AI-assisted Simulated Patient for Clinical Communication Skills Training
Explore outputs from this project supported by the AI Teaching and Learning Exploratory Fund in 2024-25
Project members
Sofia I R Pereira, Sumathi Sekaran, Anna Szabo, Dimitri Gavriloff, Philip Drennan, Vedas Thakrar, Damion Young, Jack Amiry, Suzanne Stewart (MSD).
Project summary
An LLM-based chatbot will simulate patient interactions in an OSCE-like format, creating a low-pressure environment where medical students can practice communication, clinical reasoning, taking medication histories, and counselling patients, while receiving AI-driven feedback to enhance their clinical pharmacology skills and confidence in prescribing medications.
View final project report (PDF) Appendix 1 - Prompts and marking rubric (PDF)
Appendix 2 - Case studies (PDF) Appendix 3 - Guidance (PDF)
Appendix 4 - Transcript of conversation_annotated feedback (PDF)
Appendix 5 - Chatbox Usability Questionnaire (CUQ) and scoring guidelines (PDF)
Appendix 6 - Chatbox Usability Questionnaire (CUQ) results (XLSX)
AI in Teaching and Learning at Oxford Knowledge Exchange Forum, 9 July 2025
Findings from projects supported by the AI Teaching and Learning Exploratory Fund in 2024–25 were presented at the AI in Teaching and Learning at Oxford Knowledge Exchange Forum at Saïd Business School on Wednesday, 9 July 2025.
Project team members each presented a lightning talk to all event participants, and hosted a series of small group discussions.
Follow the links below to view the lightning talk recording and slides for this project.
View presentation slides (PDF)
Project case study
An AI-assisted simulated patient (AI-SP) tool was developed to support postgraduate students in practising clinical history-taking and communication skills. Students were involved in two stages of the tool’s development, providing informal feedback that helped shape its design and implementation. The tool was trialled in three formats: live online, in-person, and offline as a self-directed resource.
In the live online format, students were placed in breakout rooms and given 10 minutes to conduct a simulated telephone consultation with the AI-SP. This was followed by a group debrief to reflect on the interaction and discuss key insights and challenges. A similar approach was used in person, with students using headphones to avoid cross-room interference. The offline version allowed students to access the tool in their own time.
Using the tool live online proved most effective. It enabled real-time troubleshooting, structured peer learning, and minimal audio interference. In-person sessions worked well with suitable equipment, though background noise may pose a challenge. In contrast, independent use without tutor facilitation resulted in limited participation, suggesting that structured delivery encourages deeper engagement.
The AI-SP offers clear educational benefits. It provides a safe, low-stakes environment where students can repeatedly practise clinical conversations, enhancing their learning through iterative rounds of structured automated feedback and reflection.
We also faced a few challenges. The AI sometimes generated slightly inaccurate responses (ie not exactly as described in the case study). Additionally, the quality of automated transcription affected the reliability of the automated feedback, particularly within the ‘interpersonal skills’ section.
Overall, student feedback was overwhelmingly positive. Students were receptive to the tool and eager for further development, suggesting the addition of varied voice types, personalities, patient-initiated questions, and decoy information to increase complexity. They also emphasised the value of immersive, interactive tools in clinical education.
Looking ahead, there is strong potential to expand the case library, improve transcription accuracy, integrate the tool into Canvas, and evaluate its impact on learning outcomes. This experience underscored the potential of AI-driven simulation in developing communication skills and demonstrated the value of collaborative, student-informed innovation in medical education.
