Studying sleep: A DIY diagnostic approach

Article By : Brian Dipert

Can a consumer device deliver an acceptable approximation of the results of a doctor-directed sleep quality analysis study?

In my previous post, I discussed how my doctor and I discovered I might have sleep apnea (a sleep disorder characterized by pauses in breathing or periods of shallow breathing during sleep), the multiple testing steps I undertook to confirm the initial suspicion, and the equipment I obtained to alleviate the condition. In reading it, you might have noticed how lengthy and complicated the process from diagnosis to remedy was, as well as how costly it was both to my healthcare insurance provider and myself. If you’ve taken to heart my “strong encouragement to follow in my footsteps and get your sleep quality checked out if at all in doubt,” but you’d like to try testing yourself out first prior to engaging with a medical professional for official analysis, a recent consumer-tailored equipment experience I had may be instructive.

While waiting for my continuous positive airway pressure (CPAP) machine to arrive, I ironically received a review-interest-inquiry email from the PR representative for a product called EverSleep, whose company ended up being less than a half hour’s drive away from me. Their product, which started out as an Indiegogo project, is conceptually akin to the first sleep monitoring device I’d previously gotten from the doctor, which solely leveraged a pulse oximeter to assess respiration and other vital signs. However, whereas the device I’d previously used had included embedded storage for data logging, in this case EverSleep leverages Bluetooth Low Energy connectivity to an app running on an Android- or iOS-based mobile device, where the data is archived and analyzed.

I’ll begin with a few unboxing shots:

Along with the unit itself were a few other goodies in the box including a roll of tape to keep the pulse oximeter sensor in place on the fingertip, a USB-to-micro USB charging cable, and an AC-to-USB charger:

The USB connector on the charger was misaligned inside, but still functioned:

Three pieces of business card-sized literature, along with a one-page instruction sheet, completed the suite.

Here’s a backside shot. Perhaps obviously there’s no photoplethysmography optical heart rate sensor as with a smartwatch or fitness tracker, since the fingertip-mounted sensors accomplish this and more … and equally perhaps obviously, there’s a 2G 12-bit accelerometer inside to measure motion during slumber.

And here it is in situ (minus the tape), alongside my Lenten reminder bracelet:

When the unit’s charging, its top- and bottom-located LEDs glow red and when fully charged, depending on its mode, they either steadily or intermittently blink green:

Now for the software side of the story. The initial setup process includes allowing the application access to your mobile device’s microphone so that, if it’s bedside next to you, it can listen for your snoring:

There’s also an upfront CYA qualifier:

The app is resolution-designed for smartphones, but I have it installed on my “Retina” iPad mini 2. As such, the low power mode setting mentioned in the configuration instructions isn’t offered on my particular iOS-based device (note, too, that I was instructed not to explicitly pair the two devices during initial setup):

At this point, it’s time to set up a user profile. It includes fields for your height, weight, gender, and bedtime, and questions about your breathing, anxiety, pets, and sleep habits. There are also sections on what the doctor has diagnosed you with and the medications and therapies you use.

Every night before using the device, you need to fill out a short survey that will tailor the results both to the day you’ve just had and your current condition. It asks about allergies, naps, using screens before bed, caffeine, alcohol, food, medications, and exercise.

At that point, pairing begins:

The pairing process is even smart enough to know if you haven’t charger-tethered your mobile device:

Following that, data recording commences. So how’d EverSleep do? Big-picture impressions first; across multiple nights of testing, it never disconnected from my iPad mini 2, and I never drained its 150 mAh lithium battery (I did recharge it every day). Two upfront oddities:

  • My wife started calling me “Rudolph the Red Nosed Reindeer” because of the fingertip sensor’s red glow; unlike some other pulse oximeters I’ve used, this one’s fairly translucent
  • There’s no way to manually turn it off, although it’ll auto-shutoff after around 1 hour of non-use

Now for the per-day data details. The first night, I intentionally slept with no oxygen assistance equipment whatsoever. To clarify, that’s 48 breathing events per hour, folks … but hey, I told you so, no snoring:

Night #2 I used my oxygen concentrator. My hourly breathing events were way down, but so too was my total sleep, so the statistical significance of the results was uncertain:

And for the third night of measurement, I used my CPAP for the first time. Hourly breathing events were up slightly from the previous night, but nowhere near where I’d been at first. Overall sleep duration and quality was also way up:

My conclusions? The CPAP is doing the sleep quality improvement job it’s been challenged with, and the EverSleep device is reasonably accurately measuring my sleep vital signs. Again, this isn’t a full-blown sleep study unit, such as the ResMed ApneaLink Air I used during the diagnosis portion of my situation, which measures respiration not only via pulse oximetry but also via chest oscillations and nasal cannula airflow patterns. But it also doesn’t require the significant cost and lengthy acquisition/use/analysis delays commensurate with a unit like that, either.

What, you might ask, about sleep trackers such as those built into latest-generation fitness trackers and also available via third-party smartwatch apps such as AutoSleep Tracker for Apple WatchOS and Sleep as Android for Google Wear OS and other Android-compatible wearables? They do the best they can with the data available to them, but it’s limited. Sure there’s an accelerometer to measure motion, along with a pulse rate monitor, but there’s no oxygen saturation measurement capability, for example. They can give a rough estimate of the percentage of total logged time that you’re in various sleep phases (along with being awake), but they can’t discern possible oxygen-deprived apnea situations, for example.

I’ll once again close with a strong encouragement to get your sleep quality checked out if at all in doubt, at minimum via a consumer device such as EverSleep, with results subsequently correlated in conjunction with professional medical assistance. And now I’ll once again turn the microphone over to you for your comments and questions. Sleep well!

Brian Dipert is Editor-in-Chief of the Embedded Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.

Related articles:

Leave a comment