One of the many problems with sleep apnea treatment, is that we're treating AHI. AHI is an imperfect proxy for all the things that really happen on account of sleep apnea. Not breathing, underbreathing, micro arousals, clustering of apneas etc. Things like UARS can go undetected nearly forever.
You can have an AHI of 1 and still get zero REM sleep because every time you start going into REM, a respiratory event knocks you out of it.
Apneas and Hypopneas have to last for 10 seconds or more to be included in the AHI count. You can literally have hundreds of respiratory events and have an AHI of zero, but you're still going to feel pretty bleedin' awful.
There is also an RDI, Respiratory Disturbance Index, which includes the AHI events but also includes arousals caused by breathing issues. It is typically much higher than AHI in un(der)treated patients. It's also imperfect, and it's harder to measure.
Then, of course, we come to the CPAP machine reported AHI, which seems to be what insurance and doctors are interested in. In short: the number is garbage. These machines are absolutely TERRIBLE at correctly identifying events. Using those numbers as the basis for any kind of decision is awful.
Maybe one day we'll get CPAP devices with EEG devices in the head strap and with unobtrusive SpO2 sensors that actually work. Maybe then we can start reporting useful numbers which can actually be used in support of treatment.
Until then, it's very much possible to keep feeling like shit even though your sleep apnea is allegedly "under control".