Ringconn generating AHI?

[LJClark]

Ringconn is inviting their ring users to undertake “Sleep Apnea Monitoring”.  This involves a higher sampling level for three nights, along with a warning that the ring will be using power at a faster rate.  The “AHI” is based on SpO2 and pulse rate data.  Actual apnea and hypopnea data, such as collected from the PAP airway and which forms the basis for the AHI, is not collected. 
 
Allow me mention that I am a wee bit skeptical, so I decided to compare the results with my PAP machine (ResMed AirSense11 with F&P Evora full face mask and the Ninon 3150BT pulse oximeter (4-second sampling for SpO2, simultaneous data logging).  The comparison took place on the nights of 3, 4, and 5 December.  Each night’s data was loaded into OSCAR. 
 
The readings -- ResMed & OSCAR / Ringconn: 
AHI:    
3rd -- 3.99 / 0.4 
4th -- 4.33 / 0.7 
5th -- 4.51 / 0.6 
SpO2 Drops (SD):
3rd -- 5 / 0 
4th -- 9 / 3 
5th -- 3 / 5 
Events (CA, OA, HP, & UA from the PAP, Unknown from Ringconn): 
3rd – 25 / 2
4th – 22 / 3
5th – 26 / 3  
 
I’m unsure what Ringconn’s Gen 2 sampling rates are.  With their Gen 1 rings SpO2 was 5 minutes and 30 minutes.  The Sleep Apnea stuff was added with the introduction of their Gen 2 rings a few months ago, and especially the two most recent app updates.  A moderator on the Ringconn Facebook page mentioned 1 second sampling, but I’m not sure that wouldn’t drain the tiny battery in those rings. 
 
Ringconn has reduced the amount of data available to users and also stepped up their emphasis on averages for numeric information and more use of curve smoothing for chart displays.  For the sleep apnea assessment, you can look at one chart and, moving your finger along the plot, see the results at 3-minute intervals.  Of course, most of us understand how much can happen during 3 minutes. 
 
What worries me is how people can misinterpret the Ringconn information.  I’m not sure it is accurate and complete enough to encourage someone to consult with their health care provider.  One of the Facebook group members declared that his mild sleep apnea might be in remission based on the Ringconn results.

[Sleeprider]

Interesting. Sounds like RingConn is prone to significantly under-count events. Its appeal is a very small and techy looking form factor. The device is basically for bio-feedback and no therapy or diagnostic claims are made by the company, they leave that to their customer testimonials.

[LJClark]

Interesting. Sounds like RingConn is prone to significantly under-count events. Its appeal is a very small and techy looking form factor. The device is basically for bio-feedback and no therapy or diagnostic claims are made by the company, they leave that to their customer testimonials.

I know that there may be some slightly different opinions regarding the AHI "scores" (use of weighted sums) and how some clinicians use the AHI in developing individual patient therapies.  But are you comfortable with a company assigning an AHI to a data collection process which is not capable of collecting the data we generally accept as necessary for that AHI?

Ringconn is seriously trying to promote the ring's accuracy for SpO2, stating that it is generally as good as medical grade instruments.  (A photo in one of their papers' summary showed a study subject wearing what appears to be a Nonin 3150 -- capable of 1, 2, and 4 second sampling.)  "Accuracy" is how you choose to define it, but in our case, the ability to quantify events seems to be just a little bit important.

[Sleeprider]

Most sleep doctors and clinics won't recognize a CPAP device as accurate for AHI due to clinical criteria for classifying events with polysomnongraphy. It's pretty obvious a device with a direct channel to respiratory flow data is better than one that infers events from changes in SpO2 with doubtful accuracy. I'm not in the device approval business,a nd while ringconn may have some papers, I don't think they have an FDA medical device approval. I have no interest in the device, and see it as being less proficient than most smart or fitness watches

The RingConn claims can be found on an article on their site "Understanding Sleep Apnea and How RingConn Can Help". In part they say"

RingConn Gen 2 offers a unique solution for those concerned about Sleep Apnea. With 24/7 monitoring, RingConn tracks your sleep stages and provides real-time data with medical-grade accuracy. Here’s how we help:

90.7% detection accuracy:

Reliable detection of potential Sleep Apnea episodes.

Nighttime blood oxygen tracking:

Continuous SpO2 monitoring ensures you stay informed of any drops that could indicate Sleep Apnea.

AI Health Coach personalized insights:

The RingConn app analyzes your data with the AHC feature and offers personalized advice to improve your sleep quality (coming soon).

I have been unable to find the so-called scientific studies on sleep apnea that support the claims of a 90.7% accuracy. They are on a lot of trade publications announcing their breakthroughts, but none of the actual studies are cited or referenced.