 9.5% vs 71.5% ± 10.2% vs 63.8% ± 11.4%; F = 48.52, p < 0.001, η² = 0.25) and case-based inference accuracy (91.2% ± 7.8% vs 78.5% ± 9.3% vs 67.3% ± 10.5%; F = 67.14, p < 0.001, η² = 0.31).
  • Questionnaire outcomes favored the AI-Enhanced Group for digital tutor recognition (4.90 ± 0.56 vs 4.44 ± 0.62 vs 4.45 ± 0.62; F = 15.32, p < 0.001) and autonomous learning habits (3.34 ± 0.76 vs 2.98 ± 0.84 vs 3.00 ± 0.86; F = 7.23, p = 0.001); 89.0% of the AI-Enhanced Group rated autonomous learning ability as “relatively consistent” or above, versus 72.0% and 68.3% (χ² = 23.56, p < 0.001). No significant difference was found for knowledge organization ability (F = 2.86, p = 0.06).
  • Blended versus Traditional differences were not statistically significant for module scores (p = 0.09) or final comprehensive scores (p = 0.07), but were significant for knowledge graph comprehension, retention, and case-based inference accuracy.

Methods

  • Three-arm parallel randomized controlled trial among first-year nursing students in the 2025 cohort at Cangzhou Medical College, China; 362 initially enrolled and 301 included in final analysis, allocated to AI-Enhanced (n = 100), Blended Teaching (n = 100), or Traditional Teaching (n = 101) groups. Randomization used a computer-generated sequence, with allocation performed at the class level.
  • All groups followed the same 16-week course structure based on Gagné’s Nine Events of Instruction. The AI-Enhanced Group used a generative AI-powered digital tutor integrated with a knowledge graph; the Blended Group used the same knowledge graph resources with instructor-led tutoring; the Traditional Group used lecture-based instruction with textbooks, slides, and physical models.
  • Outcomes included module examination scores, final comprehensive scores, knowledge graph comprehension, knowledge retention rate, case-based inference accuracy, and a 34-item Blended Teaching Effectiveness Questionnaire. Analyses used one-way ANOVA with Bonferroni post-hoc tests, eta squared effect sizes, and chi-square tests in SPSS 26.0.

Limitations

  • The authors state the study was conducted at a single medical college, limiting generalizability, and the intervention lasted a single semester with no long-term outcomes beyond the one-month follow-up test.
  • The authors state they did not independently validate the technical development of the AI tutor, systematically audit the factual accuracy of AI-generated responses, or separately examine how differences in students’ prompting behaviors may have influenced learning outcomes; they also note that excessive reliance on AI may reduce independent problem-solving abilities.
  • Evident from the text: group allocation was performed at the class level rather than individually, and the questionnaire was developed specifically for this study and relies on self-report, which may introduce clustering and response biases; blinding procedures are not described.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s12909-026-09469-0

Cites

42 of 42 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

31 of 58 cited sources checked · 3 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    These subjects are widely recognized as cognitively demanding for students because they involve complex spatial structures, intricate tissue morphology, and dynamic developmental processes.

    “However, these subjects are widely recognized as cognitively demanding for students because they involve complex spatial structures, intricate tissue morphology, and dynamic developmental processes [2, 3].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [2] Barbagallo MS, Porter JE, Abdelkader A, James A. A creative approach for undergraduate nursing students to learn anatomy and physiology: a qualitative exploratory study. Int J Nurs Educ Scholarsh. 2022;19:20220011. source
    • cites [3] Joseph MA, Roach EJ, Natarajan J, Karkada S, Cayaban ARR. Flipped classroom improves Omani nursing students performance and satisfaction in anatomy and physiology. BMC Nurs. 2021;20:1. source

      The Background says nursing students struggle with anatomy vocabulary, complex physiological concepts, and applying anatomy and physiology knowledge, supporting the general point that these subjects are demanding. It does not attribute that difficulty to complex spatial structures, intricate tissue morphology, or dynamic developmental processes. The Methods describe anatomical models and list histology as course content, but do not make those causal claims. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • Premisecitation

    Many studies on AI in education rely on observational designs or short-term interventions, lacking rigorous experimental evidence such as randomized controlled trials.

