EPR setting correlation with sleep stages

[sleepyp]

Has anybody observed this?

No that my set-up has settled and my average AHI is 0.4-1.3 every night, I've started exploring more of the finer details, I'm an engineer and very inquisitive. I like to understand things and really optimise systems.
For me EPR = 0 produces a better Deep/REM split, more like 12%/26%. My hyps are higher and AHI is closer to 1.3, so technically worse from an OSA perspective, but better from a sleep structure perspective.
EPR = 3 produces more like 6%/35%. Here my hyps are often zero and AHI is closer to 0.4 average.
These are averages, obviously there's nightly variances but the trend is observable.
Both scenarios are way better than no CPAP.
I do feel better after more deep sleep, so I'm leaning towards no EPR even though the AHI is not as good, and hypopneas happen about 1 per hour.

Has anyone else observed this? Is there a logical reason why it would be the case or am I imagining it.
I'm using Fitbit and Samsung Galaxy Watch to measure sleep stages, I believe them to be about 70% accurate, and quite consistent and so good enough to establish a trend over a week or two.

[multicast]

I strongly doubt that those gadgets you mention deliver reliable data about your sleep stages. Would that be so "easy" the DME would surely use them as standard.

Mike

[sleepyp]

During trials they have been shown to have accuracy of about 70-75%, not good enough for a medical diagnostic tool but enough to build a reliable trend over time.
It does roughly correlate to perceived energy levels. 10% deep sleep might be somewhere between 8 and 12, but it’s not 20% for example. Many people don’t wear them correctly, but when worn properly they’ve been proven pretty accurate and consistent.

I personally wouldn’t place much value n a DME’s opinion on anything other than selling CPAP equipment, and even then they’re not that knowledgeable.