Funded by a Heart Foundation Vanguard grant, a prospective observational study was conducted during which people recently diagnosed with heart failure at Western Health wore the “Narrative Clip” camera for up to 30 days. Images were captured every 30 seconds, resulting in approximately 100,000 images. These images have been annotated using bespoke software, with only preliminary analysis undertaken. This image database provides a unique opportunity to apply machine learning analysis to better understand self-management in heart failure, and inform future interventions.
The overall aims are to apply machine learning techniques to an existing database of wearable cameras images to determine patterns of self-management in people with heart failure, which will inform a future nurse-led intervention. The novel application of machine learning algorithms to better determine patterns of self-management practices and to determine whether self-management behaviours predict rehospitalisation are critical to identify opportunities for future intervention to improve self-management. In doing so, this work has the potential to improve patients’ quality of life, reduce their unnecessary frequent hospitalisation, and associated economic costs. We have key collaborations with the Insight group at Dublin City University, Ireland. This will transform our understanding of home-based care for heart failure patients by shifting the focus from reactive care to proactive self-management support, which has strong relevance to CRE.
Dr Teketo Tegegne
Prof Ralph Maddison