Many people rightfully ask
> Why has this not been adopted everywhere _yesterday_?
A book I read a few years ago might have the answer.
_Catastrophic Care: How American Health Care Killed My Father—and How We Can Fix It _ 
Here's the description, to evaluate if you want to give it a read:
> In 2007, David Goldhill's father died from a series of infections acquired in a well-regarded New York hospital. The bill was for several hundred thousand dollars--and Medicare paid it.
> These circumstances left Goldhill angry and determined to understand how it was possible that world-class technology and well-trained personnel could result in such simple, inexcusable carelessness--and how a business that failed so miserably could be rewarded with full payment.
> Catastrophic Care is the eye-opening result.
> Goldhill explicates a health-care system that now costs nearly $2.5 trillion annually, bars many from treatment, provides inconsistent quality of care, offers negligible customer service, and in which an estimated 200,000 Americans die each year from errors. Above all, he exposes the fundamental fallacy of our entire system--that Medicare and insurance coverage make care cheaper and improve our health--and suggests a comprehensive new approach that could produce better results at more acceptable costs immediately by giving us, the patients, a real role in the process.
We also need transparent pricing and reviews. Sites like ZocDoc are trying to fill the niche of something like Yelp but for doctors. There's lots of easy low hanging fruit here.
I'm a bit perplexed that something so simple, can reduce post surgical deaths by a third.
For example, why would you need everyone in the room to say their name and what they do?
Simple: everyone is wearing a mask and often surgical scrubs that are the same color. By having everyone say "Hello, my name is Dr. Jones and I'm the anaesthesiologist" you can quickly determine who is the person to direct questions to about the patient being under.
It gets even crazier when you hear stories about doctors going into the wrong operating room (especially at bigger hospitals where there are several ORs). A simple "State who you are and why you are here?" costs very little and helps avoid costly mistakes.
Another example: having a checklist allows a junior person (e.g. a nurse) to challenge more senior people when they make a mistake.
Example without checklist:
Nurse: Dr, I think you forgot to do X.
Dr: I know what I'm doing, don't question me.
Example WITH checklist:
Nurse: Dr, you missed step #4.
Dr: I'm sorry, you are correct. We all agreed that was a necessary step and I missed it.
As have others in sibling threads, highly recommend Gawande's Checklist Manifesto.
Example without checklist:
Nurse: Dr, I think you forgot to do X.
Dr: Oh, you are right, thanks a lot
Example with checklist :
Nurse: Dr, you missed step #4.
Dr: I have done this list every day for the past 15 years, I know I ddid not forget anything, don't question me
You're not wrong, but if the hospital policy is to follow checklists, then the doctor overriding it would have to answer some questions afterwards if something were to go wrong.
Presumably a review would occur, and the fact that the checklist was ignored could be used as 'evidence' against the doctor.
My understanding is that if the checklist empowers, say, 5% of the nurses to speak up and 5% of surgeons to acknowledge the error that is still a huge net positive over thousands of cases.
My favorite small nudge example is pill bottles vs "blister" packs. Having pills in blister packs has shown to have a significant effect on reducing suicide as it makes it much harder to get the number of pills needed. Yes, some people will still succeed but there is a segment of the target population who will give up after a couple pills.
But medical school is going to heavily select for conscientiousness (a personality trait that checklists work with), and also checklists are essentially another authority in the room.
So, yes, you could potentially have discussions over whether or not something has actually been done. But you aren't going to have discussions about whether it should be.
If a junior person thinks that a senior person is about to make a terrible mistake, then the word "Stop!" should be used, along with the reason. That word will often make even the most arrogant person pause and think about what they are doing.
Expertise and practice makes some of those things automatic, but not enough to reduce it all below 5-7. Further, in any surgery there will be other things going on that require cognitive attention, decreasing capacity for other things.
Everything you can do to reduce cognitive overhead makes a process smoother.
I teach the LSAT. One section logic games (officially, analytical reasoning) tests precisely this cognitive load. Students must work with 4-6 rules + what the situation calls for on question.
The rules are impossibly simple. But, in the heat of things, students just aren't capable of working with that many items unless they create a structure using diagrams. And even seemingly tiny efficiencies have an outside effect on speed and correctness.
Don't forget that these surgeons are often tired busy and stressed, three factors that worsen performance. Having a clear list that says "do this now dummy!" massively helps keep you on track even when you're a wreck.
It's like help desk asking you if it's plugged in. You may think of yourself as a technical person and that the problem is more complex like the driver isn't loaded or the monitor isn't properly configured but the problem could just be that you forgot to plug the video cable in. Just having those basic checks occur before anything can go wrong makes it so that when things do go wrong, time doesn't need to be wasted on going over the simple stuff or even worse, forgetting about the simple stuff and going down the wrong route.
Because it didn't. It attributes the entirety of a decade long decline to the check list, which is obviously nonsense. I'm not saying that checklists aren't good, but they're not miracle workers.
