Heart failure is one of the biggest health problems facing older adults around the world. It happens when the heart can no longer pump blood as well as it should.
This does not mean the heart has stopped working, but it does mean that the body does not receive enough oxygen and nutrients to meet its needs. Over time, people may develop tiredness, shortness of breath, swollen legs, difficulty exercising, and repeated hospital visits. In severe cases, heart failure can shorten life.
Around 64 million people worldwide are living with this condition, and the number continues to grow as populations age. Because the disease often develops slowly over many years, doctors have been searching for better ways to identify people at risk before symptoms appear.
Researchers from the Technion Faculty of Biomedical Engineering in Israel have now developed a new artificial intelligence system that could help solve this problem. Their study, published in npj Digital Medicine, describes an AI model called DeepHHF that can identify people who are likely to develop heart failure years before they receive a diagnosis.
The research was led by Professor Joachim Behar and PhD student Eran Zvuloni together with scientists and doctors from Rambam Health Care Campus, Shaare Zedek Medical Center, the Hebrew University of Jerusalem, and Leumit Health Services.
Current methods usually detect heart failure only after damage has already occurred or symptoms become obvious.
Early detection matters because doctors can often slow disease progression through healthier lifestyles, better blood pressure control, diabetes management, medications, and regular follow-up. Finding high-risk patients earlier could help prevent serious complications and reduce hospital admissions.
The researchers trained DeepHHF using about 70,000 Holter ECG recordings collected through Leumit Health Services.
A Holter monitor is a small portable device that records the heart’s electrical activity continuously for 24 hours while people go about their normal daily activities. It is already widely used because it is painless, safe, and does not require surgery or hospital admission.
Instead of looking only for obvious heart rhythm problems, the AI searched for tiny patterns hidden within the ECG recordings. These subtle signals are usually impossible for the human eye to detect. By analysing these hidden patterns, the system estimated whether someone faced a higher chance of developing heart failure in the future.
According to Professor Behar, the model can predict heart failure risk up to five years before clinical disease develops. If confirmed in further studies, this would give doctors valuable time to intervene before permanent heart damage occurs. Earlier treatment could improve quality of life, lower healthcare costs, reduce emergency hospital visits, and potentially save lives.
Artificial intelligence is becoming an increasingly important tool in medicine because it can recognise complex patterns in very large datasets much faster than humans.
However, AI is designed to support healthcare professionals rather than replace them. Doctors still need to consider a patient’s symptoms, medical history, physical examination, and other test results before making decisions.
The researchers believe their model could fit naturally into existing healthcare because it relies on a routine test that many patients already undergo. This means hospitals and clinics may not need expensive new equipment if the technology is eventually introduced into clinical practice.
Although the results are very encouraging, more research is still needed. Scientists will need to test the model in different countries and among people from a wider range of backgrounds before it can become part of everyday medical care.
Future studies will also need to show whether acting on these early warnings truly reduces heart failure and improves long-term outcomes.
Overall, this study represents an exciting step toward preventive medicine. Instead of waiting until patients become sick, doctors may one day identify hidden risk years in advance and take action earlier. The research combines a simple, non-invasive heart test with advanced artificial intelligence to provide information that was previously unavailable.
While the technology still requires further validation, its potential to improve patient care is considerable. If future clinical studies confirm these findings, DeepHHF could become a valuable tool for reducing the global burden of heart failure.
If you care about health, please read studies about the benefits of low-dose lithium supplements, and what we know about egg intake and heart disease.
For more health information, please see recent studies about potatoes and high blood pressure, and results showing 6 best breads for people with heart disease.
Source: Technion.
