Apple Watch can detect diabetes with an 85% accuracy

Date:14 February 2018 Tags:,

The Apple Watch can detect diabetes in those previously diagnosed with the disease with an 85 percent accuracy, according to Cardiogram founder Brandon Ballinger’s latest clinical study.

Although the outcome of this study confirms facts already established, it reveals the potential of using a heart rate monitor together with Cardiogram’s AI-based algorithm to detect diseases before they have been diagnosed.

The study is part of the larger DeepHeart study with Cardiogram and UCSF, using data from 14,000 Apple Watch users. By using the Watch’s heart rate sensor, the same type of sensor other fitness bands using Android Wear also integrate into their systems, the researchers were able to detect that 462 of the users had diabetes.

Diabetes is a huge — and growing — problem in South Africa. According to Dr Larry Distiller, founder and managing director of the Centre for Diabetes and Endocrinology in Johannesburg, “The diabetes tsunami is here. And we in South Africa are in trouble.” “Three-and-a-half million South Africans (about 6% of the population) suffer from diabetes and there are many more who are undiagnosed,” he cautioned in an interview on Health24, last year..

If Ballinger and his colleagues’ findings can help people detect diabetes early on, it could help in cutting down on diabetes-related diseases before they progress. There have been other attempts to build special-purpose glucose-sensing hardware however, this is the first large-scale study showing that ordinary heart rate monitors—when paired with an artificial intelligence-based algorithm—can identify diabetes with no extra hardware.

Previously, Ballinger and his colleagues were able to use Apple’s Watch to detect hypertension with an 82 percent accuracy, an abnormal heart rhythm with up to a 97 percent accuracy, sleep apnea with a 90 percent accuracy when paired with Cardiogram’s AI-based algorithm. Most of these discoveries have been published in clinical journals or abstracts and Ballinger intends to publish the latest findings after presenting at the (Association for the Advancement of Artificial Intelligence (AAAI) 2018 conference this week.

Going forward, Ballinger and his colleague on the study Johnson Hsieh, mentioned they could be looking at a number of diseases to detect through heart sensors, possibly even gestational diabetes.

Hsieh also cautions that anyone who thinks they might have diabetes should go to their doctor, not just rely on the Watch to tell them what’s going on, especially as those tested were already known to have diabetes or pre-diabetes.

 

Source: TechCrunch