One Algorithm's RERA another's Hypopnea?

General Discussion on any topic relating to CPAP and/or Sleep Apnea.
User avatar
BP
Posts: 199
Joined: Fri Aug 19, 2005 4:09 pm
Location: Atlanta, GA

One Algorithm's RERA another's Hypopnea?

Post by BP » Fri Oct 21, 2022 10:18 am

I have been using PR machines for almost two decades now. As I mentioned in an earlier post I am working through transitioning to Resmed. I seem to have small 10 minute or so clusters of events 3 to 4 times per night.

These clusters were mostly identified as RERAs on my System One 60, but it appears that Resmed "sees" them as Hypopneas. My AHI was almost always below 1 when using my System One, but is it possible that the algorithmic identification was incorrect?

Here are some examples below from my Oscar data. BTW, I'm not claiming that I am right, just trying to gain insight into how the algorithms differ. Since my AHI numbers were mostly under one using my System One I never really looked more closely. Now I'm thinking maybe I should have. While I have felt pretty good the last few years, I want to make sure I'm doing all I can to optimize my therapy.

Note in this PR image several REs that could be Hypopneas using Resmed's algorithm? Also, the red arrows also appear to be Hypopneas but are completely missed.
https://drive.google.com/file/d/1P3hNd8 ... sp=sharing

Note in this Resmed image that their algorithm seems to be more accurate? aggressive? at identifying Hypopneas?
https://drive.google.com/file/d/1bj-5NF ... sp=sharing

Any thoughts/suggestions would be greatly appreciated.

Thanks, BP

User avatar
lazarus
Posts: 1397
Joined: Sat Mar 12, 2022 3:21 pm

Re: One Algorithm's RERA another's Hypopnea?

Post by lazarus » Fri Oct 21, 2022 10:33 am

The estimated event numbers from home treatment machines are meant to be used for trending, to see whether a change made in pressures or other therapy-relevant settings cause the numbers to increase or decrease over time. Therefore, the proprietary labels attached to what is flagged is not as crucial as might be the case during an actual sleep study. The point is to choose the best settings for getting reported numbers as low as possible (but no lower than that :wink: ) over time.

APAP algorithms may use criteria involving breathing shapes that are less than "events" for choosing pressures. And any event reports involving "effort" in treatment machines are not based on belts anyway. But the trending numbers are still very valuable, regardless of the specifics of how "events" are labeled, reported, or used by algorithms.

User avatar
Pugsy
Posts: 64998
Joined: Thu May 14, 2009 9:31 am
Location: Missouri, USA

Re: One Algorithm's RERA another's Hypopnea?

Post by Pugsy » Fri Oct 21, 2022 10:41 am

Are you aware of the different criteria for naming of a flagged event that must be met to earn a flag?

Each brand has their own criteria so it may not be so much one brand "missing" something as it is simply a difference in criteria that needs to be met to earn a flag.

It might be as little as a slight difference in the amount of reduced air flow either just from normal night to night variations or the brand's own criteria.
Like maybe a 50% reduction in air flow vs 40% reduction....or 30 % vs 40 %.

RERAs are a totally different animal and aren't necessarily hyponeas mislabeled.
RERAs are the machine making an educated guess that you experienced an arousal from sleep because of something happening in your airway that makes your breathing look like someone's breathing who has arousals from airway events.
The machine's can't be sure though because to have an arousal you have to have sleep and the machine is unable to know for sure if you were asleep or not.

_________________
Machine: AirCurve™ 10 VAuto BiLevel Machine with HumidAir™ Heated Humidifier
Additional Comments: Mask Bleep Eclipse https://bleepsleep.com/the-eclipse/
I may have to RISE but I refuse to SHINE.

User avatar
Rubicon
Posts: 1756
Joined: Sun Feb 20, 2022 6:59 am

Re: One Algorithm's RERA another's Hypopnea?

Post by Rubicon » Fri Oct 21, 2022 10:45 am

I think it's because in PR
The hypopnea detection algorithm required the presence of two recovery breaths that nominally were at least 75% to 80% of the baseline airflow.
Freeze this moment a little bit longer.
Make each sensation a little bit stronger.
Experience slips away.

User avatar
BP
Posts: 199
Joined: Fri Aug 19, 2005 4:09 pm
Location: Atlanta, GA

Re: One Algorithm's RERA another's Hypopnea?

Post by BP » Fri Oct 21, 2022 10:57 am

Pugsy wrote:
Fri Oct 21, 2022 10:41 am
Are you aware of the different criteria for naming of a flagged event that must be met to earn a flag?
No definitely, I am not familiar with the details of each manufacturer's criteria. And I take your point.

