Nurse faces prison for a deadly error. Her colleagues worry: Could I be next?
npr.org
npr.org
An overwhelming number of alerts were routinely displayed, leading to a form of banner blindness and frequent overrides. An overzealous autocomplete suggested the wrong drug with the same first two letters. There was apparently not a permission system in place that blocked her from administering the drug for this patient.
The patient died soon after administration, and the nurse is now facing criminal prosecution for reckless homicide. Other nurses across the country are concerned, especially given their overwhelming workloads during and after covid and the frequent room for error.
A fire alarm that triggers every day for no reason is useless. In weeks people will ignore it and, when a fire actually happens, they'll die.
But there were also a bunch of systematic factors contributing to the error. The UX stuff was actually buried very deep in the article, I just thought it was an interesting part of healthcare that's rarely discussed.
The medication she administered was in a different form, even (powder instead of liquid, like kool-aid). But that still didn't stop the mistake.
Everyone in healthcare right now is exhausted all the time, understaffed and overworked, and enough small errors can accumulate into major catastrophes. Didn't seem like there was a process in place to prevent this, either in software or training or day to day.
She's not trying to avoid responsibility, but also, should she get prison time for a mistake vs losing her license and job? And who else in the chain of command should be held responsible? How will this ripple across the healthcare industry?
We have a hard time filling nursing positions already, and the other nurses interviewed for the article all worried the same could've happened to them.
What has happened is this: nurse tasked to administer Versed. Versed - what's that, don't care. <goes to medicine cabinet> VE-clickedy-click. Vecuronium? Dunno what's that, but it's good enough. Big warning! Medicine be dangerous! Medicine is administered, patient dies. Nurse says, mistakes were made, but Jesus loves me!
Enough of that.
You might be surprised to hear that most people you know are doing what they can to make a living.
It seems like you may not see it that way, while instead making a moral judgement of who “really cares”. I understand your sentiment, but I would advise caution in such situations.
Some jobs are simply not for those whose personailty does not mesh.
Another example would be an introvert, taking a job as an MC. Not the best fit there.
GP is not talking about the nurse which got it wrong, by accident, but about those that clearly are going to have difficulty doing the job properly. There's more to a job than cramming info.
Are you trying to tell me, if you were a teacher, and you were to pass or fail a student who you knew would end up unable to perform, you'd just say "Oh well, hope they get better!", and pass them?
I'm not sure what you're trying to argue here. I think you've misread the GP's statement.
That's precisely what they want, they want to be understaffed and underworked to create the myth of the suffering heroic doctor, so that they can justify all the other privileges they get and the gigantic salaries, bribes from multi-level marketing, and disgusting hedonistic trips to the Caribbean.
There's good doctors too and guess what? They're hated by mainstream doctors. I love those guys, they're so good.
If doctors were actually good the absolute first thing they would instantly do is allow there to be ten times as many doctors, that would make everyone in America much healthier instantly.
These guys are like Google, but a pentillion times more expensive. In fact one time I went in, I started talking to the doctor, he just started googling right in front of me.
That's not how it works. The process of becoming a doctor is much more formalised than you seem to insinuate and mostly does not depend on a single individual or a group of individuals other than indirectly.
I do agree that some doctors have a Messiah complex and are self-important to the point of absurdity, but they still mostly tend to do a good job.
I would argue that the limiting factor of why there is a general shortage of doctors and nurses is that the job is not the easiest one, physically or mentally, people tend to work long hours and awkward shifts and the education itself is challenging and long.
Both of those are professional masochism. In particular older doctors make sure would-be doctors and young doctors go through decades of shit in order to keep the old doctor's privileges and salaries incredibly high, without making it obvious it is a completely cornered market. If doctors could, they would close the medical schools. But they know they would get lynched by a mob, for real, really in real life, the laws against it would stop nothing, so they have no choice but to leave medical schools open producing a symbolic number of replacements with an excrutiating amount of hazing and harmful education, which very much makes them worse doctors. And lots and lots of attrition, the old doctors love that, mmm attrition, all the benefits of saying there's slots in medical schools without the actual competition from petulant young doctors.
