Technology has changed how we monitor our health with a click of a few buttons.
When tech company Apple decided to make a smartwatch, the premium smartphone-maker wanted it to be more than just a device to stay connected. It impressed fans around the world with the in-built health tools such as fall detection, electrocardiogram (ECG), atrial fibrillation (A-fib) detection, heart rate and oxygen level monitor that could keep people aware of health emergencies round the clock. Time and again, users have documented how their Apple Watch saved their lives.
But, owning an Apple Watch alone won’t keep you up to date about your health. A new study has found that the accuracy of the smartwatch can vary from person to person.
According to The Canadian Journal of Cardiology study, the Apple Watch can show an accurate diagnosis of A-fib only in a limited number of patients with similar clinical profiles. Apple developers still have a long way to go when it comes to patients who have a variety of coexisting ECG abnormalities.
Published by Elsevier, the findings show that smartwatch detection of A-fib has great potential, but can be challenging in patients who have a pre-existing cardiac disease.
“With the growing use of smartwatches in medicine, it is important to know which medical conditions and ECG abnormalities could impact and alter the detection of A-fib by the smartwatch to optimise the care of our patients,” said lead investigator Marc Strik from LIRYC institute, Bordeaux University Hospital, Bordeaux, France.
The investigators believe better algorithms and machine learning can help these tools provide more accurate diagnoses.
For the study, the researchers collected data from 734 hospitalised patients. Each patient underwent a 12-lead ECG, immediately followed by a 30-second Apple Watch recording.
The smartwatch’s automated single-lead ECG A-fib detections were classified as ‘no signs of atrial fibrillation’, ‘atrial fibrillation’, or ‘inconclusive reading’. In approximately one in every five patients, the smartwatch ECG failed to produce an automatic diagnosis.
The risk of having a false positive automated AF detection was higher for patients with premature atrial and ventricular contractions (PACs/PVCs), sinus node dysfunction, and second- or third-degree atrioventricular block.
The app correctly identified 78 per cent of the patients who were in A-fib and 81 per cent who were not in A-fib. The electrophysiologists identified 97 per cent of the patients who were in A-fib and 89 per cent who were not.
“These observations are not surprising, as smartwatch automated detection algorithms are based solely on cycle variability. Ideally, an algorithm would better discriminate between PVCs and A-fib. Any algorithm limited to the analysis of cycle variability will have poor accuracy in detecting AT/AFL. Machine learning approaches may increase smartwatch AF detection accuracy in these patients,” Dr Strik noted.
The smartwatch algorithms for the detection of A-fib in patients with cardiovascular conditions are not yet smart enough. But they may soon be, the study said.
Last week, the Apple Watch surprised everyone when it reportedly detected a woman’s pregnancy before she was even aware of it.
The 34-year-old woman took to Reddit to share her experience and said that the watch indicated that her average resting heart rate had significantly increased in just a few days. She realised there was something off about the development.
“Usually, my resting heart rate is about 57 and my heart rate has increased to 72. It’s not a big jump, but it showed up on an alert that it’s been higher for 15 days. I started trying to figure out why. The watch knew I was pregnant before I knew it! I would have never tested without wearing my watch because I have not had a period to be late on one,” she wrote on the platform.
Meanwhile, the tech giant unveiled Apple Watch Series 8 last month. The new and improved series introduced best-in-class health features such as Crash Detection for severe car accidents and an innovative temperature sensor that will make it easy for women to monitor their health.