Hacker News Reader

Comment by technofiend | Hacker News Reader

>One group runs a gauntlet of fire by sleep deprivation during pre-med training: with on-call work hours up to 120 hours a week.

I would argue this is the source of injury and death to the patients that needs to be addressed rather than powered through in some sort of macho hazing ritual. Anecdotally a very tired doctor is allegedly how my wife lost her mother.

View on HN

I used to work in hospital IT at a level 1 trauma center from 2004-2006. While I was there, an article was released which studied the relationship between time on shift and errors made by staff. As I recall (and it has been 10 years) the findings were:

* Through roughly 9 hours, the error rate was about the same. Say, x%

* In the 10th hour, the error rate increased. Say x + y%.

* In hours 11 and 12, the amount the error rate increase doubled the 10th hour's increase. So x + 2y%.

* In hours 13 and 14, the amount the error rate increase quadrupled the 10th hour's increase. So x + 4y%.

* In every subsequent hour studied, the error rate increase doubled again. x + 8y% in hour 15, x + 16y% in hour 16.

I want to say the study went to 18 hours, but I don't recall if it was 16 or 18 at this point. The paper made the recommendation that shift lengths should be limited to 10 hours maximum.

Every health care professional I spoke to about the article (resident doctors and nurses, maybe 10 overall, mostly in the ER) said the same three things:

1. They had personally witnessed someone make an error they could attribute to tiredness.

2. They themselves had never made a mistake due to tiredness.

3. They didn't want to change their schedule. Either they liked having 4 days off every week, or they thought it wasn't a significant problem.

2. They themselves had never made a mistake due to tiredness.

3. They didn't want to change their schedule. Either they liked having 4 days off every week, or they thought it wasn't a significant problem.

I think those two items and the tribal knowledge that handoffs are more dangerous to the patient helps the overwork model persist. Sounds like it's past time the medical industry prove that handoffs are more dangerous to patient outcomes. Doctors are trained to be problem owners and problem solvers, but that doesn't make them good team players. And lowering handoffs also limits oversight and prevents second guessing which is great if you're convinced you're always right, but clearly doctors are not always right and patients often pay the price.

Medicine is a conservative field - unfortunately that also includes ignoring anything harmful to their cult aspects even if it is trivially obvious like "wash your hands" after doing a damn autopsy before delivering a baby. Or yes infants can feel pain - not using anesthesia because the control isn't fine enough to not have a better chance of killing them is a grim but justifiable thing initially but inexcusable once it becomes possible to do so safely.

And the sleep deprivation and hazing "I did it so everbody else must too!" is definitely a cultic thing.

I use cult very deliberately to point at the reasoning being entirely irrational and social as opposed to underlying value. And also because actual cults use slerp deprivation.

Peter Attia (MD) had a podcast episode where he talked about being a resident and how it almost killed him (asleep while driving) and was bad for patient.

Also explained that the guy who created the residency program was a cocaine addict who rarely slept, and since then all doctors have to try to follow his crazy schedule for no good reason..

Yeah. I mean, I get that hospitals are 24-hour operations that need doctors & nurses available at all times. So some people are going to get the crap end of that stick and have to do night shifts. But is it really necessary for them to work 80+ hours while they're at it? It seems like there's enough people trying to be doctors that you could cut that down to a healthier 40 hours+ 15 hours on call if it's really necessary to. Heck, even 50 hours.

You would likely have to extend residency based on everything that you need to learn for a given specialty. Med school graduates have an average of over $180k in student loan debt - from med school alone - and resident salaries in the 2-4 year programs are mid five figures.

Given the choice, I'm not sure someone whose 4-year earning potential is capped at $60k with $200k in student loans would want to extend that to 5/6/7 years.

I believe the GP is talking about increasing the number of doctors, not their years in residency.

You learn how to do your entire specialty during residency. If you cut the number of hours, you have the same amount to learn in less time.

That's on the assumption that amount learned per hour remains constant as hours worked increases.

And that sleep deprivation has no effect on learning ability.

>You would likely have to extend residency based on everything that you need to learn for a given specialty.

I would challenge that assumption because I don't believe there's any consistent number of hours worked by residents in rotation, is there? I mean there are published schedules and then there are actually the number of hours worked which at least according to the other posters is even more than scheduled. So if there's already an element of randomness here and different doctors are getting different numbers of in-rotation hours then it's plausible hours could be made consistent and reduced, isn't it?

You may want to look into why med school is so expensive, compared to other majors.

Here in the UK, the way I heard it from a friend who was doing their pre-med, is that there was a built-in cost incentive: hospitals paid doctors who were on-call at one third of their regular hourly rate for out-of-hours on-call coverage. (That's not regular hourly rate plus a third; that's one third of normal wages for hours after the first 40.) So the hospital administration had a solid reason to work their interns and house officers into the ground rather than hiring extra junior doctors.

The original rationale was that the "on call" hours were not supposed to be busy and the duty doctors could spend most of them sleeping in a bunk or studying: but by the late 1980s (when I heard about things) they were working more or less constantly through their shifts.