They also want every doctor to be minimally competent and that goes completely against the goal of having as many doctors as possible, or, better, highlight the fact that this is how many doctors (or healthcare professionals) we can have considering cost of education, both monetary and in time.
If you really want more doctors, you could also pay software engineers and lawyers less. I was married to a doctor and I know how much she studied to be one. It's much easier to become an average software engineer or lawyer than becoming an average doctor. And you wouldn't want to meet an average doctor right after totalling your car with you inside. Changing the financial returns would make the profession more attractive to those who are in the game for money.
If I could make what I make counting squirrels in my backyard, do you think I'd think twice before leaving software engineering? ;-)
Why should there be? Versed is a Schedule IV drug, vecuronium is not scheduled-if anything, the system worked since she was able to dispense a drug that wasn't controlled as well as the drug she had permission to dispense.
In the acute care space there's lots of instances where an end user will need to override something for some reason or another. It's ultimately the end user's fault for neglecting the errors.
DEA scheduling is supposedly about abuse potential, not safety. Vecuronium is one of the most dangerous drugs in the hospital. It's used to paralyze essentially every muscle in the body except the heart to keep a patient from moving during surgery or to get a breathing tube through spasming vocal cords. These muscles include the ones used for breathing. Most physicians aren't allowed to give it. Those that are have trained for years before they're allowed to give it independently.
There are no circumstances when a nurse should be giving this drug without a physician closely supervising.
Specifically:
* Imagine being fully conscious but being unable to move or breathe. Imagine the panic as you suffocate. That's how this patient died.
* The drug the nurse intended to give (Versed) was a sedative. The standard of care for sedation includes monitoring. There was no monitoring as there should have been. A simple pulse oximeter would have caught the error before it was fatal.
* Vercuronium is especially dangerous and thus tightly controlled. Only attending anesthesiologists and emergency medicine physicians give it independently. Residents give it under supervision. The only nurses who give it are CRNAs under direct supervision. The nurse had to go out of her way to get around restrictions designed to prevent her from giving it. The drug has warnings. And she had to mix it up herself, unlike every other time she gave Versed.
From the article, it sounds like those restrictions were on the level of severity of a click-through that one is required to do many times a day. Those quickly become meaningless.
Even if the nurse had administered the correct medication, this patient would have been in danger. Versed is a benzodiazepine, and not without its own risks — especially in an elderly patient. What was this nurse doing paying so little attention to a patient she thought she had just sedated?
Why can the person who wants to withdraw meds override the machine? Overrides should need at least a 2-person check, ideally a supervisor as the second
> she and others say overrides are a normal operating procedure used daily at hospitals
If you train people to be alert blind.. they'll be alert blind.
The entire situation sucks, I hope it's determined this is a systemic issue and not the individual
Can't help but think if it was a pharmacist instead of a machine this would have been prevented too
Those folks are generally busy somewhere else, and there are few benign drugs in these situations. There are also a lot of legitimate urgent situations where an extra minute or two can get someone killed who otherwise would not be.
Crash carts and the room in general have all the meds needed to handle urgent issues don't they? You don't often see someone crashing and a nurse running off to grab a script
Something I've personally seen when I was doing my EMT clinicals was a geriatric stroke patient came in (time is key here), and staff struggling to get the anticoagulants and another medicine I can't remember (but an important part of the stroke treatment protocol) from the machine in time - which they ended up not being able to do, as no one was around who had the countersigning authority in time.
Which was at that point past the time the patient really needed it. They still got it, and who knows if it was the deciding factor or not for the patient. The stent they later tried to install also failed during installation after a rather large dose of radiation to the patient, so it was unfortunately just not their day.
It was a sad situation all the way around.
A lot of these processes are designed as CYA/risk aversion methods (which make sense from that perspective), but then also have the unstated goal of allowing cutting already overworked staff to the bone. And have the secondary effect of having people get even more exhausted, and make even more dumb mistakes, and also get blocked when it is a legitimate need by a process/approval. Though admin's ass is covered (as long as no one thinks about it too much), since the staff had to bypass a bunch of safeguards before it blew up.
