The 'Artificial Intelligence (AI) for healthcare professionals' e-learning session has now gone live on the NHS England portal, which will describe AI technologies and how they relate to healthcare.
The session was commissioned by the Society of Radiographers in partnership with NHS England E-Learning for Healthcare (ELFH), in anticipation of the Health and Care Professions Council (HCPC) updating the Standards of Proficiency for Allied Health Professionals.
SoR and ELFH contacted Dr Sonyia McFadden, a senior lecturer in diagnostic radiography and imaging at Ulster University and Chair of the SoR AI advisory team, to develop a “beginner’s guide” to the concepts of AI relevant to imaging and healthcare professionals.
The e-learning session uses examples of AI in day-to-day life to introduce fundamental concepts and terminology, before using these concepts to show how AI could be applied to the realm of healthcare.
For example, the session explains how an AI might be trained to recognise fractures in X-rays by being provided a dataset of images labelled whether it was a fracture or not – these images are fed into a machine learning algorithm.
(Image provided by Geraldine Doherty, Ulster University)
As part of this algorithm, however, the AI may “learn something unexpected.”
Machine learning models cannot be restricted in terms of the data within each dataset they use, and so may end up focusing on “unexpected and inappropriate signals” in the data to provide an answer.
The “Introduction to Data Science and AI for senior researchers: Problems with AI” provides a recent real life example of this, where a machine learning model was used to distinguish cases of Covid on X-rays that were classed as either clinically severe versus clinically moderate.
In the training data, the X-rays of Covid positive patients who were most severe were taken lying down/supine during their chest X-ray examination, and so the model learned to distinguish X-rays of people lying down versus erect.
The session also focuses on potential ethical and regulatory concerns, and enables users to develop the skills to critically appraise literature on AI, and contribute to discussion around these tools.
Geraldine Doherty, diagnostic radiographer and PhD student at Ulster University, said: “We hope that the session serves as a sound foundation to allow users to better understand applications of AI in relation to their work, and to facilitate informed, safe and effective use of AI tools.”
The session is now available here, along with a pre and post learning survey.
If you are interested in contributing to the ongoing discussion around education on artificial intelligence for medical imaging staff, there are opportunities to participate in focus groups and interviews on the topic. For more information, download the focus group participant information sheet or contact Geraldine Doherty via email at [email protected].