Abstract 1

AI-powered Clinical Decision Support Technology in a Virtual Patient Case

Ashwin Venkatesh (Blizard Institute, Queen Mary University of London),
Soham Bandyopadhyay (Nuffield Department of Surgical Sciences, University of Oxford);
Rakesh Patel (School of Medicine, University of Nottingham).

Background:

Novices underperform during clinical reasoning (CR) compared to experts due to insufficient knowledge and clinical experience. The Covid-19 pandemic may have inadvertently exacerbated this problem for medical students at a critical time in their development due to limited access to the workplace. Virtual patient (VP) technologies augmented by clinical decision support (CDS) software may offer a novel solution for developing CR in a scalable and sustainable way. A unique academic-industry (Isabel Healthcare) partnership has developed a novel AI-powered CDS software that enables the collaborative creation, storage and iterative development of VP cases. Despite evidence that CDS improves CR performance at the point-of-care in the workplace, the effectiveness of CDS for improving CR performance among students is unknown.

Method:

A multi-centre, parallel-group, non-inferiority randomised controlled trial investigated the effect of a CDS software on the CR performance and subjective experiences of medical students on a VP case. Students were scored on four validated CR assessment tools: the Assessment of Reasoning Tool (ART), the interpretive summary, differential diagnosis, explanation of reasoning, and alternatives (IDEA) assessment tool, the Clinical Reasoning Task (CRT) checklist, and the Summary Statement Assessment Rubric (SSAR). Semi-structured interviews explored student perceptions about using the technology during the CR process.

Outcomes:

The primary outcome measure consists of a composite outcome of ART, IDEA, CRT, and SSAR. The secondary outcome comprises qualitative data regarding student perceptions of the platform and, in the intervention arm, the CDS software. The results will be available by the time of the conference.