    “First, many studies on AI in education rely on observational designs or short-term interventions, lacking rigorous experimental evidence such as randomized controlled trials [18].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [18] Liu K, Zhang W, Li W, et al. Effectiveness of virtual reality in nursing education: a systematic review and meta-analysis. BMC Med Educ. 2023;23:710. source

      The Methods (Inclusion and exclusion criteria) concern virtual-reality instruction for nursing students and specify that the review included only randomized controlled trials. Searches for AI in education, observational designs, and short-term interventions found no passage assessing those claims. The paper therefore does not establish the cited statement about AI studies. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
  • Premisecitation

    AI-supported learning environments may promote self-directed learning behaviors.

    “Finally, AI-supported learning environments may promote self-directed learning behaviors [35, 37].”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [35] Ricotta DN, Richards JB, Atkins KM, et al. Self-directed learning in medical education: training for a lifetime of discovery. Teach Learn Med. 2022;34:530–40. source
    • cites [37] Sáiz-Manzanares MC, Escolar-Llamazares MC, Arnaiz González Á. Effectiveness of blended learning in nursing education. Int J Environ Res Public Health. 2020;17:1589. source

      The Introduction (§1) says LMS environments can facilitate self-regulated learning, and the Discussion (§4) says a blended-learning design with feedback furthers students’ self-regulated learning strategies. But the paper describes AI techniques chiefly as tools for analyzing and predicting learning behaviors; it does not establish that AI support itself promotes those behaviors. The claim also uses “self-directed” where the paper discusses “self-regulated” learning. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
  • Premisecitation

    The AI tutoring system enables personalized learning pathways by analyzing students' learning behavior and performance data.

    “Second, the AI tutoring system enables personalized learning pathways by analyzing students’ learning behavior and performance data [29, 35].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [29] Holderried F, Stegemann-Philipps C, Herrmann-Werner A, et al. A language model-powered simulated patient with automated feedback for history taking: prospective study. JMIR Med Educ. 2024;10:e59213. source

      The Methods (Study Outline) say the system analyzed students’ history-taking chats and gave feedback on the topics they covered; the Discussion (Conclusions) describes that feedback as tailored. The paper does not say the system analyzed broader learning behavior and performance data to create personalized learning pathways. Page numbers were not available in our copy. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • cites [35] Ricotta DN, Richards JB, Atkins KM, et al. Self-directed learning in medical education: training for a lifetime of discovery. Teach Learn Med. 2022;34:530–40. source
  • Interpretationcitation

    The higher retention rate observed in the AI-Enhanced Group suggests that AI-supported learning environments may facilitate deeper cognitive processing and more durable memory formation.

    “The higher retention rate observed in the AI-Enhanced Group suggests that AI-supported learning environments may facilitate deeper cognitive processing and more durable memory formation [27].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [27] Abdellatif H, Al Mushaiqri M, Albalushi H, et al. Teaching, learning and assessing anatomy with artificial intelligence: the road to a better future. Int J Environ Res Public Health. 2022;19:14209. source

      Section 3 reports that AI-assisted learning has improved learners’ knowledge retention, and the Conclusions describe AI tools as a resource for deep and logical learning (page numbers unavailable). The paper does not report a comparison involving an “AI-Enhanced Group” or establish that deeper cognitive processing explains a higher retention rate. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.
  • May combine two claims. 27
  • Classified as probably not a substantial claim. 26

  • May combine two claims.

    Premisecitation

    These methods frequently fail to support students in constructing coherent knowledge frameworks or visualizing spatial anatomical relationships.

    “Although these methods can convey foundational information, they frequently fail to support students in constructing coherent knowledge frameworks or visualizing spatial anatomical relationships [5].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [5] Saab MM, Hegarty J, Murphy D, Landers M. Incorporating virtual reality in nurse education: a qualitative study of nursing students’ perspectives. Nurse Educ Today. 2021;105:105045. source

      The Discussion reports that nursing students found VR particularly valuable for visualising complex anatomical structures and processes. It does not establish that conventional methods frequently fail to help students visualise spatial relationships, and it does not address whether those methods support coherent knowledge frameworks. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Knowledge graph technology has been increasingly applied in education to represent complex knowledge systems through structured networks.

