There is a tradeoff that is used to justify the long hours. On the one hand, having a well rested doctor is obviously good for them to be making good decisions. But on the other hand, patient handoffs are dangerous. The more times you change the person responsible for a patient, the longer the game of telephone you are playing with their care. This has been measured as being bad for patient outcomes.
Now, that doesn't seem to justify the fact that long hour shifts are placed so close together. It seems like you could give doctors a longer break in between shifts than they have. Residents have the worst of it. The attending actually do get a fair amount of time on/off. Residents already work a lot less than they did 50 years ago, some think that their training should be extended to cover the loss of density.
Maybe OT, but pilots have a lot of time off between shifts and it's starting to come to light that their depression rates are much higher than anyone is comfortable admitting. This may from a different source though, since flying is very much about precise repetition and less about complex decision making (ADM is hard, but not the same level of mental stress that doctors endure), you eventually realize that you're a very highly trained bus driver.
It's way more complex than that though. The regulations exist because companies were pushing pilots to do more than they could handle, leading to a number of high profile crashes. Yes, the pilots are incentivized to manage fatigue but it's like any other profession, it's easy to get complacent.
I hate when this gets brought up, because it inherently implies that we can't improve them. Everyone talks about increases in handoffs causing increases in medical errors. I think handoffs have a long way to go, and we need to better utilize technology to help in this (ie make better EMRs).
The overworked doctor is just as bad IMO. I've been there on solo 28 hour calls going on my 11th admission. In the morning I'm next to useless and my handoff to that team was less than stellar.
I had a friend who was a QA process engineer at a regional hospital. He identified a significant source of errors - every morning there was a 'double-handoff' as the doctors handed off to nurses, who ended their shift and handed off to the next shift. The problem would be solved by moving hand-off time by a mere 15 minutes, but neither the doctors' group nor the nurses' group would budge on the matter.
I've heard a few other similar stories from him as well. Doctors have immense political power; if hospitals are grinding them to dust, it's because doctors as a group are letting them (the good old 'seniors don't care that juniors are getting crushed' problem). From my own limited experience working with them as a neuro tech, doctors will close ranks quickly against outside forces, but plenty will sell each other out within the profession. For every haggard ED doctor, there's a specialist somewhere making cushy deals with the administration.
Now, that doesn't seem to justify the fact that long hour shifts are placed so close together. It seems like you could give doctors a longer break in between shifts than they have. Residents have the worst of it. The attending actually do get a fair amount of time on/off. Residents already work a lot less than they did 50 years ago, some think that their training should be extended to cover the loss of density.