Same thing happens with doctors in the US. Doctors should know that after the first 12 hours, the doctor's not as good at doctoring anymore.

Yes. Although in talking with doctor friends, it turns out there's an interesting counter-force. A major source of errors is handoffs.

I've certainly seen this when in the hospital with family; new doctors would frequently ask me for historical information, including about what other doctors had said or done. I was fine giving it, but what if I hadn't been there?

Do they work with checklists? That seems to be the number one reason why crews in aviation can work with such low error rates. Handoffs with strict checklists would probably cause lower error rates.

Yes, they do have a lot of checklists, but it's one thing to create a checklist for an airplane with very strict requirements for validation and maintenance and another to create a checklist for a human being.

As others mention, there's a big push towards checklists. (Note: they only work if they are bottom up, not top down.) But aviation mechanics work on something that is heavily standardized, and they get to shut down the thing they are working on, which eliminates a time dimension to the problem. Neither is true for people, so checklists aren't really the solution to the information-handoff problem in medicine.

Can you expand on top down vs bottom up? Do you literally mean reading from the bottom of the page, or starting with the most granular items and decreasing granularity?

Sorry, I mean in an organizational sense. If managers tell workers "Here are checklists! Now follow them!" then they aren't as effective in improving quality as workers saying, "Let's collaborate to make sure we are doing procedure X well. How about a checklist?"

Checklists are an easily observable symptom of an team drive for improvement of outcomes. But you can't reverse that and plop down checklists that are used under threat.

I don't know much about how hospitals work, do doctors even have "managers?" I would figure the doctors would be the ones coming up with and using the checklists.

It's complicated. Doctors employed by hospitals report to two sets of managers. There are other doctors who serve as medical department heads and the chief of medicine. But those doctors are also accountable to non-physician business managers for administrative and financial issues. And then independent self-employed doctors may have privileges at the hospital and see some patients there but don't have managers in the traditional sense.

Crap, I listened to a podcasts sometime last year about how using checklists in a medical setting decreased errors by a ridiculous amount.

I'm going to try to find it. But, if I don't, you got the gist of it:

Have a checklist, and stick to it. Don't expect one person, or even multiple people, to remember all of the steps off the top of their head.

Possibly talking about "The Checklist Manifesto" - http://atulgawande.com/book/the-checklist-manifesto/

It's an -excellent- book about the application of checklists to error reduction etc.

That's exactly it. I knew it was something based on a book. I've updated my comment below, if anyone is looking for the specific radio show about it.

If you find it, post it so I can listen.

I'm having a bear of a time trying to figure out what podcast it was. It might even have been a YouTube video. I've been going through different permutations of search terms to find it.

I'm worried it might have been .NET Rocks![1] They have a great podcast, but their episodes aren't indexed by Google because of their website structure.

I'm gonna keep looking, and even check my home computer's history when I get home tonight, but I wouldn't hold your breath. Someone a bit above linked to a book that I'm pretty sure was the impetus behind the podcast I listened to, so it might be worth giving that a read.

[1] https://dotnetrocks.com/

EDIT: HA! I found it! It was actually in the press section on the book's web page.

http://www.npr.org/templates/story/story.php?storyId=1222261...

The discussion about work hours here typically seems to put effectiveness after 4-6 hours as dropping off for programming, for pilots regulations allow 8 work hours (with 10 required hours off, and at least 8 of those consecutive uninterrupted sleep). It makes one wonder at the working conditions of doctors and sailors.

Toss in architects, visual effects artists, kinda of whatever job you want in the US, especially if there's no history of a union.

As someone else said, there's a lot of ego involved too. Given that a lot of it is "brainwork", it's the only physical way they might be able to feel they're testing themselves against something.

Most of it in those jobs is just production-churn and an attempt by the higher-ups to save money on things that a) were bid low and b) have shitty margins.

Medicine as well. Medical residents can have shifts that are 24 hours long.

Recently enough a group of doctors went on strike with perhaps one of the most reasonable demands I have ever heard. They didn't want to continue to do shifts that were over 24 hours long!

Well, as long as they only let people with jobs that can kill people be sleep deprived, it'll be ok

I understand that medicine is tricky because the handoff from one doctor to the next can be dangerous for patients. You have to balance continuity with exhaustion. That's not to say that they necessarily have the right balance now, but it's a hard problem.

I've heard doctors say this, but they never cite evidence. But there is a wealth of evidence that says that doctor performance drops significantly with motor performance being harmed so much as to increase the number of errors and excess movements. I'll add the citation at the end, but it's not about the single study; it's about the fact that there is little evidence to the contrary and lots of evidence saying it's the case. If this were a drug, we'd be prescribing it every which way.

And most importantly, know that this is not necessary. American doctor supply is restricted by the AMA. America could suck the best doctors out of many other nations if it so desired and its healthcare practitioners would be better rested, more mentally well, and far less error prone.

http://www.bmj.com/content/323/7323/1222.1.short

> I've heard doctors say this, but they never cite evidence.

Seems pretty well researched by a quick google. Maybe they're eliding the support not because it's an assumption, but because it's so thoroughly established they take it for granted?

http://www.nejm.org/doi/full/10.1056/NEJMsa1405556#t=article...

https://psnet.ahrq.gov/primers/primer/9/handoffs-and-signout...

...

> But there is a wealth of evidence that says that doctor performance drops significantly with motor performance being harmed

That's true too, but I worry you're trying to rebut (X & Y) with evidence for Y.

If we find ourselves slipping into a battle of emphasis, maybe we face a problem of optimization?

From what I'm seeing, your hypothesis about this being a question of optimisation seems well supported. I worry that medical professionals resist changes here out of cultural belief or a resistance to process. That wouldn't be without precedent. Even something as well-established as washing hands between patients isn't done without explicit process (compliance is near 13% without explicit reinforcement for something that everyone learns the legend of Semmelweis for), and surgeons refuse even _private_ outcomes tracking.

Agreed, I can't know from first principles if the right shift length is 4 or 8 or 30. I'm sympathetic to your suggestion that it would be a surprising coincidence if doctors landed on the perfect number through historical accident. Maybe they have a feel for safety in shift length, but they don't in other areas like washing, so maybe not.

It probably varies by specialty and even individuals, probably a monster of a problem.

Do you think tech in Silicon Valley is somehow above that, though?

Maybe? I've only been here about 2 years. First half was a bootcamp, which while intense certainly allowed for enough sleep (I would leave the bootcamp at 8pm, get home ~9, gives me an hour to relax, wake up at 7, and commute for an 8am arrival).

Second half has been at a company with a very pro-family-time CEO, and as a bachelor that just means a lot of time to explore this awesome new city :) So I haven't yet experienced the "utter madness startup" I guess?