    “In parallel, knowledge graph technology has been increasingly applied in education to represent complex knowledge systems through structured networks [12, 13].”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [12] Murali L, Gopakumar G, Viswanathan DM, Nedungadi P. Towards electronic health record-based medical knowledge graph construction, completion, and applications: a literature study. J Biomed Inf. 2023;143:104403. source
    • cites [13] He Y, Xie Z, Li J, et al. Global trends in mHealth and medical education research: bibliometrics and knowledge graph analysis. JMIR Med Educ. 2024;10:e52461. source

      The Methods (Data Analysis and Visualization; page unavailable) describe knowledge mapping used to analyze research on mHealth and medical education. The Results (Keyword Co-Occurrence and Research Trends; page unavailable) show a network of research keywords, not knowledge graphs used in education to represent complex knowledge systems. Searches for educational knowledge graphs and structured networks found no passage addressing that claim. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    In medical education, knowledge graphs have been used to support clinical reasoning training, curriculum design, and intelligent learning resource recommendation.

    “In medical education, knowledge graphs have been used to support clinical reasoning training, curriculum design, and intelligent learning resource recommendation [15].”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [15] Xu X, Chen Y, Miao J. Opportunities, challenges, and future directions of large language models, including ChatGPT in medical education: a systematic scoping review. J Educ Eval Health Prof. 2024;21:6. source

      The Results (Synthesis of results) discuss ChatGPT as a tool for personalized learning and curriculum development, while the Discussion addresses its potential effects on clinical reasoning. A search of our copy for “knowledge graph” and the claimed applications found no passage linking knowledge graphs to clinical reasoning training, curriculum design, or learning-resource recommendation. The cited paper’s discussion of ChatGPT does not establish uses of knowledge graphs. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Knowledge graphs can help learners move beyond isolated memorization and visualize how different concepts are connected, thereby facilitating knowledge organization, cognitive mapping, and conceptual integration.

    “By making such relationships explicit, knowledge graphs can help learners move beyond isolated memorization and visualize how different concepts are connected, thereby facilitating knowledge organization, cognitive mapping, and conceptual integration [14].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [14] Deng H, Jiang Y, Han Q, et al. The impact of COVID-19 on online medical education: a knowledge graph analysis based on co-term analysis. BMC Med Educ. 2023;23:203. source

      The Method (Research overview; page unavailable) describes visualizing relationships among co-occurring subject terms, which supports the claim that a graph can make connections visible. It uses those graphs to analyze research literature, however; it does not say that learners use them to move beyond memorization or improve knowledge organization, cognitive mapping, or conceptual integration. (LLM, conf=0.79)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Previous studies have shown that AI-assisted learning systems can enhance students' engagement, promote self-directed learning, and improve academic performance in various disciplines.

    “Previous studies have shown that AI-assisted learning systems can enhance students’ engagement, promote self-directed learning, and improve academic performance in various disciplines [10, 11].”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [10] Lee H. The rise of ChatGPT: exploring its potential in medical education. Anat Sci Educ. 2024;17:926–31. source

      The Abstract (page unavailable) says ChatGPT has the potential to increase student engagement, but explicitly says research is needed to confirm this. The cited paper does not establish that previous studies showed improved engagement, self-directed learning, or academic performance across various disciplines. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [11] Gilson A, Safranek CW, Huang T, et al. How does ChatGPT perform on the United States medical licensing examination (USMLE)? The implications of large language models for medical education and knowledge assessment. JMIR Med Educ. 2023;9:e45312. source

      The Discussion (Principal Findings; page unavailable) describes ChatGPT as a potential aid for students studying independently, and the Conclusions envision an interactive learning environment. The paper does not show that AI-assisted systems increased student engagement or academic performance, nor does it establish those outcomes across disciplines; it calls for further study of educational efficacy. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    AI-powered educational tools can support adaptive learning, intelligent tutoring, automated feedback, and personalized learning pathways.

    “AI-powered educational tools can support adaptive learning, intelligent tutoring, automated feedback, and personalized learning pathways [8, 9].”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [8] Fazlollahi AM, Bakhaidar M, Alsayegh A, et al. Effect of artificial intelligence tutoring vs expert instruction on learning simulated surgical skills among medical students: a randomized clinical trial. JAMA Netw Open. 2022;5:e2149008. source
    • cites [9] Wang M, Sun Z, Jia M, et al. Intelligent virtual case learning system based on real medical records and natural language processing. BMC Med Inf Decis Mak. 2022;22:60. source

      The Results section describes AIteach responding to students’ questions, checking diagnoses, scoring work, and marking missing diagnostic factors; the Conclusion says it supports learning without a clinical tutor. These passages support interactive tutoring and automated feedback, but they do not establish adaptive learning or personalized learning pathways. (LLM, conf=0.85)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    AI digital tutors provide real-time feedback and adaptive learning support, which are difficult to achieve in traditional classroom settings.

