Philippine researchers develop AI model for accessible heart monitoring

Advanced heart monitoring no longer confined to hospital walls
The AI model could bring reliable cardiac assessment to clinics and health centers that currently lack specialized equipment.
Mark

Why does it matter that this works with skin sensors instead of the old equipment?

Mimi

Because the old equipment is expensive, requires trained specialists, and only exists in big hospitals. A rural clinic can't afford it. This system uses data any clinic can collect and lets algorithms do the analysis.

Mark

But 97.78% accuracy—is that good enough for doctors to actually trust it?

Mimi

It's remarkably high for a medical AI. Doctors will want to validate it in their own patient populations, which is exactly what the researchers plan to do next. But yes, that accuracy suggests it could be reliable enough to guide real clinical decisions.

Mark

The story mentions young people getting heart disease now. Why is that happening?

Mimi

Obesity, diabetes, high blood pressure—these metabolic conditions are spreading globally, and they're showing up in younger people than they used to. The Philippines has 27 million overweight or obese adults. That's the soil where heart disease grows.

Mark

So this AI tool could catch disease earlier in people who wouldn't normally get screened?

Mimi

Exactly. If a provincial clinic can now do reliable cardiac assessment with a few skin stickers, they can identify at-risk people before they have a heart attack. That's the real potential here—prevention, not just treatment.

Mark

What's the next hurdle?

Mimi

Proving it works across different populations—different ages, body types, health backgrounds. And simplifying it further. The researchers want to see if they can get the same accuracy with even fewer measurements.

  • Cardiovascular disease is accelerating among young Filipinos, with 27 million overweight or obese and nearly 101,000 deaths from ischemic heart disease last year alone — yet advanced heart monitoring remains locked inside major city hospitals.
  • The gap between who needs cardiac assessment and who can access it is not a technical problem but a structural one, leaving rural and underserved communities with the highest risk and the least recourse.
  • Ateneo researchers answered this inequity with an AI model that replaces expensive hemodynamic analyzers with simple skin sensor stickers, processing basic physiological signals through neural networks to predict cardiac index at 97.78% accuracy.
  • The findings, published in a peer-reviewed journal in April 2026, signal that the barrier to reliable heart monitoring may now be lower than the medical establishment has long assumed.
  • The team is pressing further — testing the model across diverse populations and working to reduce the number of measurements needed, aiming for a future where a provincial clinic can offer the same cardiac clarity as a Manila hospital.

At Ateneo de Manila University, researchers have built an artificial intelligence system capable of assessing heart function with 97.78% accuracy using only noninvasive skin sensors — no specialized equipment, no hospital required. The work arrives as cardiovascular disease quietly reshapes itself into a younger, broader crisis: in the Philippines alone, ischemic heart disease claimed nearly one in five lives last year, while millions more carry undiagnosed risk. What this team has done is not merely engineer a diagnostic tool, but attempt to redraw the boundary between who receives careful cardiac attention and who does not.

A research team at Ateneo de Manila University has developed an AI system that assesses how well a heart is pumping blood using nothing more than small sensor stickers placed on a patient's skin. Led by Patricia Angela Abu from the university's Department of Information Systems and Computer Science, the model predicts cardiac index — the standard clinical measure of heart function — with 97.78% accuracy. Their findings appeared in the April 2026 issue of Bioengineering.

The significance lies in what the system replaces. Measuring cardiac index has traditionally demanded expensive hemodynamic analyzers, controlled hospital environments, and trained specialists. This model instead draws on basic physiological data — heart rate, stroke volume, cardiac output — collected noninvasively and processed through modern algorithms, removing the need for equipment most clinics simply do not have.

The timing is not incidental. Cardiovascular disease is no longer confined to older populations. The WHO has documented rising rates of heart disease among people in their twenties and thirties, driven by the global spread of obesity, hypertension, and diabetes. In the Philippines, the burden is acute: ischemic heart disease caused nearly one in five deaths last year, more than 13% of adults have elevated blood pressure, and an estimated 4.7 million Filipinos have diabetes — with another 2.8 million undiagnosed.

Yet the monitoring that could catch these conditions early remains concentrated in large urban hospitals. Rural areas and smaller facilities lack both the equipment and the expertise, meaning those most at risk are often least able to access care. The AI model offers a path around this bottleneck — bringing reliable cardiac assessment to clinics that currently cannot offer it.

The researchers plan to validate the model across more diverse populations and to simplify the process further, reducing the number of measurements required. Their broader aim is a world in which a patient in a provincial health center receives the same quality of cardiac assessment as someone in Manila — not as an aspiration, but as a practical reality.

A team of researchers at Ateneo de Manila University has built an artificial intelligence system that can measure how well a heart is pumping blood—a critical piece of medical information—using nothing more than small sensor stickers placed on a patient's skin. The model, led by Patricia Angela Abu from the university's Department of Information Systems and Computer Science, achieved 97.78% accuracy in predicting cardiac index, the standard measure clinicians use to evaluate heart function and decide on treatment.

What makes this work significant is what it replaces. Traditionally, assessing cardiac index requires expensive hemodynamic analyzers, a controlled hospital environment, and specialized medical staff trained to operate the equipment and interpret the results. The new system uses basic physiological data—heart rate, stroke volume index, cardiac output—collected through noninvasive sensors and processes it through modern algorithms. The researchers published their findings in the April 2026 issue of Bioengineering, a peer-reviewed journal, under the title "Robust Non-Invasive Cardiac Index Prediction via Feature Integration and Data-Augmented Neural Networks."

The timing of this development matters. Cardiovascular disease is no longer a problem confined to aging populations. The World Health Organization has documented a troubling shift: people in their twenties and thirties are developing heart disease at rising rates, a trend the organization attributes to the global spread of metabolic conditions like obesity, hypertension, high cholesterol, and diabetes. In the Philippines specifically, the numbers are stark. Ischemic heart disease—the most common type—killed nearly 101,000 people last year, accounting for almost one in five deaths nationwide. More than 13% of Filipino adults have elevated blood pressure. Roughly 27 million Filipinos are overweight or obese. The International Diabetes Federation estimates that 4.7 million Filipino adults have diabetes, with another 2.8 million undiagnosed.

Yet access to the kind of detailed heart monitoring that could catch disease early remains deeply unequal. Advanced cardiac assessments are concentrated in major hospitals in large cities. Rural areas and smaller health facilities lack both the equipment and the expertise. This means people at highest risk—those with metabolic conditions, those in underserved regions—are often the least able to get the monitoring that could help them. The AI model offers a way around this bottleneck. By reducing the need for specialized equipment and expertise, it could bring reliable heart assessment to clinics and health centers that currently cannot offer it.

The researchers are not stopping here. They plan to test the model across more diverse populations to ensure it works equally well across different ages, body types, and health backgrounds. They are also exploring whether they can reduce the number of measurements required even further, making the process faster and simpler. The vision they are working toward is one in which advanced heart monitoring is no longer something you have to travel to a major hospital to receive—where a patient in a provincial clinic or a rural health center could get the same quality of cardiac assessment as someone in Manila.

Advanced cardiovascular assessment is no longer confined to the walls of large hospitals and specialized clinics
— ADMU Research Communications Section
Reliable cardiovascular assessment may not always require complex or resource-intensive procedures
— ADMU Research Communications Section
Contattaci Domande frequenti