Actionable companion · Self-regulation & metacognition

Learning with agency: Mapping artificial intelligence in problem-based and case-based medical education: a bibliometric analysis (2019–2026)

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paper Search paper agent CC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Higher education. Health-professions learning must be distinguished from K–12 instruction.

Original authors’ abstract, reproduced under the linked source license:

Abstract Introduction Problem-based learning (PBL) has been a cornerstone of medical education since its introduction at McMaster University in the 1960s. Since the public release of ChatGPT in November 2022, artificial intelligence (AI) tools have increasingly been applied to PBL and case-based learning (CBL) contexts, yet the research landscape at this intersection remains poorly characterized. This study aimed to map the growth trajectory, thematic structure, and collaboration networks of AI-PBL/CBL research from 2019 to 2026. Methods A comprehensive search of Scopus and Web of Science was conducted on June 2, 2026, combining AI-related terms with PBL/CBL frameworks and medical education contexts. Using a PRISMA-guided bibliometric review workflow, 1,616 records were identified; after deduplication and eligibility screening, 735 unique publications (2019–2026, original articles, reviews, conference papers, and other eligible indexed document types) were included. Bibliometric analyses employed VOSviewer for network visualization (keyword co-occurrence, co-authorship, co-citation), CiteSpace for citation burst detection, and Bibliometrix for thematic mapping, three-field plot, and factorial analysis. Results Publication output grew from 25 papers in 2022 to 254 in 2025, with 206 papers indexed by June 2, 2026. The United States (n = 70) and China (n = 64) led publication volume. Keyword co-occurrence analysis identified four thematic clusters: a central AI-focused cluster, a medical education and clinical reasoning cluster, a nursing and simulation-oriented cluster, and an educational technology cluster. Kung et al.'s 2023 study evaluating ChatGPT's performance on the USMLE was the most frequently co-cited reference (62 co-citations, betweenness centrality = 0.11). Thematic mapping positioned machine learning as a motor theme, while clinical reasoning, medical education, and self-directed learning appeared in the basic themes quadrant. The country/region collaboration network comprised 33 countries/regions and was led by the United States and China, although collaboration patterns remained uneven across regions. Conclusion To our knowledge, this is the first bibliometric study specifically focused on the intersection of AI technologies with PBL/CBL in health professions education. The findings reveal rapid growth after 2023, four distinct but interconnected research clusters, and a collaboration network led by the United States and China, with uneven regional participation. These results may inform curriculum design and research priorities in AI-enhanced medical education. Keywords: artificial intelligence - AI, bibliometric analysis, case-based learning (CBL), ChatGPT, clinical reasoning, large language models (LLM), medical education - graduate, PBL (problem based learning) model

What kind of evidence is this?

Supporting AI learning science · Bibliometric / mapping study

Stage: Higher education. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Study methods · source passage

For this descriptive bibliometric review/analysis, a comprehensive literature search was conducted on June 2, 2026, using Scopus (Elsevier) and Web of Science (WoS) Core Collection (Clarivate Analytics). Both databases were searched without language restrictions. The search strategy combined three concept clusters using the Boolean operator AND:

Limitations and uncertainty · source passage

The four thematic clusters identified through keyword co-occurrence analysis highlight distinct research emphases within the field. The nursing and simulation-oriented cluster, centered on nursing education and virtual reality, reflects the substantial attention AI has received in clinical skills training for nursing students, an emphasis that may be driven by the protocol-driven, hands-on nature of nursing curricula where AI-assisted procedural training and feedback systems show immediate applicability (24). The AI and clinical reasoning cluster examines how machine learning and natural language processing can augment aspects of the diagnostic process traditionally facilitated by PBL tutors—arguably the most complex pedagogical function that AI tools are being asked to support. The educational technology cluster, encompassing e-learning, generative AI, and active learning, represents a convergence of AI capabilities with established educational modalities. The LLM-focused cluster, anchored by ChatGPT and large language models, is the most emergent and rapidly expanding: as generative AI tools become more prevalent in clinical practice, PBL curricula are evolving to teach students about AI's capabilities, limitations, biases, and ethical dimensions alongside learning with AI tools (25). The relative separation of these clusters in the network suggests that AI literacy education and AI-assisted pedagogy are developing as parallel rather than integrated streams, a gap that curriculum designers may wish to address.

Design the system before the pilot

Use this study to inform a learning mechanism or implementation question. It is not direct evidence that autonomous agents improve learning. Keep the proposed agent behavior separate from what the source actually tested.

Use plan–attempt–check–reflect prompts and compare confidence with actual performance. Keep learner choice and reflection visible.
Use common rubrics, blind scoring where feasible and a fresh unaided reasoning task. Document permissible assistance and avoid treating AI-detector scores as proof of misconduct.

A 90-day district pilot

Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
Days 16–30 · Design. Choose one disciplinary task. Require a first attempt, a documented AI consultation, a critique of the output and an independent explanation. Align the comparison group and scoring rubric with the same objective. Name an accountable owner and set the baseline, success threshold and budget before implementation.
Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
Days 61–90 · Decide. Blind-score reasoning and accuracy on a new task without AI. Distinguish model examination scores, student perceptions and actual learner gains; only the last directly tests educational benefit. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Qingyuan Tan; Yukui Ma; Jichun Zhao; Huanrui Hu. Mapping artificial intelligence in problem-based and case-based medical education: a bibliometric analysis (2019–2026). Frontiers in Medicine (2026-07-30). https://doi.org/10.3389/fmed.2026.1901004

Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.