Agentic education · Learning science

Actionable companion · Assessment & reasoning

Learning with agency: Artificial intelligence–powered virtual standardized patients in teaching history-taking skills to medical students: a randomized controlled trial

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 paperSearch paper agentCC 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 Background Artificial intelligence (AI) has emerged as a promising tool in medical education. It offers opportunities to enhance learning experiences, particularly through the development of virtual standardized patients (SP) integrated with AI for training purposes. These virtual patients are especially useful for teaching history-taking skills (HTS). However, evidence comparing AI-powered virtual standardized patients (AI-VSP) with traditional SP in teaching HTS remains unclear. This study aimed to evaluate the effectiveness of AI-VSP and compare it with conventional SP in teaching HTS to undergraduate medical students. Methods A randomized controlled trial (TCTR20251202012) was conducted among third-year medical students at Duy Tan University, Vietnam. This study included participants in the pre-clinical skills curriculum without prior formal HTS training. Students were randomized at the class level to intervention or control groups via a computer-generated sequence. The intervention group practiced with AI-VSP, while the control group learned with conventional SP. Primary outcomes included pre-test, post-test, and objective structured clinical examination (OSCE) scores. Examiners were blinded to group allocation. Student satisfaction was assessed as a secondary outcome using a 5-point Likert scale. Results A total of 67 medical students were included, comprising 34 in the AI-VSP group and 33 in the SP group. Both groups demonstrated significant improvement in post-test scores compared with pre-test results (p 0.05). Conclusions Learning outcomes and student satisfaction did not differ significantly between the AI-VSP and SP instructional methods. Trial registration TCTR20251202012 (registered on 20/11/2025); retrospectively registered; https://www.thaiclinicaltrials.org/show/TCTR20251202012. Supplementary Information The online version contains supplementary material available at 10.1186/s12909-026-09305-5. Keywords: Artificial intelligence, Virtual standardized patient, History-taking skills, Medical education

What kind of evidence is this?

Supporting AI learning science · Randomized study reported

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

This randomized controlled trial (Trial registration number: TCTR20251202012) was conducted at Duy Tan University in Da Nang, Vietnam. The study took place from April 2025 to October 2025. The study protocol was approved by Duy Tan University and fully adhered to the ethical principles of the Declaration of Helsinki (2013 version). Reporting of this trial followed the Consolidated Standards of Reporting Trials 2025 (CONSORT 2025) guidelines [15].

Limitations and uncertainty · source passage

To address these limitations, the standardized patient (SP) method was developed as an alternative. SP provides students with immersive, clinically realistic experiences within a controlled environment. Numerous studies have demonstrated that SP-based teaching offers a consistent, reliable, and safe learning environment. It also effectively enhances the teaching of HTS. This method helps students gain confidence and procedural familiarity at simulation centers before transitioning to real clinical settings [3–6]. Despite its pedagogical strengths, this method demands substantial resources to achieve optimal results. Training SP requires significant time and manpower, and maintaining these programs incurs ongoing financial and logistical costs [7]. Over time, these challenges have prompted educators to explore more sustainable yet equally effective alternatives.

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 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

  1. 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.
  2. 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.
  3. 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.
  4. 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

Hai Nguyen Ngoc Dang; Thang Viet Luong; Hoa Thi Ha Vo; Linh Thi Khanh Nguyen; Trung Nguyen Tran; Hong Thi Anh Pham; Huong Thanh Truong; Hoa Tran; Toan Thanh Tran; Tien Anh Hoang; Thang Chi Doan; Quan Huynh; Thomas H Marwick. Artificial intelligence–powered virtual standardized patients in teaching history-taking skills to medical students: a randomized controlled trial. BMC Medical Education (2026-04-30). https://doi.org/10.1186/s12909-026-09305-5

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.