Researchers from the University of Surrey, University of Warwick and University of Florence respectively have developed a way AI could save human lives rather than posing as a risk to them.
Using a method that makes use of neural networks the team have trained an AI machine learning system that can detect congestive heart failure (CHF) from just a single heartbeat, with results that are 100% accurate. This development for early and efficient detection would be embraced by doctors, healthcare systems and other clinical practitioners alike as CHF can put a strain on care systems due to sustained healthcare costs.
The costs are growing with a significant part of healthcare budgets going on hospitalization for patients suffering from CHF. Due to the gravity of the situation researchers believe that healthcare systems across the world “urgently require efficient detection processes”.
It is thought that CHF affects around 26 million people around the world and because it is a degenerative condition it typically worsens with age as the degree of severity reaches a peak – up to a 40% increase in mortality. The condition itself affects how blood is circulated throughout the body and can go on to cause other issues and complications as a consequence, such as making it more difficult for the liver to function properly. Therefore, early detection is vital as it can lessen the burden on the infrastructure of healthcare systems but also preserve the lives of those at risk or suffering from CHF.
The latest breakthrough is hoped to solve the ongoing concerns doctors and practitioners have with regards to patient’s lives and the other consequences of CHF. The AI neural network system used throughout the research makes for compelling reading when identifying patterns and structures in data sets due to its effectiveness.
We trained and tested the CNN model on large publicly available ECG datasets featuring subjects with CHF as well as healthy, non-arrhythmic hearts. Our model delivered 100% accuracy: by checking just one heartbeat we are able detect whether or not a person has heart failure. Our model is also one of the first known to be able to identify the ECG' s morphological features specifically associated to the severity of the condition.
Dr Sebastiano Massaro, Associate Professor, The University of Surrey
The research itself has been largely carried out on datasets from CHF patients with severe progression of the disease but the technology does allow for “rapid interventions”. The team believe that by continuing their research the technology would not only benefit those suffering from CHF but also provide early detection schemes and become an asset for other groups at risk of degenerative illness.
In addition to their findings published in Biomedical Signal Processing and Control Journal, the team have discovered that this form of CHF detection vastly improves on current methods as it “yields prominent accuracy” and is also “error-free”. Thus, doctors and clinical practitioners would have the tools at their disposal to make a significant difference concerning early and reliable diagnosis. This technology could also reduce the financial burden on global healthcare systems as it could help impact emergency situations and hospitalization as well as being time efficient.
Massaro and his team hope that through continued research and development of their approach this technique will soon be applied to everyday scenarios and practices. This promises to be what Massaro calls, “an open breakthrough frontiers for both clinical research and practice.”