Worked a 3 year old a couple weeks ago with an unknown downtime in a pool (up to 15 minutes). He was dead from the outset (initial rhythm was asystole), and he stayed dead (much warmer water in this case), but you still try...
This is especially true for cold-water drownings. Downtimes even longer than 42 minutes with full recoveries are not unheard of. You're not dead until you're warm and dead.
It's a bit of both... I have an 'obvious death' protocol that can apply if someone is, well... obviously dead. If they are not obviously dead, and do not have a DNR (or a 'MOLST' in New York) then we'll begin resuscitative efforts (using either ACLS[1] or PALS[2] guidelines, as appropriate). If the patient's heart has any sort of electrical activity going on, we'll generally spend 15-20 minutes trying to get it working properly again before calling a medical control physician to get permission to stop efforts and pronounce death.
If there is no electrical activity, then we'll go through a couple 'rounds' of ACLS, and it there is no response, we'll just pronounce them. Very rarely do we transport an adult in cardiac arrest. CPR in a moving ambulance is very difficult to perform adequately, and in most cases we would just be bringing them to a facility that would be providing the same level of care that we are in the field (albeit with more hands available to help, but the tradeoff in time and poor quality CPR isn't worth it).
The only exception to this are cases of traumatic arrests. If something is physically broken, no amount of CPR, drugs, or electricity are going to help, and their only hope is bright lights and cold steel (in a surgeon's hand). Needless to say, the survival rates for out of hospital traumatic arrests are _vanishingly_ small.
Kids are different though... They're easier to move, and easier to do CPR on. This means in most cases a pediatric arrest is a 'scoop and run', with any interventions being performed on the way in. This is in part because the resiliency of kids means they are a _little_ more tolerant of the time, and scene's with dead kids tend to be extremely chaotic. It can be very difficult to concentrate with an (understandably) distraught parent yelling at you.
Worked a 3 year old a couple weeks ago with an unknown downtime in a pool (up to 15 minutes). He was dead from the outset (initial rhythm was asystole), and he stayed dead (much warmer water in this case), but you still try...
This is especially true for cold-water drownings. Downtimes even longer than 42 minutes with full recoveries are not unheard of. You're not dead until you're warm and dead.