Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality?
Introduction: This research study evaluated the effectiveness of utilising patients simulated through artificial intelligence (AI) for medical education, and the role of AI as a learning tool compared with traditional peer-to-peer formats. A medical education platform, SIMPAT, which generates timed...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Future Healthcare Journal |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2514664525001778 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850071615097798656 |
|---|---|
| author | Pegin Poulose |
| author_facet | Pegin Poulose |
| author_sort | Pegin Poulose |
| collection | DOAJ |
| description | Introduction: This research study evaluated the effectiveness of utilising patients simulated through artificial intelligence (AI) for medical education, and the role of AI as a learning tool compared with traditional peer-to-peer formats. A medical education platform, SIMPAT, which generates timed clinical scenarios of AI-simulated patients with preset clinical backgrounds ranging across a variety of medical specialities, was incorporated in this study. The aim was to understand how this learning style impacted the clinical confidence and knowledge acquisition of participants, and whether it was a time-efficient alternative. Furthermore, the study also obtained the participants’ perceptions of the platforms realism, convenience and intuitiveness. Materials and methods: Medical students and recent medical graduates (FY1 level) from diverse geographic and academic institution backgrounds used SIMPAT over a 1-month period (13 January 2025–13 February 2025) to practice responding to AI-generated clinical scenarios. The participants, 40 in total, were encouraged to take a medical history from the simulated patients, who had been programmed to respond in a humanlike manner. A feedback form was handed to the participants electronically, which included quantitative ratings for intuitiveness and realism of responses (scores graded from 1–5, with 5 being the maximum grade) and free-text feedback on the platform’s strengths, weaknesses and recommendations, and their experiences compared with peer-based learning. Results and discussion: Participants highlighted that learning through AI-simulated patients had many advantages, a common theme being the ability to study a large quantity of clinical scenarios with ease. Many mentioned how effective this method is in reinforcing knowledge, because a large number of simulated patient cases are available for most common clinical scenarios seen in medical school exams. Participants highlighted how this is limited practically when studying with peers. 80% of participants rated the platform’s intuitiveness 4 or higher, while realism ratings were moderate (55% scored 3, 30% scored 4). Participants highlighted the platform’s strengths in providing challenging but relevant questions, immediate feedback, identifying knowledge gaps and boosting confidence. Compared with peer-to-peer learning, SIMPAT was perceived as more time-efficient, because, traditionally, one peer within a pair had to be the ‘patient actor’. However, reduced realism in patient interactions and limited empathy training, along with a few technical errors, were noted as drawbacks. Despite this, 95% of participants recommended the platform, viewing it as a supplement to traditional peer-to-peer methods rather than a replacement. Conclusions: The involvement of AI-simulated patients in medical education and continuous learning is more than just a theoretical tool for improving clinical communication skills, particularly in areas such as efficiency of study, knowledge acquisition and building confidence. Further advancements in this technology can improve the functionality of medical education in the modern age, although addressing its limitations, namely realism and lack of empathy in AI responses, is vital to ensure successful adjunction with conventional strategies within medical education. |
| format | Article |
| id | doaj-art-9000828fb9f549bb80caefb1cc2cbf3e |
| institution | DOAJ |
| issn | 2514-6645 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Future Healthcare Journal |
| spelling | doaj-art-9000828fb9f549bb80caefb1cc2cbf3e2025-08-20T02:47:17ZengElsevierFuture Healthcare Journal2514-66452025-06-0112210039810.1016/j.fhj.2025.100398Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality?Pegin Poulose0WWLIntroduction: This research study evaluated the effectiveness of utilising patients simulated through artificial intelligence (AI) for medical education, and the role of AI as a learning tool compared with traditional peer-to-peer formats. A medical education platform, SIMPAT, which generates timed clinical scenarios of AI-simulated patients with preset clinical backgrounds ranging across a variety of medical specialities, was incorporated in this study. The aim was to understand how this learning style impacted the clinical confidence and knowledge acquisition of participants, and whether it was a time-efficient alternative. Furthermore, the study also obtained the participants’ perceptions of the platforms realism, convenience and intuitiveness. Materials and methods: Medical students and recent medical graduates (FY1 level) from diverse geographic and academic institution backgrounds used SIMPAT over a 1-month period (13 January 2025–13 February 2025) to practice responding to AI-generated clinical scenarios. The participants, 40 in total, were encouraged to take a medical history from the simulated patients, who had been programmed to respond in a humanlike manner. A feedback form was handed to the participants electronically, which included quantitative ratings for intuitiveness and realism of responses (scores graded from 1–5, with 5 being the maximum grade) and free-text feedback on the platform’s strengths, weaknesses and recommendations, and their experiences compared with peer-based learning. Results and discussion: Participants highlighted that learning through AI-simulated patients had many advantages, a common theme being the ability to study a large quantity of clinical scenarios with ease. Many mentioned how effective this method is in reinforcing knowledge, because a large number of simulated patient cases are available for most common clinical scenarios seen in medical school exams. Participants highlighted how this is limited practically when studying with peers. 80% of participants rated the platform’s intuitiveness 4 or higher, while realism ratings were moderate (55% scored 3, 30% scored 4). Participants highlighted the platform’s strengths in providing challenging but relevant questions, immediate feedback, identifying knowledge gaps and boosting confidence. Compared with peer-to-peer learning, SIMPAT was perceived as more time-efficient, because, traditionally, one peer within a pair had to be the ‘patient actor’. However, reduced realism in patient interactions and limited empathy training, along with a few technical errors, were noted as drawbacks. Despite this, 95% of participants recommended the platform, viewing it as a supplement to traditional peer-to-peer methods rather than a replacement. Conclusions: The involvement of AI-simulated patients in medical education and continuous learning is more than just a theoretical tool for improving clinical communication skills, particularly in areas such as efficiency of study, knowledge acquisition and building confidence. Further advancements in this technology can improve the functionality of medical education in the modern age, although addressing its limitations, namely realism and lack of empathy in AI responses, is vital to ensure successful adjunction with conventional strategies within medical education.http://www.sciencedirect.com/science/article/pii/S2514664525001778 |
| spellingShingle | Pegin Poulose Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality? Future Healthcare Journal |
| title | Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality? |
| title_full | Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality? |
| title_fullStr | Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality? |
| title_full_unstemmed | Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality? |
| title_short | Evaluating the role of AI-simulated patients compared with peer-to-peer learning models in the enhancement of medical education: is it beyond theoretical functionality? |
| title_sort | evaluating the role of ai simulated patients compared with peer to peer learning models in the enhancement of medical education is it beyond theoretical functionality |
| url | http://www.sciencedirect.com/science/article/pii/S2514664525001778 |
| work_keys_str_mv | AT peginpoulose evaluatingtheroleofaisimulatedpatientscomparedwithpeertopeerlearningmodelsintheenhancementofmedicaleducationisitbeyondtheoreticalfunctionality |