The first thing we care about is the patients. Most studies that have looked at longer hours vs more handoffs have shown equipoise in terms of death and serious adverse events. So, in that setting, most of us feel comfortable to think about what we want for ourselves as trainees and educators.

From the perspective of a trainee, on the rotations where duty hours even become a question, I was working 60-80 hour weeks. You can achieve that in a number of ways. One way is to work 5-6 days per week of 12-14 hour shifts. This sounds reasonable until you realize, a few months in, that you are at work from 7a-7-9p (assuming you're working daytime shifts) 6 days a week. It's monotonous and exhausting.

You can achieve the same by mixing some 24h days into the mix. During my training, this could only happen starting in our second year and beyond. Suddenly, I had several days with ample free time to buy groceries, do laundry, think about the world around me.

Medical training is extremely long and time intensive. Partitioning time into 24 hour chunks can surprisingly be preferable when 80 hour weeks are the rule (when safe for patients).

But if you do 80 hour weeks and 24 hour shifts, then you will end up doing Monday, Wednesday, Friday and Sunday, so you have two 24 hour shifts back to back, resulting in a 48 hour shift? Is that correct? Or do you do 4 days one week and 3 days the other? Do you get to nap during your shift?

The rule is no more than 80 hours per week averaged over a 4 week period, so some weeks you pull more, others less. Not every block requires 80 hours/week. Generally though, 24h shifts are placed into your schedule ("call day"), and other days you work a usual/shorter shift.

Did they study the effect of sleep deprivation on the quality of handoffs?

The FIRST study found that increased scheduling flexibility (including allowing longer shifts) was not worse than forcing shorter shifts (in terms of patient outcomes).

http://www.nejm.org/doi/full/10.1056/NEJMoa1515724