    “First, AI digital tutors provide real-time feedback and adaptive learning support, which are difficult to achieve in traditional classroom settings [28, 33].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [28] Badyal DK, Bala S, Singh T, Gulrez G. Impact of immediate feedback on the learning of medical students in pharmacology. J Adv Med Educ Prof. 2019;7:1–6. source

      The Methods (Intervention; page unavailable) describe computer-based modules that give immediate feedback when a student selects an answer, and the Introduction says feedback in traditional written assessment is often late or absent. The paper does not study AI digital tutors or establish that they provide adaptive learning support; its comparison concerns written assessment rather than classroom settings generally. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [33] Julian D, Smith R. Developing an intelligent tutoring system for robotic-assisted surgery instruction. Int J Med Robot. 2019;15:e2037. source

  • Classified as probably not a substantial claim.

    Premisecitation

    Knowledge graphs provide a structured visual representation of relationships among concepts, enabling students to organize knowledge more effectively.

    “Knowledge graphs provide a structured visual representation of relationships among concepts, enabling students to organize knowledge more effectively [12, 14].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [12] Murali L, Gopakumar G, Viswanathan DM, Nedungadi P. Towards electronic health record-based medical knowledge graph construction, completion, and applications: a literature study. J Biomed Inf. 2023;143:104403. source
    • cites [14] Deng H, Jiang Y, Han Q, et al. The impact of COVID-19 on online medical education: a knowledge graph analysis based on co-term analysis. BMC Med Educ. 2023;23:203. source

      The Method (Research overview; page unavailable) describes visual maps of relationships among co-occurring terms. The paper uses these maps to analyze research literature, however, and does not say that they help students organize their knowledge more effectively. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationcitation

    The results highlight the importance of combining AI technologies with structured knowledge representation tools, such as knowledge graphs.

    “Second, the results highlight the importance of combining AI technologies with structured knowledge representation tools, such as knowledge graphs [12, 13, 14].”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [12] Murali L, Gopakumar G, Viswanathan DM, Nedungadi P. Towards electronic health record-based medical knowledge graph construction, completion, and applications: a literature study. J Biomed Inf. 2023;143:104403. source
    • cites [13] He Y, Xie Z, Li J, et al. Global trends in mHealth and medical education research: bibliometrics and knowledge graph analysis. JMIR Med Educ. 2024;10:e52461. source

      The Methods section describes knowledge mapping with VOSviewer, and the Results section mentions AI as an example of a digital-health technology. Neither passage addresses combining AI with knowledge graphs or says the results highlight the importance of doing so. Searched for knowledge graphs and AI integration; page numbers were not available in our copy. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [14] Deng H, Jiang Y, Han Q, et al. The impact of COVID-19 on online medical education: a knowledge graph analysis based on co-term analysis. BMC Med Educ. 2023;23:203. source

      The Method (Research overview; page unavailable) describes using CiteSpace and VOSviewer to visualize bibliometric networks. Searches for AI, artificial intelligence, and integration with knowledge graphs found no passage in our copy saying that combining them is important. (LLM, conf=0.73)

      Experimental — accuracy not yet measured
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    AI-assisted educational systems can improve student engagement, provide adaptive learning support, and promote deeper understanding of complex concepts.

    “Previous research has also indicated that AI-assisted educational systems can improve student engagement, provide adaptive learning support, and promote deeper understanding of complex concepts [8, 9, 11].”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites [8] Fazlollahi AM, Bakhaidar M, Alsayegh A, et al. Effect of artificial intelligence tutoring vs expert instruction on learning simulated surgical skills among medical students: a randomized clinical trial. JAMA Netw Open. 2022;5:e2149008. source

      The Methods (Interventions; page not provided) describe feedback tailored to learners’ performance, and the Discussion reports improved simulated surgical performance with AI tutoring. The Results (Emotions and Cognitive Load; page not provided) report more positive emotions during training, but do not establish improved engagement relative to other instruction. The paper does not assess deeper understanding of complex concepts. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [9] Wang M, Sun Z, Jia M, et al. Intelligent virtual case learning system based on real medical records and natural language processing. BMC Med Inf Decis Mak. 2022;22:60. source

