A program evaluation published in BMC Medical Education reports that brief, individualized instruction successfully prepares fourth-year medical students to construct prompts for artificial intelligence systems in evidence-based medicine. Conducted by researchers Erik Langenau and Hsinliang Chen, the investigation examined an instructional model aimed at developing prompt engineering as an emerging competency in medical education. The study evaluated whether focused, one-on-one guidance could teach medical students structured frameworks for formulating queries within a closed interdisciplinary database to address clinical questions.
The educational initiative enrolled nineteen fourth-year medical students who participated in a structured three-part curriculum facilitated by faculty members. Each participant completed three separate forty-five-minute, one-on-one learning sessions designed to provide personalized instructional support. Rather than relying on open-ended general artificial intelligence applications, the curriculum integrated prompt construction exercises directly within Scopus AI, a closed, interdisciplinary database environment. The instructional team structured the sessions around practical clinical inquiry, guiding students through the process of querying the system to support evidence-based medicine during point-of-care scenarios.
During the faculty-led sessions, students received formal training in two distinct prompt engineering frameworks: PICO, which organizes queries around Patient, Intervention, Comparison, and Outcome, and CLEAR, which structures prompts to be Concise, Logical, Explicit, Adaptive, and Reflective. The curriculum aimed to determine how effectively these two conceptual frameworks could be applied to translate complex medical questions into productive artificial intelligence searches. To assess the educational efficacy of the instructional approach, the researchers tracked participant development using weekly evaluations administered after each session and conducted structured group interviews upon completion of the entire program.
Student confidence in prompt generation was evaluated using a four-point scale ranging from one, representing strongly agree, to four, representing strongly disagree. Across the three weeks of individualized instruction, reported mean confidence remained consistently high, registering at 1.16 following the first week, 1.32 after the second week, and 1.05 upon concluding the third week. These longitudinal evaluation metrics indicated sustained self-assurance among the students as they developed and refined their querying skills.
The evaluation also highlighted clear student preferences regarding the practical utility of the two prompt frameworks. Participants favored the PICO framework for clinical workflows, reporting that its components were more straightforward and directly applicable for rapid literature searches compared to the CLEAR methodology. Participants specifically valued Scopus AI for delivering credible, verifiable, and peer-reviewed references, contrasting this closed environment with open general artificial intelligence tools such as ChatGPT that lack equivalent academic source validation.
The study authors concluded that brief, individualized instruction offers an effective educational design for advancing artificial intelligence literacy and prompt engineering competencies among medical trainees. Utilizing a closed artificial intelligence system served to mitigate potential informational inaccuracies while systematically reinforcing student confidence in applying evidence-based medicine principles to clinical queries. The authors noted that this three-part, one-on-one format provides a viable curricular framework for embedding prompt engineering into fourth-year clinical rotations to aid point-of-care decision-making. However, the evaluation emphasized that expanding such individualized instructional models to broader student populations will require balancing institutional faculty time demands against overall educational value.
Subject of Research: AI prompt engineering instruction in medical education
Article Title: A program evaluation for teaching AI prompt engineering for evidence-based medicine to fourth year medical students
Article References: Langenau, E., & Chen, H. (2026). A program evaluation for teaching AI prompt engineering for evidence-based medicine to fourth year medical students. BMC Medical Education. https://doi.org/10.1186/s12909-026-10790-x
Image Credits: AI Generated
DOI: 10.1186/s12909-026-10790-x
Keywords: medical education, prompt engineering, evidence-based medicine, artificial intelligence, Scopus AI, PICO framework, clinical rotations, large language models, program, evaluation, teaching, prompt