The EU Working Hours Directive was supposed to fix this by banning workers from putting in more than about 50 hours a week without very specific protections being enforced, but one of the first things the UK's Conservative government did in 2010 was to stop enforcing this.

I wonder how much of it is financially motivated in this way and how much is the inertia of this hazing-style culture. From what you've said, I'm sure hospitals would object to a change out of financial self-interest, but it seems the ingrained culture stops the issue from getting a big push to begin with.

I know labor unions (sometimes rightfully) get a bad rap, but it seems this is exactly the type of abuse they were designed to stop. There are some [0] but the rate is low, less than 15%, and there's a sort of self-censorship style of pressure against pushing harder for them.

[0] https://www.theatlantic.com/business/archive/2017/02/doctors...

But Doctors have a very strong labor union, which tends to advocate FOR this system. Probably even though the AMA represents residents, doctors who have succeeded in the residency system have more sway within it.

I'm in the US. After some searching, I can't find any information about most doctors having unions. I found one that appears to mostly work with California doctors, but that was about it. Either way, I was speaking about residents, who don't have particularly strong or ubiquitous unions. Again, that's in the US. Other countries may vary.

I was being tricky here but I mean the AMA.

It controls a great deal about doctor's education and working conditions. It does not collectively bargain, so it's not strictly a union. But it's more powerful than most unions at this point. So a glib, "maybe doctors should get a union to represent them" answer to poor working conditions for residents doesn't really make sense. They already have a powerful organization that should represent them.

I think that there is at least some group of physicians who really think that poor working conditions for residents improves patient outcomes and doctor training.

That is very true, the AMA is not formally a union, but it does exert a significant level of control and influence in standardizing the practices of doctors.

As for patient outcomes, I'd love to see a study of them for the roughly 15% of residents that have a union and very modest work place improvements, compared to outcomes for the rest of residents. You're right, many doctors do seem to "feel" the traditional method is superior, but I'd like to see hard data.

I know I've seen some before, I think The Atlantic had an article about it in the past year.

I found this paper: http://www.acgme.org/Portals/0/PDFs/Position%20Papers/Commit...

It's citations have some hard data. Maybe the most interesting part for me was this:

"There exists instead a widespread belief that physicians can be trained to defy the biology of sleep and that safeguards are in place so that patients and residents are not harmed by work schedules that are unheard of in any other workplace, let alone a hospital. That belief is most evident today in the FIRST and iCOMPARE studies that set out to prove that there is no difference in patient outcomes from residents who work 16 or 30 hour shifts. The principal investigators were so convinced that no harm would come of these experiments that they determined it wasn’t necessary to obtain informed consent from either patients or residents in the hospitals where the studies were conducted. This determination has been widely disputed and is now under investigation by the Office of Human Research Protections."

> It seems like there's enough people trying to be doctors

Not in the US. The doctor per pop count is very low.

https://www.nationmaster.com/country-info/stats/Health/Physi...

I think the key word in the sentence is trying to be doctors. The doctor per pop count in very low because med schools have super low acceptance rates[1]. That would probably be the best place to implement some sort of reforms if we want more doctors.

[1] https://www.accepted.com/medical/med-selectivity-index

There are plenty of people trying to be doctors, but there aren't enough programs to train them at the moment.

I imagine if med schools made it easier to get through their programs, the end result would be a significant lowering of the average pay for their profession, as more doctor's hit the job market. It sounds like it's in the best monetary interest of doctors to keep their professional supply low, allowing the demand for them to be high.

The counter argument that I have heard is that patient handoffs are where a disproportionate number of errors occur. Increasing the number of shifts means that more patients in the ED or on the floor will have care fragmented between providers, making it more likely that results will not be followed up or that changes in a patient's status will not be recognized.

I don't know at what point the errors from sleep deprivation exceed the errors from patient handoffs. People seem to take different views depending on what side of the work hours debate they fall on.

This would be easy to do a controlled experiment to compare.

My girlfriend is a second-year OB/GYN resident (which is a 4 year program) and while that field's residency is less insane than some other specialties like ER, she still works 12 hour shifts Mon-Fri (6a-6p on paper but generally 6-7:30), one 24-hour weekend shift a month, and one 12-hour weekend free clinic shift a month. This is on top of the "extras" that are not work but are required to graduate the residency programs. Weekly rotating presentations to the rest of her group (4 other second-year residents so one ~45 minute presentation every 4 weeks), research, generally keeping up with the state of the art in her field, etc.

So she's only "scheduled" for ~67 hours a week averaged throughout the month, but realistically it is in the 85-90 range.

It's easy to see how a more demanding or emergent field could seriously select for folks who are more able or willing to work on less sleep.

I've always assumed that the failure rate for handing off care to another person and the rate because of long hours must go in favor of long hours. I don't have anything to back this up except the MCAT being designed to check how you perform when tired. Now with electronic records I've been wondering if long hours are still be necessary, assuming I'm right about the failure rates.