      The Methods section ('Cultivating students’ clinical thinking with various difficulties'; page unavailable) describes difficulty levels students can choose to suit their knowledge and says the cases can improve understanding. The Results ('Short-term learning test results'; page unavailable) report improved clinical-thinking scores. These support tailored learning and improved understanding in medical training, but the paper does not establish improved student engagement or the broader claim about complex concepts across AI-assisted educational systems. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [11] Gilson A, Safranek CW, Huang T, et al. How does ChatGPT perform on the United States medical licensing examination (USMLE)? The implications of large language models for medical education and knowledge assessment. JMIR Med Educ. 2023;9:e45312. source

      The Discussion (Principal Findings; page unavailable) describes ChatGPT giving personalized responses that may help students structure their knowledge and explore the reasoning behind difficult questions. The Conclusions describe an interactive learning environment as a potential use, not an established improvement. The paper does not report improved student engagement or demonstrate that AI-assisted systems improve learning outcomes. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    Traditional teaching methods often present these concepts in fragmented ways, which may hinder students from constructing coherent cognitive frameworks.

    “Traditional teaching methods often present these concepts in fragmented ways, which may hinder students from constructing coherent cognitive frameworks [4, 5].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [4] Nuuyoma V, Mukerenge NF, Kuugongelwa S. Fostering understanding of anatomy and physiology among undergraduate nursing students for integration into clinical practice: an integrative literature review. Nurse Educ Pract. 2025;88:104575. source
    • cites [5] Saab MM, Hegarty J, Murphy D, Landers M. Incorporating virtual reality in nurse education: a qualitative study of nursing students’ perspectives. Nurse Educ Today. 2021;105:105045. source

      The Discussion (page unavailable) says VR could help consolidate previously acquired information and supplement teaching of complex anatomy and physiology. Neither that section nor the passages on traditional learning strategies say that traditional teaching presents concepts in fragmented ways or hinders students from forming coherent cognitive frameworks. Searches for those ideas found no supporting passage in our copy. (LLM, conf=0.70)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    Anatomy and histology & embryology involve highly structured knowledge systems with complex hierarchical relationships among concepts.

    “Anatomy and histology & embryology involve highly structured knowledge systems with complex hierarchical relationships among concepts [2, 3].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [2] Barbagallo MS, Porter JE, Abdelkader A, James A. A creative approach for undergraduate nursing students to learn anatomy and physiology: a qualitative exploratory study. Int J Nurs Educ Scholarsh. 2022;19:20220011. source
    • cites [3] Joseph MA, Roach EJ, Natarajan J, Karkada S, Cayaban ARR. Flipped classroom improves Omani nursing students performance and satisfaction in anatomy and physiology. BMC Nurs. 2021;20:1. source

      The Background says nursing students encounter complex physiological concepts and a large volume of anatomy vocabulary. The Methods list histology as course content, but the paper does not describe anatomy, histology, or embryology as structured knowledge systems with hierarchical relationships among concepts. (LLM, conf=0.79)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    Excessive reliance on AI technologies may potentially reduce students' independent problem-solving abilities.

    “Third, although AI tools provide valuable learning support, excessive reliance on AI technologies may potentially reduce students’ independent problem-solving abilities [41, 42].”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [41] Khadjooi K, Rostami K, Ishaq S. How to use Gagne’s model of instructional design in teaching psychomotor skills. Gastroenterol Hepatol Bed Bench. 2011;4:116–9.

      The sections on assessing performance and enhancing retention discuss independent practice of a clinical procedure, but not AI tools or their effects on problem-solving. Searches for excessive reliance on AI and for AI-related dependence and problem-solving found no passage addressing the attributed claim. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [42] Wang J, Li W, Dun A, et al. 3D visualization technology for learning human anatomy among medical students and residents: a meta- and regression analysis. BMC Med Educ. 2024;24:461. source

      The Discussion describes benefits and limitations of 3D visualization in anatomy education, but does not address excessive reliance on AI or students’ independent problem-solving abilities. Searches for overreliance, AI, and independent problem-solving found no passage addressing that claim. Page numbers were not available in our copy. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    The AI-Enhanced Group achieved the highest final comprehensive scores (84.2 ± 6.8), followed by the Blended Group (78.9 ± 7.5) and the Traditional Group (75.2 ± 7.9).

