AI and machine learning technologies are changing how we study and interpret electrocardiograms (ECG). To ensure they are fit for purpose, it is important that their development is closely aligned with the priorities of those making clinical decisions and patients receiving care, from the earliest stages.

At ICE 2024 we presented a study investigating clinicians’ attitudes towards AI technologies and in particular the potential of a novel approach of pseudo-colouring ECGs to expose life-threatening changes (QT-prolongation and STEMI), which show how an automated, explainable, algorithm is working.

We conducted a series of interviews with junior doctors and clinicians from five specialties (emergency medicine, anaesthesia and critical care, cardiology, primary care, and paramedics) in the UK, spanning all career stages. Our study: (i) documents their approach to computational ECG interpretation, (ii) explores the potential of our novel pseudo-colouring approach (representing abnormal changes in signal duration/amplitude using a sequence of colours) and other future ‘human-like’ computing approaches to facilitate ECG interpretation and support clinical decision making, and (iii) elicits their opinions about the importance of explainability and trustworthiness of AI algorithms. The results show that clinicians have an overwhelmingly positive attitude toward digitally enhanced and AI supported ECGs and our pseudo-colouring visualisation approach is their preferred method of presentation, when compared with other contemporary approaches, such as fully automated read-outs and ECG measurements.

Based on our preliminary work and the stakeholder engagement study, ECG technology based on AI- driven human-like computing applications seems to be a promising way both to facilitate interpretation of ECGs, and to highlight difficult to spot or overlooked pathologies, helping to make ECGs easier to understand for clinicians and laypeople alike.