Doctors may soon have a safer way to track blood pressure in critically ill patients. Engineers at Johns Hopkins University have built a wearable sensor system that uses artificial intelligence to estimate blood pressure almost as accurately as the invasive tubes used today. The new research appears in the journal Computers in Biology and Medicine.
The Problem With Current Monitoring
Patients in intensive care units often need constant blood pressure checks. Their levels can swing quickly. High blood pressure raises the risk of stroke, heart attack, and kidney damage. Low blood pressure can starve the brain and organs of blood.
To catch these swings, hospitals use arterial lines. These are thin tubes inserted directly into an artery, usually in the arm or groin. They give doctors a live, constant readout of blood pressure. But they carry serious risks. Patients can develop bleeding, blood clots, or infections. The tubes also restrict how much a patient can move.
Regular blood pressure cuffs are much safer. But they only take occasional readings, not a continuous stream. That gap can be dangerous when a patient’s condition changes fast.
A New Approach
A team led by PhD student Carl Harris set out to close that gap without using needles or tubes. They created a system called MOSAIC. It relies on two small sensors. One sits on the chest and tracks the heart’s electrical signals. The other sits on a finger and tracks blood flow.
Both sensors send their data to a deep learning model. The model combines the signals and builds a full blood pressure waveform. This is a continuous, moving picture of blood pressure over time, similar to what an arterial line produces.
“We wanted to find a better way,” Harris said, describing the risks that come with standard arterial catheters.
Early Results Are Promising
The researchers tested MOSAIC on 28 patients in the Johns Hopkins Hospital intensive care unit. The sensor readings closely matched the results from traditional arterial lines.
“We’re very close to hitting that gold standard,” Harris said. He called the results a strong sign that the approach works.
Senior author Robert Stevens, who leads informatics and innovation efforts at Johns Hopkins Medicine, said the system offers real accuracy without the danger. “We reconstruct waveform data in a way that’s meaningful, accurate, reliable and, most importantly, non-invasive,” he said.
The team is now testing the system on a larger group of ICU patients to confirm these early findings.
The researchers see a much bigger future for this technology. They believe it could eventually reduce the need for arterial lines altogether. It might also let hospitals monitor blood pressure in general wards, not just intensive care.
There’s also a case for home use. Millions of people worldwide live with high blood pressure, one of the most common and deadly conditions on the planet. The team hopes people with hypertension could wear these sensors every day, the same way many diabetics wear glucose monitors.
Even healthy people could benefit. Currently, doctors know very little about how blood pressure shifts during a normal day.
“We observe sick patients in the intensive care unit, but we have no idea what’s going on with blood pressure in a healthy person who’s just living their life, going to work and being with their family,” Stevens said. “Nobody really knows.”
Who Worked on This
Besides Harris and Stevens, the study included Johns Hopkins researchers Bright Nnadi, Sampath Rapuri, and John Rattray, along with engineers Francesco Tenore and Ralph Etienne-Cummings from the Johns Hopkins Applied Physics Laboratory.
The project received funding from the Johns Hopkins Institute for Clinical and Translational Research, the National Institutes of Health, and the National Science Foundation.
