I've carried one for a couple of years for EMS cases. The basic answer is no, it probably won't help you predict AMS. I've seen a huge range of O2 saturation and haven't really been able to spot a trend. By the time someone is symptomatic, that's the clue.
There's some help, maybe, in comparing a group of people all going at the same rate. I suppose someone who's out of range for the group might be at higher risk.
I now use it mostly to give me a good pulse and the O2 sat as another bit of data in deciding how serious someone is. But I've never really used it to determine if someone should be flown out or stay -- I stick to signs and symptoms (how they look, how they report feeling, trend -- getting better or getting worse, pulse, & respiration).
University Medical Center in Fresno was doing a small study on the Sierra Crest taking pulse/ox, but hasn't yet published, as far as I know. Don't know what they came up with.
George