Apple Watch Able to Detect Abnormal Heart Rhythm With 97% AccuracyMay 12, 2017
The Apple Watch’s built-in heart rate monitor is 97 percent accurate when detecting the most common form of an abnormal heart rhythm when paired with an algorithm to sort through the data, according to a new study conducted by the University of California, San Francisco and the team behind the Cardiogram app (via TechCrunch).
There were 6,158 participants in the study, all of whom used the Cardiogram app on the Apple Watch to monitor their heart rate. Most were known to have normal EKG readings, but 200 suffer from paroxysmal atrial fibrillation (an occasional irregular heartbeat).
Data from these participants, along with data taken from normal Cardiogram users, was used to build a neural network that could recognize the abnormal heart rhythms solely from data collected by the Apple Watch. As of today, Cardiogram says its algorithm can almost always successfully determine when a patient is in atrial fibrillation.
In order to validate the model, we obtained gold-standard labels of atrial fibrillation from cardioversions. In a cardioversion, a patient experiencing atrial fibrillation is converted back to normal sinus rhythm, either chemically or with a shock to the heart. 51 patients at UCSF agreed to wear an Apple Watch during their cardioversion.
We obtained heart rate samples before the procedure, when the patient was in atrial fibrillation, and after, when patient’s heart was restored to a normal rhythm. On this validation set, our model performed with an AUC of 0.97, beating existing methods.
Cardiogram is a startup that’s aiming to garner more information from the data collected by the Apple Watch. The study, which Cardiogram has raised funding for, started in March of 2016 and will continue as UCSF and Cardiogram work to refine the neural network and detect other conditions beyond atrial fibrillation.
Cardiogram plans to put in additional work before using its algorithm to start notifying Cardiogram users of arrhythmias. The company needs to conduct further testing to make sure the algorithm works in a variety of conditions and it needs to work on scaling it so it can be used continuously by all Cardiogram users.