One of the most challenging bits with that is going to be transmitting RF underwater. And an interesting thing to think about is whether or not such a device would end up causing even more accidents, when people start just assuming that "it's ok if I go past my limits, someone will save me"
Underwater RF is no trouble in this case. You have the signal backwards. As long as there is no problem, regularly transmit an identity beacon. Loss of signal is the alert.
How does the regular beacon transmit if it's underwater? Someone standing in the shallow end of the pool or treading water in the deep end may pretty easily keep their wrist underwater for minutes at a time.
There's two things that such a device would need to have: quite low false positives (so that the lifeguard isn't distracted by false alarms - leading to a boy-who-cried-wolf scenario), and vanishingly small false negatives (the first time a kid dies because the device didn't work, it'll be a bad day). In-water RF would be very hard to do to make those two conditions true.
Edit: Don't get me wrong, I think it's a fabulous idea. Wearing my EE hat though, I haven't thought of a good way to do it that would work reliably enough for a life-saving device.
