This recent sleep study in Biological Psychology caught my eye: “Effects of pre-sleep simulated on-call instructions on subsequent sleep“. These guys studied people in a sleep laboratory and told them they would need to respond to a noise at some random time during the night (like a doctor who is on-call would have to). So the subjects went to sleep expecting that they would have to wake up sometime in the middle of the night to respond to a “call”. However, unbeknownst to the subjects, the researchers never planned on waking them with any noise, and let them sleep a normal night through.
You would probably guess that these people might subjectively have a worse night’s sleep. But the cool thing is that the researchers objectively noticed this, as well. The people expecting to be woken up had a longer time to fall asleep, woke up more times after falling asleep, and generally had lower sleep efficiency (the number of minutes of sleep divided by the number of minutes in bed).
Not really what you want for a doctor that might have to function the next day (and by function I mean do stuff like open heart surgery).