Apple Watch now take ECG readings and detect blood diseases
Despite obvious shortcomings, Apple Watch has long grown from the wrist receiver of incoming notifications to a full medical device. Already, the clock can not only measure your pulse but also identify diseases such as diabetes, hypertension and sleep apnea. Recently this small but confident list has been replenished with the ability to take ECG readings and measure the potassium concentration in the plasma.
Unlike hypertension, sleep apnea and diabetes, which are diagnosed exclusively with software, a KardiaBand bracelet is required to conduct an electrical cardiac activity (ECG) study and to detect hyperkalemia. On its body, there is an electrode that reads data on the work of the heart, on the basis of which artificial intelligence makes conclusions about the potassium concentration in the patient’s blood.
In order to teach KardiaBand to reveal signs of hyperkalemia, an analysis of the ECG data of two million people collected over the last 20 years was carried out. Due to the vastness of the study, the accuracy with which the bracelet diagnoses an increased potassium concentration is between 90 and 94%. The uniqueness of the expert’s work lies in the ability to perform the analysis in a non-invasive way, which does not require actual blood sampling.
Apple Watch has taught to identify diabetes at an early stage
Apple Watch is able to diagnose diabetes at an early stage with an accuracy of 85%, found scientists at the University of California in San Francisco in conjunction with specialists from the start-up Cardiogram. The study examined data from 14,000 users, 463 of whom had clocked undiagnosed diabetes.
According to Johnson Xie, founder of the Cardiogram start-up, to diagnose diabetes, Apple Watch does not need to have access to the user’s blood. All manipulations are non-invasive. To detect signs of the disease, the clock uses the heart rate monitor and specialized software.
“Our heart is connected to the pancreas through the autonomic nervous system,” he says. – People suffering from diabetes, regardless of the stage, are characterized by a variability in the heart rate. A study conducted by the Framingham Heart Study in 2015 showed that low heart rate variability combined with tachycardia often precursors of diabetes. “
As in the case of hypertension and sleep apnea, self-learning neural networks are used for diagnosis, before analyzing data from more than 33,000 people suffering from diabetes and other diseases this will allow the clock not only to detect signs of diabetes but also not to confuse it with other ailments.