I guess I'm trying understand if my therapy was actually equivalent between the two machines, even though the AS11 AHI is 1.98 and the System One was only 0.62, just to use examples. Do I keep trying to drive the numbers down or not?

User avatar
Pugsy
Posts: 64998
Joined: Thu May 14, 2009 9:31 am
Location: Missouri, USA

Re: One Algorithm's RERA another's Hypopnea?

Post by Pugsy » Fri Oct 21, 2022 11:26 am

BP wrote:
Fri Oct 21, 2022 10:57 am
though the AS11 AHI is 1.98 and the System One was only 0.62, just to use examples. Do I keep trying to drive the numbers down or not?
I don't worry about minor differences in AHI myself because 90% of the time any extra flagged events are more related to arousal breathing or SWJ (sleep/wake/junk) than they are real asleep airway related collapses but if you aren't sleeping well or feeling decent then it might be worth a little bit of tweaking to see if it changes either the numbers or how you sleep and how you feel.

There comes a point though where chasing lower numbers is counter productive.

Now chasing how I sleep or how I feel during the day.....that's worth evaluating IMHO but not so much the numbers. YMMV.

_________________
Machine: AirCurve™ 10 VAuto BiLevel Machine with HumidAir™ Heated Humidifier
Additional Comments: Mask Bleep Eclipse https://bleepsleep.com/the-eclipse/
I may have to RISE but I refuse to SHINE.

User avatar
BP
Posts: 199
Joined: Fri Aug 19, 2005 4:09 pm
Location: Atlanta, GA

Re: One Algorithm's RERA another's Hypopnea?

Post by BP » Fri Oct 21, 2022 11:36 am

Pugsy wrote:
Fri Oct 21, 2022 11:26 am
I don't worry about minor differences in AHI myself because 90% of the time any extra flagged events are more related to arousal breathing or SWJ (sleep/wake/junk) than they are real asleep airway related collapses but if you aren't sleeping well or feeling decent then it might be worth a little bit of tweaking to see if it changes either the numbers or how you sleep and how you feel.
That makes good sense.

This whole processed has been a mess, and I am probably obsessing more that I should. Like many others, I was really happy with my System One, then PR threw a wrench into it. Now I am struggling with insomnia, and adjusting to the AS11. It's strange but the quietness of the machine is bothering me, not sure what that is all about, but I really loved the woosh sound of the System One. Kind of like a built in white noise generator.

User avatar
ozij
Posts: 10444
Joined: Fri Mar 18, 2005 11:52 pm

Re: One Algorithm's RERA another's Hypopnea?

Post by ozij » Fri Oct 21, 2022 8:46 pm

lazarus wrote:
Fri Oct 21, 2022 10:33 am
The point is to choose the best settings for getting reported numbers as low as possible (but no lower than that :wink: ) over time.
Suggested re-phrase:
The point is to choose the best settings for getting reported numbers as low as necessary to give you refreshing sleep (but no lower than that :wink: ) over time.

_________________
Mask: AirFit™ P10 Nasal Pillow CPAP Mask with Headgear
Additional Comments: Machine: Resmed AirSense10 for Her with Climateline heated hose ; alternating masks.
And now here is my secret, a very simple secret; it is only with the heart that one can see rightly, what is essential is invisible to the eye.
Antoine de Saint-Exupery

Good advice is compromised by missing data
Forum member Dog Slobber Nov. 2023

User avatar
lazarus
Posts: 1397
Joined: Sat Mar 12, 2022 3:21 pm

Re: One Algorithm's RERA another's Hypopnea?

Post by lazarus » Fri Oct 21, 2022 9:27 pm

ozij wrote:
Fri Oct 21, 2022 8:46 pm
lazarus wrote:
Fri Oct 21, 2022 10:33 am
The point is to choose the best settings for getting reported numbers as low as possible (but no lower than that :wink: ) over time.
Suggested re-phrase:
The point is to choose the best settings for getting reported numbers as low as necessary to give you refreshing sleep (but no lower than that :wink: ) over time.
Sounds good.

Fully agree that the subjective self-assessment can be a significant adjunct to the reported data for those sensitive to feeling a difference.

But, for me, if a change to my therapy caused my trending numbers to start climbing, I would likely be suspect of the change, myself, whether I immediately felt a difference or not.

User avatar
chunkyfrog
Posts: 34545
Joined: Mon Jul 12, 2010 5:10 pm
Location: Nowhere special--this year in particular.

Re: One Algorithm's RERA another's Hypopnea?

Post by chunkyfrog » Fri Oct 21, 2022 9:41 pm

I'm so grateful I don't need to understand respironics craziness.

_________________
Mask: AirFit™ P10 For Her Nasal Pillow CPAP Mask with Headgear
Additional Comments: Airsense 10 Autoset for Her