Researchers at the University of New South Wales (UNSW) have developed an ultra-lightweight wearable sensor that can continuously monitor patients with heart and lung conditions at home. The so-called AusculPatch records the subtle vibrations of the heart, lungs and blood vessels and, in future, when combined with AI, could alert healthcare professionals at an early stage if the health of the person wearing the monitoring sensor deteriorates.
According to the researchers, the technology could help shift the focus from occasional hospital check-ups to continuous monitoring in the home. This could enable conditions to be detected earlier and potentially prevent hospital admissions.
An alternative to the stethoscope
Cardiovascular diseases and chronic lung conditions remain among the leading causes of death worldwide. Yet many patients are assessed only during brief consultations, meaning that significant changes in their health may go unnoticed. “What we have developed is a small, portable device that can be attached to the chest and records heart and breathing sounds,” says lead researcher Hoang-Phuong Phan of UNSW. “Technically speaking, it is intended as an alternative to the stethoscope, which doctors use to assess heart and lung conditions.”
This could be particularly beneficial for people who live far from a hospital or are reluctant to see a doctor for mild symptoms. According to co-researcher Anthony Sunjaya, conditions are sometimes only detected once they have already worsened significantly. Furthermore, regular consultations provide only a snapshot lasting around fifteen minutes, meaning abnormalities can easily be missed.
Subtle vibrations
Weighing just 3.2 grams and measuring 20 by 47 millimetres, the AusculPatch is considerably more compact than many existing wearable monitoring systems. The flexible patch is secured to the chest or over a peripheral artery using medical tape. The sensor contains an ultra-thin silicon chip that detects extremely small mechanical vibrations transmitted through body tissue from the heart, lungs and blood vessels. Unlike conventional microphones, the sensor also records very low-frequency vibrations that are difficult to detect with existing wearables.
As a result, the patch can record not only heart sounds but also breathing patterns, pulse waves and blood flow vibrations. The researchers also developed a design that largely suppresses ambient noise. According to initial tests, the sensor continued to record reliable heart sounds even during conversations or in noisy environments.
Although the technology has so far only been tested on a limited number of healthy volunteers, the measurements closely matched those of existing clinical instruments, including electrocardiograms (ECGs), ultrasound, blood pressure monitors and digital stethoscopes. The sensor also continued to collect reliable data continuously during everyday activities such as walking, climbing stairs, working and eating.
Early warning
According to the researchers, AusculPatch differs from smartwatches and fitness trackers in that it not only measures vital signs but also directly collects mechanical information about the functioning of the heart and lungs. This could offer new possibilities for monitoring chronic conditions, sleep disorders and heart valve abnormalities.
A key part of the further development is the use of AI. As the patch collects large amounts of physiological data, machine-learning algorithms will, in future, be able to recognise patterns that indicate the early onset of a disease’s progression. “We can apply machine learning to recognise abnormal signals, alert patients and, at the same time, inform their doctor,” says co-researcher Chi Cong Nguyen. “The ultimate goal is a system that automatically flags worrying changes before serious symptoms arise.”
In addition to cardiovascular applications, the scientists also investigated other possibilities. For example, the sensor proved capable of recording vibrations of the vocal cords. In a proof-of-concept, an AI model even succeeded in recognising spoken words, which could be used to wirelessly control a robotic arm. In the long term, this technology could provide support to people with speech impairments or severe physical disabilities.
The researchers are now preparing larger clinical trials involving around two hundred patients with conditions including heart valve disorders and implanted cardiac support systems. This will be followed by trials involving around a thousand patients to further refine the AI algorithms. According to Phan, it is expected to take another four to five years before the technology can be used as a medical device. Consumer versions aimed at general health monitoring could potentially become available sooner.