Missing in the press release or indictment of the staff member though is that EVERYONE had to bypass the safeguards all the time because of the sick system they were in.
The only real way to address this (IMO) is to tackle the internal structural problems causing massive overwork and lack of staffing in the medical field. Good luck with that though.
>> While testifying before the nursing board last year, foreshadowing her defense in the upcoming trial, Vaught said that at the time of Murphey's death, Vanderbilt was instructing nurses to use overrides to overcome cabinet delays and constant technical problems caused by an ongoing overhaul of the hospital's electronic health records system.
As for missing the warning on the vial cap, were nurses trained to look for messages there? I can easily imagine that the process of extracting medicine into a syringe would become so routine that an unusual exception like this could be missed without some kind of check process.
High pressure, intensive work comes with an inherent risk. Attempting to reduce this risk by adding additional pressure with threats of punishment to the people who do this extremely important work every day seems grotesque to me.
Reduce the pressure and intensiveness of the work as much as possible. Accept that these people are already inherently motivated not to kill others or themselves by accident or neglect.
Do they not want ambulance drivers in Denmark? That is mind blowing to me.
I'd argue that someone hitting an ambulance crossing a red light should be considered on an individual case, and we should accept that 99% of the time, everyone (who are not driving an emergency vehicle) are on mental autopilot and it works Just Fine (tm) and that when the rare case comes, we should accept that, we shouldn't punish people for being humans and not hyper vigilant machines.
Sure, it's easy to argue that "when you drive, you should DRIVE!" yeah, sure, outside of the race-track, that's just not how human brains work, without some boost of adrenaline, there's simply no way to keep the kind of vigilance required to act correctly 100% of the time.. (heck, I've driven out early during races because I've felt myself going back to "just driving around" mode).
In short: it's the technology. That article changed my views on warnings, and I think it should be required reading for anyone designing a system with alarms and overrides.
While testifying before the nursing board last year, foreshadowing her defense in the upcoming trial, Vaught said that at the time of Murphey's death, Vanderbilt was instructing nurses to use overrides to overcome cabinet delays and constant technical problems caused by an ongoing overhaul of the hospital's electronic health records system.
Murphey's care alone required at least 20 cabinet overrides in just three days, Vaught said.
"Overriding was something we did as part of our practice every day," Vaught said. "You couldn't get a bag of fluids for a patient without using an override function."
This clearly speaks to a larger systems issue and highlights a problem with the way medical software is designed to deflect liability from the hospital to individual employees. Why is a nurse able to get a paralytic agent from a PYXIS or similar machine outside of critical care? That falls on the hospital, pharmacy, and the nurse in terms of responsibility.
Clearly she should never work in healthcare again from incompetence but if she burns due to this being criminal: everyone else above her that lead to this should burn too.
When near every interaction flags the same alert you become numb to it. If you are involved in this field I caution you: get some actual clinicians and stakeholders involved early even if your customer is the hospital because the priorities are different.
I know critical care nurses that get flagged on almost every medication with overrides and alerts because it is physically delivered late and they have more than one patient. The system don't measure the metric of when the medication actually arrives on the ward rather when it is ordered and delivered so fundamentally almost everything is late. What are you going to do? Not take care of your other patients, balance priorities or will the time and medication into place? It's a systems issue of being under resourced and still having to deliver care.
As a physician I run into a version of this every day in my clinical practice with alerts for interactions and pharmacy faxes for interactions that are clinically irrelevant based on a database flagging it. After a while it becomes numbing and you start to get cognitive biases. It's really not that different from the circumstances that led to the Challenger disaster.
This will not help the overworked state and the next issue I can see is confusing two drugs (marketing _loves_ to confuse people, given enough phonemes it would be easy to type the full name in wrong); potentially asking for a brief description of what the caregiver expects to happen, with a quick NLP pass comparing the "effect description attached to drug" paragraph and the 2-3 sentence "expected effect description from caregiver". Yes more paperwork, but saving lives is usually worth the extra effort. Everyone already knows the long-term overwork solution is cut down on admin & hire more personnel.
