Enhancing TBL Application Exercises with AI-Powered Patient Personas

Abstract

From static cases to interactive clinical simulations

Team-Based Learning Application Exercises in clinical education often rely on static, text-based case presentations. While these cases are useful for delivering information, they can limit student engagement, reduce opportunities for active inquiry, and provide minimal real-time feedback.

This study explores the use of AI-powered patient personas to transform passive case analysis into interactive, team-based clinical simulations. Two personas, a simulated patient and a simulated doctor, guided students through history-taking, triage, diagnostic reasoning, and structured physical examination.

Conceptual framework

Two AI personas supporting different stages of reasoning

The design centres on two AI personas: a simulated patient, Mr Lim, and a simulated clinician, Dr Tan. The patient persona responds realistically and requires students to actively elicit information through questioning. The clinician persona supports reasoning by prompting students to justify decisions and conduct structured examinations.

The approach shifts the AE from passive case consumption towards active knowledge construction, where students must ask, interpret, justify, and refine their clinical reasoning as a team.

Between stages, students reflect on their performance, adding a metacognitive layer that helps them evaluate not only what they concluded, but how they reached those conclusions.

Student workflow

AI-supported clinical AE sequence

Stage 01

Interact with the AI patient

History-taking and triage

Inquiry-driven

Students begin by interviewing the AI patient. Their task is to take a thorough patient history, ask appropriate questions, listen carefully to the answers, and gather relevant information for triage and clinical assessment.

Ask

Teams formulate targeted clinical questions rather than receiving all case details upfront.

Listen

Students interpret the patient’s responses and identify what information is still missing.

Triage

Teams use the information gathered to assess urgency and likely clinical priorities.

Collaborate

Students build a shared understanding through team discussion and decision-making.

Stage 02

Present findings to the AI doctor

Clinical reasoning and structured examination

Reasoning-focused

Students then present the history they have taken to the AI doctor. They explain their clinical reasoning, discuss possible interpretations, and conduct a structured physical examination with guidance from the AI clinician.

Present

Teams summarise the patient history clearly and clinically.

Justify

Students explain the reasoning behind their triage decisions and priorities.

Examine

Teams work through a structured physical examination process.

Reflect

Students review their performance and identify improvements for future practice.

Learning design value

The two-persona structure mirrors a realistic clinical workflow, moving students from patient interaction to professional handover, reasoning, examination, and reflection.

Participants

88 undergraduate nursing students took part in the intervention.

Data collection

Qualitative feedback and observations captured the student experience.

Focus areas

Engagement, reasoning, communication, realism, and learning outcomes.

Learning mode

Conversational, inquiry-driven, team-based clinical problem-solving.

Findings

Engagement, realism, and questioning skills

Students reported high engagement and realism when interacting with the AI patient. The conversational format encouraged teams to ask questions actively, interpret responses, and construct their clinical understanding progressively.

The AI doctor stage reinforced theoretical knowledge and supported structured learning. Students valued the opportunity to present findings, explain their reasoning, and work through examination processes in a more authentic format.

Identified challenge

Many students struggled with formulating effective questions, indicating that additional scaffolding is needed before and during the activity.

Discussion

Extending the TBL Application Exercise model

The AI-powered persona approach aligns strongly with TBL principles because it requires students to collaborate, ask questions, justify decisions, and respond to feedback. Instead of consuming a static case, students actively construct the case through dialogue.

The intervention also introduces continuous feedback into the AE process. However, to maximise educational value, students need support in developing questioning strategies, clinical communication skills, and structured reasoning processes.

Conclusion

AI personas can make clinical AEs more authentic and adaptive

AI-powered personas can enhance student engagement, clinical reasoning, and communication skills within Team-Based Learning Application Exercises. They offer a scalable and realistic way to create feedback-rich clinical simulations that better reflect real-world practice.

Further refinement is needed, particularly around scaffolding student questioning, but the approach demonstrates strong potential for transforming clinical case-based learning into a more active, collaborative, and reflective experience.

Contact and resources

Project contacts

Ernie Ghiglione

Senior Research Fellow, LAMS Foundation

ernieg@lamsfoundation.org

Tanushry Roy

Senior Lecturer, Programme Director for Bachelor of Science (Nursing Practice) – Year 1, Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore

nurtr@nus.edu.sg

Resources

Presentation