    “Students in the AI-Enhanced Group achieved the highest final scores (84.2 ± 6.8), followed by the Blended Teaching Group (78.9 ± 7.5) and the Traditional Teaching Group (75.2 ± 7.9).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The AI-Enhanced Group achieved the highest knowledge graph comprehension score (4.54 ± 0.50), followed by the Blended Group (3.81 ± 0.70) and the Traditional Group (2.96 ± 0.88).

    “The AI-Enhanced Group achieved the highest comprehension score (4.54 ± 0.50), followed by the Blended Teaching Group (3.81 ± 0.70) and the Traditional Teaching Group (2.96 ± 0.88).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The AI-Enhanced Group achieved the highest module scores (82.5 ± 7.3), followed by the Blended Group (77.6 ± 8.1) and the Traditional Group (74.1 ± 8.5).

    “The AI-Enhanced Group achieved the highest module scores (82.5 ± 7.3), followed by the Blended Teaching Group (77.6 ± 8.1) and the Traditional Teaching Group (74.1 ± 8.5).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The AI-Enhanced Group had the highest knowledge retention rate (82.3% ± 9.5%), followed by the Blended Group (71.5% ± 10.2%) and the Traditional Teaching Group (63.8% ± 11.4%).

    “The AI-Enhanced Group demonstrated the highest retention rate (82.3% ± 9.5%), followed by the Blended Teaching Group (71.5% ± 10.2%) and the Traditional Teaching Group (63.8% ± 11.4%).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The AI-Enhanced Group achieved the highest case-based inference accuracy (91.2% ± 7.8%), followed by the Blended Group (78.5% ± 9.3%) and the Traditional Teaching Group (67.3% ± 10.5%).

    “The AI-Enhanced Group achieved the highest accuracy (91.2% ± 7.8%), followed by the Blended Teaching Group (78.5% ± 9.3%) and the Traditional Teaching Group (67.3% ± 10.5%).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    Of 362 initially enrolled students, 301 completed and were included in the final analysis.

    “A total of 362 first-year nursing students were initially enrolled in the study. During the study period, students who withdrew from the course, transferred to other programs, or failed to complete all assessments were excluded. Finally, 301 participants were included in the final analysis.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The three groups did not differ significantly at baseline in gender, age, pre-test scores, or self-assessment of learning foundation.

    “No statistically significant differences were observed among the three groups in terms of gender distribution (χ² = 0.126, p = 0.94), age (F = 0.582, p = 0.56), pre-test scores (F = 0.143, p = 0.87), or self-assessment of learning foundation (F = 0.472, p = 0.62). These results indicate that the three groups were comparable at baseline.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Randomization was conducted using a computer-generated randomization sequence.

    “Randomization was conducted using a computer-generated randomization sequence.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    The study was a three-arm parallel randomized controlled trial.

    “This study employed a three-arm parallel randomized controlled trial to evaluate the effectiveness of an AI-enhanced teaching model integrating a generative AI–powered digital tutor with a knowledge graph in anatomy and histology & embryology education for nursing students.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    The study received ethics approval and all participants provided informed consent.

    “The study protocol was reviewed and approved by the Ethics Committee of Cangzhou Medical College. All participants provided informed consent before participation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The Blended Teaching Effectiveness Questionnaire subscale for learning habits had good internal consistency (Cronbach's α = 0.845).

    “Learning habits (12 items) Evaluating behavioral patterns related to previewing, reviewing, and learning tool usage (Cronbach’s α = 0.845).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Participants were randomly assigned to three groups with final sample sizes of 100, 100, and 101.

    “The participants were randomly assigned to three groups: the AI-Enhanced Group (n = 100), the Blended Teaching Group (n = 100), and the Traditional Teaching Group (n = 101).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductreasoning

    Group allocation was performed at the class level using the generated random sequence to minimize contamination.

    “To minimize potential contamination between students within the same classroom environment, group allocation was performed at the class level. Each class was assigned to one of the three teaching models according to the generated random sequence.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisereasoning

    Many students experience difficulties in knowledge integration, long-term retention, and the transfer of theoretical knowledge to clinical problem-solving contexts.

    “As a result, many students experience difficulties in knowledge integration, long-term retention, and the transfer of theoretical knowledge to clinical problem-solving contexts.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Human Anatomy and Histology & Embryology are fundamental courses in nursing education, providing essential knowledge for understanding human structure, physiological mechanisms, and clinical nursing practice.

