Project title

A multi-site investigation into the effectiveness of an artificial intelligence-powered clinical decision support technology on students’ clinical reasoning when using virtual patients.

Country

UK

Background

Making a diagnosis is arguably the most important professional skill in the General Medical Council’s Outcomes for graduates. One of the key steps in this process – known as clinical reasoning (CR) is gathering information and integrating it with one’s previous knowledge to generate a list of possible diagnoses and choosing the most likely one. Challenges for novices include gathering key information, generating the list, managing diagnostic uncertainty, and avoiding diagnostic error. Access to reference materials or experts can mitigate the problem, however, clinical decision support (CDS) technologies provide a novel, safe, and accessible alternative. CDS systems work by providing clinical and other health information through matching person-specific data inputted by patients or clinicians using machine learning and artificial intelligent (AI) software.

Summary

This project seeks to investigate the effectiveness of CDS technology to support the ability of final-year undergraduate students to accurately identify and refine a list of diagnostic possibilities (including differential diagnosis) as they transition from medical school into clinical practice. Whilst evidence exists that shows CDS improves CR performance at the point of care in the workplace, there has been no published research investigating the use of CDS among medical students. It is also not known whether and how CDS can be employed in the teaching of clinical reasoning. Therefore, our research aims to address this knowledge gap by investigating the impact of CDS on medical students’ CR development, improving CR performance, and reducing diagnostic errors among students. This innovative project will leverage the capability of the first virtual patient (VP*) digital learning resource-powered artificial intelligent CDS technology developed in the UK – the Isabel-EPIFFANY eLearning platform.

Outcome

The anticipated results are that VP cases with CDS are an effective scaffold for CR development among medical students transitioning from medical school into clinical practice. The perceived usefulness of VP cases with CDS is likely to be favourable among medical students as well. The impact of the research will include:

  1. A step-change in the way clinical reasoning teaching is delivered longitudinally and at scale across undergraduate training programmes.
  2. A valid and reliable source of 'big data' for identifying CR components using VP cases that may predict which and when CR errors take place.
  3. A transformative way to provide remediation for novices who struggle with CR due to the capability of the resource to provide VP cases targeted to the personal needs of individuals.

*A VP is a computer-simulated real-life clinical scenario allowing learners to emulate the roles of healthcare professionals to obtain a history, conduct a physical exam, generate differentials, and make therapeutic decisions.