Diagnosing Sleep Apnea from Your Wrist?
So, after much deliberation, I ordered an Apple Watch. I ordered it because I can never hear the phone ring when it is in my pocket. The advertisements tell me that the watch will blink and vibrate and yell at me when the phone rings, and hopefully that will be enough to get my attention.
But blinking and vibrating and yelling at old people is not all that the watch will do. It will remember your wife’s birthday and your daughter’s phone number. It will tell you how to get to the hardware store. It will tell you when to get up and move around so that you don’t develop a deep venous thrombosis. And now there is an initial report that shows the Apple Watch can do even more.
A recent abstract from the University of California, San Francisco Health Group, working in collaboration with Cardiogram, Inc. (an app developer), provides preliminary data on the use of off-the-shelf wearables for the diagnosis of hypertension and (wait for it …) sleep apnea.1 It uses longitudinal heart rate variability and activity data to predict the diagnosis. The data is analyzed using a deep neural network, which is another way of saying artificial intelligence.
For sleep apnea, the diagnostic sensitivity was 90.4% and the specificity was 59.8% using the best cutoff levels. This compares favorably with home sleep testing devices, including those with effort and flow signals comparable to laboratory testing.2
Are you skeptical? You should be. This is a single abstract and not a peer-reviewed publication. It comes from a single research center. It is rife with potential bias, given that the authors are all employees or consultants for the app developer.
But there were 6,115 active users providing 33,628 person-weeks of data. Of these, 30% of the participants were used to validate the predictions and 16.6% had sleep apnea.
Sleep apnea is not the only diagnosis subject to new technology. Another recent article reports that a physician testing an ultrasound device that plugs into his iPhone was able to diagnose his own cancer.
I’ve been saying for the past five years that wearable devices will soon be able to diagnose sleep apnea. Is the abstract an inadequate demonstration of the accuracy of the algorithm? Of course, it is. But the future is clearly moving in the direction of home testing for a variety of medical disorders, and I’m sure that sleep apnea will be included in that trend. The alternative of a robust laboratory sleep testing industry for the next decade is improbable, with a p < .05 as we say in the statistics class.
Given that we can predict the home testing future with reasonable certainty, what should you, as a sleep technologist, do to prepare? The AAST has been saying this for some time now: expand your skill set, get a higher degree and certification, prepare for complex patients with medical comorbidities and integrate as a valuable member of the sleep center.
The clock is ticking. But not your Apple Watch. It’s digital and it doesn’t tick. But soon it will diagnose your patient’s sleep apnea.
- Tison, G.H., Singh, A.C., Ohashi, D.A., Hsieh, J.T., Ballinger, B.M., Olgin, J.E., Marcus, G.M. & Pletcher, M.J. Cardiovascular risk stratification using off-the-shelf wearables and a multi-task deep learning algorhythm. Circulation 2017;136:A21042.
- Collop NA; Tracy SL; Kapur V; Mehra R; Kuhlmann D; Fleishman SA; Ojile JM. Obstructive sleep apnea devices for out-of-center (OOC) testing: technology evaluation. J Clin Sleep Med 2011;7(5):531-548.