    “Human Anatomy and Histology & Embryology are fundamental courses in nursing education, providing essential knowledge for understanding human structure, physiological mechanisms, and clinical nursing practice [1].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [1] Asman O, Kagan I, Itzhaki M. Nursing students’ experiences and perceptions of an anatomy laboratory session: mixed methods study. Anat Sci Educ. 2022;15:898–909. source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The Blended Teaching Effectiveness Questionnaire subscale for self-assessment of learning foundation had good internal consistency (Cronbach's α = 0.826).

    “Self-assessment of learning foundation (12 items) Evaluating students’ abilities in knowledge organization, information analysis, and independent learning (Cronbach’s α = 0.826).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The Blended Teaching Effectiveness Questionnaire subscale for student expectations had good internal consistency (Cronbach's α = 0.863).

    “Student expectations (10 items) Assessing students’ acceptance of and expectations toward digital tutors and knowledge graph–supported learning (Cronbach’s α = 0.863).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The blended teaching model integrating a generative AI-powered digital tutor with a knowledge graph significantly improved nursing students' academic performance, knowledge retention, and clinical reasoning ability.

    “The blended teaching model integrating a generative AI-powered digital tutor with a knowledge graph significantly improved nursing students’ academic performance, knowledge retention, and clinical reasoning ability.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The ten case-based questions validly measure clinical reasoning ability.
    • Rests on an assumption: The knowledge retention rate (follow-up test score / final examination score) validly measures long-term knowledge retention.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The allocation procedure was conducted by a researcher not involved in the teaching intervention or outcome evaluation.

    “The allocation procedure was conducted by a researcher who was not involved in the teaching intervention or outcome evaluation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    Existing AI-assisted learning systems often focus on providing automated question answering or content generation but do not adequately address the challenge of structuring complex knowledge systems, which is a critical requirement in anatomy education.

    “Second, existing AI-assisted learning systems often focus on providing automated question answering or content generation but do not adequately address the challenge of structuring complex knowledge systems, which is a critical requirement in anatomy education.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    This structured learning approach may explain the large effect size observed for knowledge graph comprehension in the present study.

    “This structured learning approach may explain the large effect size observed for knowledge graph comprehension in the present study.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

  • May combine two claims.

    Interpretationreasoning

    By linking individual knowledge points within a coherent conceptual network, students are better able to understand relationships among concepts and build a hierarchical knowledge framework.

    “By linking individual knowledge points within a coherent conceptual network, students are better able to understand relationships among concepts and build a hierarchical knowledge framework.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The questionnaire results indicated that students in the AI-Enhanced Group reported higher levels of autonomous learning ability, suggesting that AI tools may encourage students to actively engage in the learning process.

    “The questionnaire results in this study indicated that students in the AI-Enhanced Group reported higher levels of autonomous learning ability, suggesting that AI tools may encourage students to actively engage in the learning process.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

  • May combine two claims.

    Interpretationreasoning

    The integration of AI-generated clinical scenarios and knowledge graph navigation may help students connect theoretical knowledge with clinical reasoning processes, thereby improving their ability to analyze and solve clinical problems.

    “The integration of AI-generated clinical scenarios and knowledge graph navigation may help students connect theoretical knowledge with clinical reasoning processes, thereby improving their ability to analyze and solve clinical problems.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

  • Classified as probably not a substantial claim.

    Premisecitation

    Previous studies have suggested that knowledge graph-based learning environments can support cognitive mapping and knowledge integration.

    “Previous studies have suggested that knowledge graph–based learning environments can support cognitive mapping and knowledge integration [23, 24].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [23] West DC, Pomeroy JR, Park JK, et al. Critical thinking in graduate medical education: a role for concept mapping assessment? JAMA. 2000;284:1105–10. source
    • cites [24] Brondfield S, Seol A, Hyland K, et al. Integrating concept maps into a medical student oncology curriculum. J Cancer Educ. 2021;36:85–91. source

  • May combine two claims.

    Premisecitation

    Personalized learning increases learning efficiency and motivation, particularly in complex subjects such as anatomy.

    “Personalized learning has been shown to increase learning efficiency and motivation, particularly in complex subjects such as anatomy [35, 36].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [35] Ricotta DN, Richards JB, Atkins KM, et al. Self-directed learning in medical education: training for a lifetime of discovery. Teach Learn Med. 2022;34:530–40. source
    • cites [36] Ten Cate TJ, Kusurkar RA, Williams GC. How self-determination theory can assist our understanding of the teaching and learning processes in medical education. Med Teach. 2011;33:961–73. source

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    AI digital tutors can offer individualized learning support to large numbers of students simultaneously, thereby alleviating the limitations of traditional teacher-centered instruction.