I'm sorry about your mother in law regardless.

This is the claim, but if you look at the experiments supporting it, they’re all bonkers. For example, care is no worse from doctors on 28 hr shifts vs 24. I would argue that both of these are so far from a well-rested baseline that they’re meaningless.

We know from lab experiments that performance craters way, way before that, which is why pilots, truck drivers, cops, and every other profession work shorter shifts.

All this talk is insane. I have like maybe 4 top-performing hours on my best days. The idea of working more than 8 in a life safety critical situation boggles the mind.

Super intersting to see people discussing this! As someone that holds a pilots license and talks with a lot of commercial pilots about their jobs, sleep is often a hot topic of discussion. It would be illegal, not just irresponsible, for a pilot to try pull off a 120 hour work week. Why do we treat doctors differently? Their macho behaviour in my opinion is unacceptable, and likely reduces positive patient outcomes, yet it goes on?

Medicine has different concerns than piloting, though. Pilots don't have to ensure continuity of care for their passengers, and can accomplish their work in reasonable chunks. Doubling the number of patient handoffs, on the other hand, significantly increases the risk of one of the more common sources of error.

But could the increase in errors from increased hand-offs be smaller than the increase in errors from sleep-deprived doctors, nurses and patients combined with the increase in disease spread and disease progression from the effect of sleep deprivation on the immune systems of doctors, nurses and patients? Additionally, could there be benefits from getting more eyes on a patient's condition?

I've always wondered if the problem might be in the handoff procedures. Those studies about mortality in the event of a handoff always make it sound like obviously handoffs are an inherently dangerous thing, but perhaps the modern way of doing them is just poorly implemented? It feels extremely unlikely to me that absurdly long hours are the only solution.

I think this is a rational question to ask! Often people will jump to conclusions "handoffs cause errors", and while errors are a symptom of handoffs executed poorly, it seems uncommon for people to often ask for improvements to these processes.

There's just no way efficient way to capture, log, and communicate every single little detail a doctor observes while diagnosing and setting treatment for a patient. First impressions are often wrong, and little details can become significant later. Early treatment might be designed not only to address symptoms but to exclude other diagnoses.

In a plane in level controlled flight, there is very little to hand off between two type-rated pilots. Both folks understand the machine, and the machine is working the way it is supposed to. Humans work the same way! Parents "hand off" their healthy kids to schools or babysitters or relatives every day.

But imagine a plane that is in the process of crashing; it's in a dive, controls are not responding as expected, one of the engines keeps turning off. A pilot is fighting to regain control... how comfy are you with THAT pilot handing off the aircraft to another pilot in the middle of that situation?

It's a little silly as an analogy, since plane crashes tend to be resolved pretty quickly one or the other. But conceptually, just imagine a plane that is in the process of maybe crashing for 12 hours. There's a good argument for a pilot to just see that through instead of "clocking out" at 8 hours.

> Pilots don't have to ensure continuity of care for their passengers, and can accomplish their work in reasonable chunks.

Are you aware that on long haul international flights pilots do in fact rotate who is actively "flying" the plane? Flying is in quotes because most of the work is done by automation these days. I won't draw any analogies between the autopilot doing much of the work for pilots and nurses doing it for doctors because I can't actually support the statement with any data.

Yes, but that's the whole point: there's little risk in rotating pilots, whereas rotating members of the care team carries a substantial risk of key information falling by the wayside.

I see your point, but i'd love to see a peer reviewed study to back up the claim that "it's better to have over-worked doctors and nurses than to introduce more handoffs".

There's no error in hand offs. Attending doctors are detached and spend just a little time with a patient during their stay. Nurses and PAs do most of the work.

Flying a plane is way less complex than medicine. A pilot can hand off a flying plane way, way, way more easily than a doctor can hand off a patient with a complex set of symptoms and treatments.

I should also point out that the vast majority of doctors don't work 120 hours a week continuously, rather, they experience higher-than-usual clusters of working time vs. not working time. That is, they might be on rotation for 36 hours straight but then off for 36 hours or more. And some of their shifts might only be 8 hours. Residents work longer hours, but are supervised by doctors.

I know several ER docs and they all cite the dangers of patient hand-off as the main reason they continue to support long shifts.

Many sane minds would argue that, but the ACGME is actively arguing otherwise. They are actively building rules that allow residents to work longer shifts (up to 28 hours straight) while claiming these rules improve patient safety.

Patient hand-offs between shifts tends to introduce errors. Instead of finding ways to improve hand-offs, the ACGME is simply trying to have fewer of them.

These are supposed to be some of the "top minds" in medicine and that's the best they can come up with.

That's no really addressing the problem, it's just doubling down on the mistaken notion that the treating physician is the best one to tend to the patient under any circumstance. Doctors need to learn to be team players.

There's been some study(s) which suggest that medical error is the 3rd leading cause of death. This may have been debunked so if wrong please correct me.

https://www.cnbc.com/2018/02/22/medical-errors-third-leading...

Reply on news.ycombinator.com