Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

A friend runs a residency program and I've talked with them about it a few times. This is definitely getting better, but the number of hours still seems high to me. As a developer, I'm strongly against extended hours because I know how quickly the error rate creeps up. (And I probably wouldn't know that if I weren't doing TDD and pair programming, because the first thing that goes for me is ability to notice my poor performance.) My basic question was: shouldn't doctors work 8 hours and then go home?

The big difference between writing code and doing medicine is that patients won't stay the same when a doctor leaves for the day. With 8-hour shifts and 40-hour weeks, covering a patient around the clock requires 4-5 people. Those people will have 21 handoffs during that week. Each one of those handoffs is an opportunity for information to get lost, for understanding to fade, for followups not to happen. If people work 12 hours, that's only 4 handoffs. 16 and it's 10. 24 and it's 7.

Obviously, at some point the harm from overwork outweighs the harm from handoffs. But it's not an easy decision to make. When I'm debugging some weird, urgent problem, I know how valuable it is to stay with it, to keep all the state loaded in my head until I figure it out. And hospitals are full of weird, urgent problems.



This is really important and an often missed point. The handover can cause more errors than a fatigued doctor. So you would need to solve this problem at the same time (eg what if there was a way to change doctors without a handover)


Nurses have the handoff issues as much as the doctors if not more, and yet they work strict shifts.


Realistically, nursing is a job that is far more structured than is being an MD. This allows nurses to function much better as a group. That's precisely the crux of the matter and the (questionable) reason for the hours. Nursing teams are the stable basis on which the MDs rely on to fight the huge pile of s* * * work that escapes the standard pathways in which we try to shoehorn every patient.


I find it extremely unlikely that so many patients have difficult handover and accute condition that 24h+ shifts are necessary as the norm. Even assuming some patients and situations are like that, it should not be normal (which it is) and there should be enough recovery after (which it is not really).


Never underestimate the a ability of people or systems to fuck up.

When I had back surgery, due to a miscommunication in the nursing staff turnover I wasn’t given any pain relief 3 hours after a major surgery until late the next day.

My grandmother was given PT on the wrong limb. Again, poor handoff.

A friend got a big congratulations from the OB doing rounds when she was in the hospital. Small problem: she was there for complications of miscarriage.


That does not sound like defense of current system at all. It sounds like lack of process and system, relying on individual memory and recall instead.

In particular, both OB and PT on wrong limb are not just handover errors. First I am bit surprised that babies and miscarriages mix - they don't in here (so you know which it is based on room, but I think cause of split is something else). PT on wrong limb is error of not checking what you are supposed to do before administering - but also sounds like error people are more likely to do under time pressure or when tired and falling into routine.

All in all, after major procedure you are in hospital for days and they have to move you between doctors and nurses many times.


I hope it doesn’t sound like a defense. IMO this is a broken system that needlessly kills people as a matter of bureaucratic inertia.

Surgeries performed by the Ob/Gyns are housed on a specific floor. The situation was that there wasn’t a transition as the shifts changed due to understaffing, and the person doing rounds didn’t bother to read the chart.

In my surgical recovery state, the reasoning was poor situational awareness and burden on the nurses. Data was located in 4 different EMRs and they missed it.

Half my family is engaged in various medical professions. Every one of them is dissatisfied with how these systems work.


I had foot surgery a few months ago. When I came to from GA, my nurse was worried that my heart rate was a little too high. He did an ultrasound of my bladder, turns out no one had cathed me and I had 1200 ml of urine. Apparently the discomfort of an overfull bladder was causing higher heart rate. Nurse was shocked when he saw the ultrasound, ran to get the equipment to drain the bladder.



You'd find it unlikely, but that's literally the argument residency's governing body is making towards 24hr+ shifts.


I know. I also know software development teams that believe that long night is simply necessary for every single deadline and everything is always at the last moment. And I know teams that make deadlines without long night - not always but often enough and sometimes sooner.

The "it is totally impossible without 28h+ regular shifts" sounds like team in first category.

Not blaming individuals here, these problems are cultural and systematic and largely steam from leadership.


I've heard this rationalization from medical professionals before.

They need to solve the problem with handoffs. It's a big problem, and problem requires rewriting a lot of the process, but there's nothing more to it. It's solvable.

The institution however, has beaten any imagination out of them and replaced it with resistance to change.


How would you propose to solve the problem? Most attempts thus far have relied up getting attending doctors to do detailed data entry about patient state into an EHR, but that by itself is time consuming and detracts from immediate patient care.

Maybe someday AI and NLP technology will serve to automate much of the data capture, but we're many years away from those technologies being a practical reality.


What about doubling or tripling staff and then having them work in teams of two with half their shifts overlapping with someone that will be staying longer? Would that work?


How will you pay for that? Costs make that impossible even in the richest of countries.


All I'm interested in is whether it would work, not the cost. There are radical proposals that would manage cost better, but there's no point in considering them if it wouldn't help.


Which radical proposals would manage cost better?


I could propose as many pat, armchair solutions as the next HN reader :) However, it's clearly not a problem that's conducive to quick fixes, or it would have been sorted.

That said, the arguments for the status quo are laughable. There's no other field of human endeavour where people claim that 24 hour shifts are >safer< than 8 hour shifts. It's not like handoffs have even been eliminated: handouts are just done 3 times less, but by profoundly sleep-deprived people, who have to reiterate the last 24 hours to the incoming staff.

When people defend behaviours of this nature, it's a 'culture smell': to a greater or lesser extent, the proponents have been indoctrinated to cling to a local maxima.


Your experience with coding for a long time is exactly like mine, but no one is going to die if I stay up late and write terrible code. Otherwise the clear answer would be to just be time inefficient and pick it up the next day. If there is an edge case where an emergency happens in hour 7 of an 8 hour scheduled shift I guess that's fine, you can go hard like that from time to time and be fine, but not constantly. And this really is an edge case since most medical care is not in an emergency setting.


I agree that most medical care isn't urgent. But I think all doctors should be able to handle emergencies. It's similar to how I think we should work to never have urgent production issues, but that all developers should be able to debug urgent production issues.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: