[snowwrestler]

A friend of mine is a PA who works in emergency medicine, which means he pulls 30 hour shifts sometimes.

I challenged him about how, with what we now know about sleep deprivation, he could defend that schedule.

He pointed out that much of medicine, especially emergency medicine, requires deep complex analysis of a wide variety of symptoms, some of which might seem unimportant or unrelated at first. We've all seen shows like House where it takes a genius to diagnose the root cause of a set of weird symptoms. While that is obviously exaggerated, the reality is that diagnosis is often difficult and in an ER, happens continuously with treatment.

He said there is no way that a doctor or PA can fully hand that mental flow state off to another one. So the scariest thing to him is handoff--what if he forgets to document or mention some seemingly minor detail that ends up being crucial??

Long shifts give medical personnel more continuous time with each patient, reducing the chance that handoff will come too early in treatment, when mistakes or misses have a greater impact. It also permits long periods of overlap between shifts.

"Being sleep deprived is bad for care," he admitted, "but so are handoffs." He feels that as long as the total time per week does not exceed too many hours, long shifts are good for care.

[llimllib]

1. That's weird, what the heck ED do they do 30 hour shifts in? Non-US? My wife is an ED physician in the US and that does not ever happen in any ED she's been in.

2. She's also boarded in internal medicine, where she did have to do 30 hour shifts. They are terrible for patients and physicians, and I've been in heated arguments with physicians because I think the primary reason they exist is as hazing.

3. Handoffs are a legit problem; much more in other fields than in the ED. (Though handoffs do exist in the ED). It is a balancing act for that reason, though physicians could do a lot better with the handoff process IMO.

4. http://mef.med.ufl.edu/files/2010/01/Resident-and-Attending-... :

> Handoffs, the transfer of patient care from one health care provider to another, are known to be vulnerable to communication failures8 and have been called “remarkably haphazard.”

and

> When looking specifically at malpractice cases with communication breakdowns, 43% involved handoffs.

5. It's my belief that patient-centric design and communication could eliminate nearly all of these issues while reducing the need for long shifts, but there's a big [evidence needed] tag on that

[jblok]

What, with your knowledge, could be done with technology to solve this problem of handoff?

A notepad app probably wouldn't cut it, but is there any way to move the complex thoughts in a doctors mind into a database somehow?

[llimllib]

I don't know what the exact process would be, but I think that the key would be to focus on process. Start treating each patient mishap like the FAA treats every airline mishap, take steps to ensure that each doesn't happen in the same way again, and I think you could improve things a great deal.

I think doctors focus on, and are very good at, doctoring. A little more focus on process and the totality of the hospital experience could bring big gains relatively quickly.

That's just an outside opinion though, my wife would probably disagree with me.

[tdb7893]

I doubt it would scale. The FAA regulates one thing (flying planes) and there still are a ton of checklists and regulations, presumably if you tried the same thing with medicine it would collapse under it's own weight with all the different edge cases

[pmoleri]

I experienced this twice, when my baby daughter was in observation after two episodes a week apart.

She recovered quickly so she was released from ER to observation for 48 hours. In those 48 hours we would see new doctors and nurses every 8 hours, explain everything again, so they wouldn't miss any details from the history, and every now and then a doctor would come up with a new theory and order studies without first consluting with a specialist in that area.

One of this theories was epilepsy, after that we went to see a neuropediatric and dismissed the theory inmediatly, even before seeing the study.

All of this was in a stable condition, I can't imagine going through the same in critical condition. So, altough I think 30 hour shifts are excesive, I can understand why long shifts are important.

[knz]

My pregnant wife was bounced between clinics a few years back and subsequently spent several weeks in a hospital bed. We had the same experience of having to repeat the story every shift change (at least for the first week) and ultimately became incredibly frustrated with doctors doing rounds who would directly contradict the information provided by other medical staff or what was written in her chart.

One of the biggest lessons we took away from the whole experience was the need to be your own medical advocate (understand the health issues, pay attention to what doctors tell you, and don't be afraid to ask for clarification or challenge them if they say something that contradicts other information from a reputable source) - doctors work long hours and see many patients - it's not really surprising that they make mistakes and may miss critical information when scanning similar data sets hour after hour. Unfortunately for patients can be serious consequences to these errors.