    “AI digital tutors can offer individualized learning support to large numbers of students simultaneously, thereby alleviating the limitations of traditional teacher-centered instruction.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

  • May combine two claims.

    Study design and conductnone given

    The study did not independently validate the technical development process of the generative AI-powered digital tutor or systematically audit the factual accuracy of each AI-generated response.

    “Fourth, this study evaluated the educational application of the AI-enhanced teaching model but did not independently validate the technical development process of the generative AI-powered digital tutor or systematically audit the factual accuracy of each AI-generated response.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductnone given

    The intervention was implemented over a single semester, and long-term learning outcomes beyond the follow-up test were not evaluated.

    “Second, the intervention was implemented over a single semester, and long-term learning outcomes beyond the follow-up test were not evaluated.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationcitation

    AI-supported blended teaching models may play an important role in the future development of medical education.

    “These findings suggest that AI-supported blended teaching models may play an important role in the future development of medical education [19, 39].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [19] Al-Eraky MM. Robert Gagne’s nine events of instruction, revisited. Acad Med. 2012;87:677. source
    • cites [39] Foronda CL, Fernandez-Burgos M, Nadeau C, et al. Virtual simulation in nursing education: a systematic review spanning 1996 to 2018. Simul Healthc. 2020;15:46–54. source
    • Rests on an assumption: The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The study did not separately examine how differences in students' prompting behaviors may have influenced learning outcomes.

    “In addition, the study did not separately examine how differences in students’ prompting behaviors may have influenced learning outcomes.”
    Quote in the paperNot a substantial claim
44 of 44

  • The paper relies on this without support.

    The ten case-based questions validly measure clinical reasoning ability.

    “Ten clinical case-based questions were designed to evaluate students’ ability to apply anatomical knowledge to clinical scenarios.”
    “The percentage of correctly answered questions was calculated as the case-based inference accuracy.”

  • The paper relies on this without support.

    The knowledge retention rate (follow-up test score / final examination score) validly measures long-term knowledge retention.

    “Knowledge retention rate = (follow-up test score / final examination score) × 100%.”

  • The paper relies on this without support.

    The large volume of fragmented knowledge points and the need to integrate structure–function–clinical application relationships increase learning difficulty.

    “The large volume of fragmented knowledge points and the need to integrate “structure–function–clinical application” relationships further increase the learning difficulty.”

  • The paper relies on this without support.

    The AI-generated content provided by the digital tutor was sufficiently accurate to support learning.

    “Fourth, this study evaluated the educational application of the AI-enhanced teaching model but did not independently validate the technical development process of the generative AI-powered digital tutor or systematically audit the factual accuracy of each AI-generated response.”
4 of 4

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Zhao, C., Zhu, J., Liu, J., Zhao, W., & Pang, Y.
Affiliations
Cangzhou Medical College (CN)
Corresponding author
Pang, Yin
ORCID
Liu, Jianhui
Venue
BMC Medical Education
Publisher
Springer Science and Business Media LLC
Published
2026-05-22
Volume
26
Issue
1
ISSN
1472-6920 (ISSN-L)
DOI
10.1186/s12909-026-09469-0
PMC
PMC13295220
OpenAlex
W7162120057
Semantic Scholar
25ecc8a772d926d2480bf66bea30711e92625a75
Cited by
1
References
38
Citations by year
2026: 1
FWCI
1.93
Type
journal article
Language
en
Open access
Licence
Locations
  • BMC Medical Education (publishedVersion, cc-by-nc-nd)
  • Research Square (acceptedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • PubMed Central (submittedVersion, cc-by-nc-nd)
Topics
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Physical Sciences › Engineering › Biomedical Engineering › Anatomy and Medical Technology
  • Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
Keywords
generative AI, knowledge graphs, nursing education, anatomy education, histology and embryology, blended learning, knowledge retention, clinical reasoning, self-directed learning, AI acceptance, Gagne's nine events of instruction
MeSH terms
Generative Artificial Intelligence, Anatomy, Artificial Intelligence, Computer-Assisted Instruction, Education, Nursing, Educational Measurement, Female, Humans, Male, Students, Nursing, Young Adult

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-23.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-23.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review