There are a number of factors in play here. First and foremost, for most people 30-hours is a lot (the limit is 28-hours however), residents included. That being said, let us make an analogy to running a marathon. If you take your average person off the street, and ask them to run a marathon, they would not be ready. Marathon runners, on the other hand, have trained for the distance and are psychological and physiologically ready for the event. Most people off the street would not function well after 28-hours. Likewise, most residents don't either. Even so, residents in training, at least in neurosurgery, are generally familiar with getting less sleep, and/or taking call for 28-hours. Is it safe? It seems so, given the system of checks in place. Is there a better system? At the moment, that appears to be no.
To answer your question regarding a shortage of neurosurgeons. That is tough to measure. From what I can tell, there does not seem to be a protracted wait for a neurosurgeon, or people going untreated. Would the number of hours be less if we had more neurosurgeons? Most likely. Would those neurosurgeons be trained as well? Undoubtedly not. The RRC has oversight over resident training and has strict requirements for the experiences of each resident with respect to volume and breadth of cases seen. Diluting the volume (and breadth) across more residents would result in lower quality training for each resident.
Finally, to address human error. Unfortunately, in the world of humans dealing with and treating other humans, human error is a factor. There have been many studies looking at duty hours and whether they play a role in these events, notably prior to and after the switch from 28-hours, to a 16-hour limit for interns, and back. My assessment of this is that most residents tend to be able to function within their range of practice at the end of their call period. Most near-misses, misses, or never events that occur, are due to a series of errors and involve multiple people.