(I hope your daughter is doing ok now!)

[pmoleri]

Yes, sorry for the cliffhanger, it turned out to be allergy to cow's milk protein (most probably as it's difficult to confirm). She was 5 month's old, and the treatment was the withdrawal of cow's milk for a year, then she was able to consume it normally.

[amorphid]

That's a really interesting point. I wonder if it'd be practical to address the handoff problem with rolling start times. So if average ER patient takes 3 hours to treat from start to finish, for an 8 hour shift, you stop taking new patients at the 5 or 6 hour mark. That being said, if an ER doc is in the "graceful shutdown" part of their shift, and a spike in patients rolls in the door, it'd be hard to say no to helping out.

[snowwrestler]

I think that is how it works, but with longer periods of time. Instead of an 8 hour shift with overlap of 2, it's a 24 to 30 hour shift with overlap of 8 to 10.

On rounds, a doc might only see each patient every hour or two... even longer for specialists.

ER docs also work some shorter shifts. So in a week he might work one 24 or 30 hour shift, and one or two 8-hour shifts. The shorter shifts provide arms and legs for simple cases, and the longer shifts provide continuity for more serious cases. The docs take turns holding down the long shift.

[mjevans]

I had similar thoughts, except with 4s.

For the first 4 hours they take patients (and shadow some of the persisting cases from other doctors).

For the next 4 hours, they're purely cleanup/handoff... EXCEPT in the case that a major crisis happens. Things extend 4 hours at a time in that case.

This would, however, mean staggering the doctors in 6 different shifts that could each handle the ingress load for their ramp up period.

Naturally doctors that are 'morning' or 'evening' people should be binned in to shifts compatible with their biological schedules.

[Spooky23]

Where do you get the 5-6 doctors to make up the difference since you're working a short shift plus pairing with another doctor for half your shift?

[mjevans]

Think of the 'doctor slots' as cutting blades with overlap.

The first four hours (half shift) the doctors are taking in new patients. In the next four hours they aren't taking them in, they're finishing processing and starting to pass them off to the next shift if they appear to be complex cases.

[imesh]

My brother just finished his residency. His opinion is they do it for cultural reasons and it's a terrible thing. I'm curious what percentage of doctors think the way they do it is a good idea.

[rwallace]

If handoffs were really the reason, a few seconds thought suggests ways to deal with that. For example, assign fewer patients per doctor, and let the doctor grab short bursts of sleep in between seeing patients. Yet somehow, I feel confident in predicting that if you start suggesting those solutions to the authorities, you'll find they are utterly uninterested.

[michaelmachine]

Yeah there is lots of literature on this issue and some ongoing studies trying to determine how to reduce medical error during the hand-off. For example: http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD...

[mpweiher]

Sounds like there's a need for pair-doctoring.

- Always have 2 doctors for everything

- Stagger/overlap the shifts

[kahrkunne]

- double the cost of already expensive medical care

[vinay427]

No, because 100% of your medical expenses surely don't go to payroll for doctors' salaries.

[maxerickson]

Did he have statistics on the typical amount of time he spent caring for patients?

Like, how many people are in the care of an ER for 10 hour stretches in a given week.

[braveo]

overlap the shifts by 2 hours.

Why is that not an acceptable solution, outside of economics?

[lostlogin]

Who wants to start the AM 2 hours earlier or finish the night 2 hours later?

[braveo]

people who get paid really well and don't want to kill other people?

Who wants to pull 30 hour shifts?

[user837387]

General AI will make all of this irrelevant. A General AI doctor never needs to sleep and would never forget anything relevant.

That is, if the government allows it to be used in consumer products. It would make a potent military weapon. Probably more potent than any military weapon invented so far.

[eropple]

Devices of uncertain possibility and certainly no timetable certainly will make such a thing irrelevant.

Meanwhile, we're discussing the world that exists.

[user837387]

>>Devices of uncertain possibility

It is certain. 100% certain. It will be simpler than the brain and way more powerful.

>>no timetable certainly will make such a thing irrelevant.

Yeah, not sure when this will happen. But it will happen. It'll happen within my lifetime.