New research shows that artificial intelligence (AI) will soon replace intrusive monitoring of glucose. The move would be especially important for people with diabetes if the technology works.
Many recent media reports suggest the pace of AI growth is slowing.
Despite this, AI engineers continue to build state-of- the-art technology that promises to make everybody’s everyday life simpler.
Some AI systems are being developed to develop therapies for particular health conditions. Scientists have recently developed an AI program which can detect low blood sugar, or hypoglycemia, for example.
The researchers hope this system will allow patients to calculate their blood glucose levels without needing an invasive method called a finger prick test. L
The team recently published in the journal Scientific Reports the findings of a pilot study.
Issues applicable to current methods
Measuring blood glucose for many people currently includes pricking a finger with a needle connected to an apparatus. Then, a continuous glucose monitor (CGM) analyzes the blood sample, which often needs to be calibrated at least twice daily.
The process can be difficult, uncomfortable, and inconvenient especially for children and people in the middle of the night who need to test their blood. As a result, some people can’t measure their levels as frequently or as accurately as necessary.
The researchers behind the current study hope that a non-invasive method will help improve compliance rates, especially among those needing to monitor their levels of glucose closely, such as people with diabetes.
The new AI technology was developed at Warwick University, United Kingdom, and hypoglycemia can be identified from the heart using electrocardiogram (ECG) signals.
The scientists showed in their report that this new technology is 82 percent accurate of the time, a quality close to that of existing CGM systems. Senior study author Leandro Pecchia, Ph.D., university’s associate professor of biomedical engineering, commented:
“Our innovation consisted in using [AI] for automatic detecting [of] hypoglycemia via few ECG beats. This is relevant because ECG can be detected in any circumstance, including sleeping.”
How’s it working out?
Hypoglycemia affects the heart’s electrophysiology, and because it has slightly different effects on the heart of each individual, an AI system allows highly personalized monitoring of the glucose levels.
The team used AI in the recent pilot study to automatically detect nocturnal hypoglycemia from a couple of heartbeat signals captured by a wearable device. The study involved healthy individuals, who were tracked for 14 consecutive days by the scientists for 24 hours a day.
This study was unusual in that the scientists independently tracked the glucose levels of the participants, whereas previous studies had evaluated participants ‘ outcomes as a group.
The authors claim that their new approach captures the tremendous variability of ECG signals among individuals which could not be adequately implemented in previous trials.
Real-time alarm system
The wave-shaped readouts from an ECG machine provide a detailed picture of the action of the heart; each segment of the wave provides information on specific heart events, such as heartbeats.
The authors behind the current study have developed a way to reliably imagine which part of the ECG wave is related to a hypoglycemic event.
This could lead to a real-time alarm system that warns people if their levels of blood sugar change dramatically. Getting such an early warning could significantly shorten the amount of time a person is experiencing hypoglycemia, which can be very dangerous for people with diabetes particularly.
The new method from the team is one example of precision medicine that could greatly enhance the way people treat diabetes. While before this technology becomes available there is still some way to go, the initial results are promising.
Since its conception the AI has been controversial. Nonetheless, while certain innovations that evoke dreams that are reminiscent of George Orwell’s 1984, they may alternatively carry the promise of a future where technological innovation complements medicine to improve the lives of millions.
If successful, the technology tested in this study could pave the way for a lot more applications of AI and heart electrophysiology. It could also be used, with highly personalized accuracy, to treat a number of conditions arising from changes in the blood.
