How Apple’s Sleep Apnea Feature Works on Apple Watch – Video

The Apple watch is getting a new health feature to detect signs of sleep apnea, a potentially serious condition where you stop breathing momentarily during the night. Let’s dig into how all of this works. It’s estimated that close to 1 billion people around the world between the ages of 3069 may have obstructive sleep apnoea left untreated. It could increase the risk of cardiovascular or metabolic health problems. Normally sleep apnoea is detected on a polysomnograph, otherwise known as a sleep study. This could be in a sleep clinic or there are at home tests that you can do. A sleep study measures your brain activity, heart rate movement and blood oxygen among other metrics to assess your sleep. Now, you can get some of this data from your wrist just by wearing a smartwatch to bed. The Apple watch uses the accelerometer and machine learning to detect interruptions in your breathing pattern. By monitoring small movements on the wrist. You wear the watch to sleep for at least 10 nights in a 30 day period. And it measures a new metric that Apple calls breathing disturbances. I spoke with Apple to dive into more about how this all works and how it was tested. My name is Matt Bianchi. I’m a neurologist and sleep disorders specialist and I work in a research unit called Health Technologies at Apple. My name is Deidre Cald Beck and I’m the senior director of Apple Watch and Health product marketing. Each night you wear your watch, we’re tracking that breathing sensor signal through accelerometer. We quantify that number and you get to see your breathing disturbance value and whether it’s elevated or not elevated each night in the health app. Now, even a person without sleep, apnea can have some interruptions in breathing at night. Sometimes it’s a little too much alcohol. Sometimes it’s, you’re sleeping more on your back than on your side. We want to actually tackle that night to night ability issue, head on by having the notification logic. Look at 30 days of data. So we get a holistic view of what’s going on. Can you elaborate a bit more on what the specific wrist movements are when that differs essentially from regular breathing to a breathing, that might indicate sleep apnoea. Imagine that when you breathe, it’s an oscillation of movement in your chest. That oscillation in and out is actually seen through apple watch accelerometer at the wrist. And so when you have an interruption in breathing, where you’re either taking more shallow breaths or you stop breathing for 20 or 30 seconds that can be read out and detected through machine learning algorithms in the wrist. How does the Apple watch determine the difference between someone just tossing and turning in bed, having a restless night compared to an actual breathing disturbance? So, one of the ways we make sure that we don’t make mistakes or sort of misconstrue movements like that is to collect an incredible amount of ground truth data across a huge variety of people sleeping in the lab in a sleep center, sleeping at home in their natural environment. In fact, our internal work involved over 11,000 nights of ground through recordings alongside watch to train these algorithms. We partnered with uh academic medical centers, hospital systems uh through uh patient care settings as well as clinical research organizations, which are private uh uh research organizations to uh recruit patients. And then there were was an additional 1500 nights collected uh through CRO for the validation study submitted to the FDA um and will be submitted to other regulatory bodies as well. What Matt is describing of course is is what we used for um development and and and the clinical validation. But um we do have a little sleep lab on campus and you know, of course to do this continued prototyping and iteration of of all of our features and including, you know, the continued enhancements of our features. Uh for sleep tracking and sleep stages, sleep apnea notifications are coming to the series 10, series nine and ultra two for people. 18 years and older in over 100 and 50 countries. So you may be wondering if this sleep apnea feature relies on the A Ramada. Why didn’t Apple roll it out on earlier Apple watches for this feature to meet our performance objectives? We needed that processing power that we introduced with the S nine SI um last year. So and that same processing power is in the S 10 sip and in ultra two. So it’s the ability to use the accelerometer data process that process that data on device and ensure that we get, you know, uh get accuracy and ensure we meet our performance objectives. And so we really needed that combination sleep apnea notifications are now available in the US following FDA clearance. You’ll go through a quick set up process and then wear the watch to sleep after each night of sleep, you’ll see the breathing disturbances. Metrics show up in the health app listed either as elevated or not elevated. The Apple watch is the second smart watch to get sleep apnea notifications following the Samsung Galaxy watch. Earlier in 2024 Apple takes a different approach to Samsung’s sleep apnea feature because the Galaxy watch only needs two nights of sleep data versus 30 days on the Apple watch. And Samsung’s watches use blood oxygen as part of this measurement. Regardless of methodology. It’s really important to note that no smartwatch is a to give a medical diagnosis of sleep apnea or any other condition. The idea is that you take the data from your watch, which in the case of Apple’s health app, lets you export it as a PDF. Then you present it to your doctor. Thanks to Deirdre and Matt from Apple for chatting with us about all things Apple watch. And thanks to the Clancy in San Francisco for letting us film and thank you for watching.