10/28/2025
Who Apple Watch Is Really Alerting to Undiagnosed Sleep Apnea
Like many sleep medicine stakeholders, I felt the stirrings of something momentous when first Samsung, soon followed by Apple, received the OK from the Food and Drug Administration (FDA) to market their over-the-counter devices that assess risk for sleep apnea in the United States. Hordes of the 23 million-plus people with undiagnosed moderate to severe obstructive sleep apnea (OSA) would effortlessly be alerted to their potential of harbouring the sleep disorder, I thought.
I have since tempered my expectations of what the risk detection features will mean for consumers who are unaware or in denial of OSA symptoms. That’s mostly because I learned neither of these smartwatch/health app combinations turns on the alert mechanism automatically.
The Samsung feature “is intended for on-demand use,” says Jeffrey Kim, a senior product manager at Samsung Electronics America. It monitors, on request, two nights of oxygen sensor data at a time. The accelerometer sensor-based Apple Watch feature runs opportunistically in the background, but only after the user sets up sleep tracking and passes checks for the FDA’s intended user criteria.
But I have high expectations for another group impacted by these consumer smartwatch sleep apnea risk alerts: sleep medicine-referring physicians.
The biggest impact will be on the doctors—from primary care physicians to cardiologists to gynecologists and more—who had been dismissing their patients’ sleep complaints, prescribing hypnotics without a full sleep evaluation, or referring patients to specialties like psychiatry for “depression”-related sleepiness.
Simultaneously, the alerts—and the evidence, like the exportable PDF Apple provides to users flagged for sleep apnea risk—will lend people the confidence to approach physicians about their sleep. And since the likelihood of a physician making a referral for OSA evaluation increased by a factor of nearly 10 when patients specifically inquired about OSA, at least according to one study’s findings,1 these alerts are poised to make a huge difference in OSA diagnosis.
OSA Screening: Benefits Versus Harms
The smartwatch features appear to work as well as (and possibly even better than) existing screeners like the Epworth Sleepiness Scale. Samsung’s sensitivity is 82.7% and modified specificity is 91.1%; Apple’s sensitivity is 66.3% and specificity is 98.5%. Both were designed with low-ish sensitivities to minimize false positives. Both were tested on large numbers of people across the spectrum of sleep apnea severities, including those with normal breathing. Caveat: This is based on their FDA submissions, not on peer-reviewed studies.
But, curiously, with these FDA nods, the country has effectively opted in to a grand screening experiment without knowing if flagging people at scale will have benefits that outweigh harms.
The US Preventive Services Task Force has researched the potential of OSA screening on asymptomatic adults numerous times, most recently in 2022, and come up empty-handed.2 According to the task force, zero studies have been published on the harms of screening the general population for OSA or on the financial costs of the OSA diagnostic pathway, and only a few have detailed the potential physical harms of OSA therapy.
For its part, Samsung, in its De Novo documents, details risks ranging from false positives and false negatives to misinterpretation of device output to electrical shocks. It also lists the methods—like specific testing and labeling—in place to mitigate the risks.
I don’t fault technology companies for pressing forward with features that make their smartwatches more appealing, but I do worry about unintended consequences.
“In the patients who get a positive screen, I think that that is a beneficial thing for them,” says Steven Holfinger, MD, MS, vice chair for the American Academy of Sleep Medicine (AASM)’s emerging technologies committee. “I also think it’s yet to be seen if the net effect is truly positive—because there are a lot of things that could potentially go wrong. For example, we don’t know how many patients who get a negative screen are not going to seek care when otherwise they should.”
Also, it’s not exactly the general population that wears smartwatches. It’s skewed toward Hispanic, Gen-Z, college-educated women with full-time employment.3
“If you’re viewing a strategy of population screening through the lens of ‘we’re going to allow this to kind of naturally happen this way where patients screen themselves—but only the patients who can afford it can do so,’ then that’s not really the optimal strategy to screen a population, I would imagine,” Holfinger says.
May 5, 2025 By Sree Roy