Leads: Professor David Clifton and Professor Peter Watkinson
Electronic records and other sources of data contain data that can be used to identify previously unrecognised health conditions. We are building on our earlier work detecting hospital patients at risk of undiagnosed chronic hypertension and confirming this diagnosis in the community. We are focusing on recognising undiagnosed atrial fibrillation, responsible for ten percent of strokes.
We are training machine learning (ML) algorithms to combine routine data from linked electronic patient records with next-generation monitoring to uncover previously unrecognised patterns of disease. After that, we will define the technology-enabled care pathways to deliver and optimise the patient care using these findings.
We are extending this approach to detect other unrecognised conditions in hospitalised patients, and digitally support early treatment in the community.