I don't know how these systems are designed and tested, and what regulatory hurdles they have to pass, but it sounds like there is a huge disconnect between how the manufacturer expects them to be used and how they are actually used, with frequent overrides, day to day. It must be a tough industry to work in, either in the patient facing side or the medical devices and software side. I'd hate to be the person who coded all the warnings in that software. "What do you mean they bypassed all 7 warnings? Even the one that said this was a paralyzing agent?!"
I wonder if some of this wasn't a procedural failure too. Like why doesn't a potentially life ending drug require at least two people to vouch? Even in retail a manager has to come and turn a key for some trivial refund, or in our field a reviewer has to approve changes first.
Is it that the hospital cheaped out on staffing so they didn't want two nurses double-checking each other? Was this use case never accounted for in software development? Are all drugs potentially life ending so there's no way for the software to reliably reduce false positives? So many questions...
There is a certain level of cognitive offloading happening with the machine that is uncomfortable to me. Its like the Tesla autopilot giving drivers a false sense of security. One could easily fall to assuming what came out of the cart is the right thing. People are much less trusting with each other.
Also, it seems like a really bad idea to me to mix trade and chemical name in that type of system. You should search one and only one type of name at a time, simply to ensure namespace tripups like this don't happen.
This sounds like cognitive load failure from workload mismanagement.
First you select a drug, then the system estimates danger and urgency(Safer drugs often used in cases where seconds count could bypass the extra check), then the cabinet displays all the purposes one might use the drug for, along with random unrelated purposes and a text field.
If you ask for a sedative, you have to say why. And if you click an unrelated purpose, it will make you retry everything, and also show up on a report that you and your supervisor can look at to see if there are any patterns of frequently being caught.
Especially dangerous drugs could even require control room authorization, one operator could potentially supervise dozens of nurses remotely.
I'm not sure if people would get even more complacent, but it seems like errors are already common enough that it might be hard to make it worse.
"Type the full name of the drug to dispense."
Or better, make sure you have processes and procedures to minimize fatality-causing errors - if others won't create them, YOU SHOULD for yourself as a professional! A VERY TRIVIAL process is having a checklist what includes: "Visually validate medication label to patient order" and even "Have one other attending nurse or doctor confirm validation". This is done by pilots as SOP!
I have ZERO sympathy for her or for "colleagues who now worry"... NONE.
Like everyone else, I don't have the full picture of what happened in this particular case nor the extent to which this nurse's activities amount to criminal negligence. But I know for certain that we don't improve safety of life-critical systems generally by having a "if you make a mistake then fuck you" attitude.
* You have been working 12 or 16 or longer hour shifts 6 days a week or more for the last two years due to COVID, and 24-hour or more shifts not infrequently.
* You have many more patients than you should, because the hospital can't or won't hire more staff (partially due to COVID, partially this is just how it is).
* You are regularly abused and belittled by patients, hospital admin, and maybe even your colleagues (medicine is often quite toxic).
No process whatsoever can make up for the level of chronic exhaustion and incapacity these circumstances produce, and shouting "there should be process for this!" is making the problem worse. You are doing the same thing as the patients who expect to be treated like royalty - expecting humans to operate like machines. If it wasn't this nurse, it would have been another one. Few people can do better under these circumstances.
That doesn't mean we don't hold people responsible, but without systemic change, this will keep happening. You have trivialized a very very large problem.
Then every time you try to adjust the flight controls, four warnings pop up, company policy is to ignore the first two and sometimes the third. You're always supposed to pay attention to the fourth alert because it's super important, but it looks exactly the same as the other 12 alerts you saw this hour, and the passenger in 12B is having an asthma attack and the one in 20C is vomiting and upsetting everyone around her. And your copilot just quit again and your planned vacation for next month got canceled and you're asked to work two doubles again.
You make a mistake, and a year later some lawyer asks you why you dismissed warning #4 nine months ago when SOP clearly says you need to pay attention. Don't you know how to do your job?!