>After testing, review and feedback from health boards across Scotland, the surgical checklist was included as one of the ten Patient Safety Essentials to be implemented across all health boards in Scotland37. The surgical checklist was not a stand‐alone intervention. It was, however, the only Patient Safety Essential that targeted surgical patients specifically during the interval studied37. Thus, the addition of the checklist to the other parameters within the SPSP may have contributed to the improvement in results observed in the present study.
>It is acknowledged that attributing causal links to the findings in population‐wide data set analysis is not possible.
What you said they didn't do is kind of exactly what they did. Yes, there is some attempts at hand waving, but it's silly to boil down the success of a large and multi pronged program to a simple checklist.
That's hardly hand-waving.
Here's the 10 safety essentials check list they are talking about:
If the checklist is responsible for all of the improvement in surgical outcomes then we can stop doing the rest of the stuff, right? No need to emphasize hand washing or sterile central line because the magic checklist will stop the infection.
Here's a link to the study.
Please point out the parts of it that you find to be scientifically failing.
There's a valid point to be made that the original article implies a greater causal link than the study, but that isn't the study's fault, and it doesn't actually diminish the effect of the checklist in the process. Poor reporting on science is always a problem, and journalists typically use layperson phrasing in order to make the information more consumable.
Article on checklists in the New Yorker
Death rate after fix: 0.46%
Number of deaths after:93
Death rate before fix: 0.46/(1-0.37) = 0.73%
Number of deaths before: 147
Number of lives saved per year: 54.
Doctor - "Smith, please provide checklist for Septal Myectomy"
Smith - "Confirming list for Septal Myectomy?"
Doctor - "Yes"
(Smith raises first point)
(Smith raises other points until the end of the list)
Smith - "You have completed the checklist for Septal Myectomy, please review it on the screen if you so desire. Is there anything I can help you with?"
Doctor - "Thanks Smith, no need for anything else"
I would be eager to support a project that would provide this kind of support to doctors and other professionals.
My favourite ones are for vacations... My wife and I will be taking our son (who will be 5 months old at the time) on his first trip that will require long haul flights (Australia - Vietnam - UK) so I'm currently working on a checklist for what to have in our hand luggage, and a separate one for the hold luggage. For me, it's part of the looking-forward aspect to such a trip, plus there are lots of things that we really don't want to forget, or run out of.
And anyone who's scuba dived will be familiar with the three item checklist that forms part of the pre-dive "buddy check".
* B A R (buoyancy, air, releases) in the UK
* A B C (air, buoyancy, clips) in the US and probably elsewhere
A potential life saver!
The problem I have twofold. A few years down the road will people still really follow the checklists religiously, or quickly just fill out the boxes.
The second part is people's jobs get dumbed down by this, I'm sure we've been in some situation where there is a structured system which you have no control, you just fill out forms, its disempowering so you learn not to think, just do your part of the system and hope it turns out OK. Its where bureaucracy and red tape starts. I've quit jobs like that because I found it demoralizing.
Will be interesting if there was a follow up 5 years later.
EDIT: Didn't you get the memo about your TPS reports? That is what I'm trying to avoid.
What people call bureaucracy is process that they follow without knowing why it's there. Every single rule in a business is there because something went wrong, and someone didn't want it to happen again.
The answer to bureaucracy is transparency. If you explain why a rule has to be followed people don't mind that there's a rule.
(You also have to regularly review the rules and get rid of ones that don't make sense any more, but that's much less common that you'd think.)
However, I want it "dumbed down". I want it to be easier to prevent mistakes. I want smart people to use more of their brainpower on things that can't be solved by some words and boxes on paper. If the surgeon is thinking about the surgery instead of what he might have forgotten about the pre-op, that's good, IMO.
Agreed. The only issue is that if the surgeon isn't thinking about the surgery because he's still doing all the routine paperwork.
I completely understand why this took effort to get implemented and why it seems obvious in retrospect.
The book also talks about the design of checklists which is not trivial at all.
Checklists I have seen used during my career are often rushed through, with more time usually being spent appropriate to the complexity of a situation. Local culture will strongly impinge on effectiveness of (yet another?) checklist. The safety culture in medicine is improving dramatically, but there is tension between protocolizing everything into checklists, and individualizing care to each patient. AI will be part of the solution to this.
An important checklist commonality in medicine is the request that _anyone_ should speak up if they think something is wrong. In analyzing situations where patients were harmed, e.g. wrong side surgery, it is common to hear that someone, a nurse, a technician, did notice, but was not empowered to speak up. As a doctor, I am used to the power to speak up and act, but I value others' eyes in many situations. More complex and highly dynamic medical situations will always require the attention, knowledge and skill of those involved.
I was hoping to get that info because the first thing that popped into my mind was that people are/were apparently dying due to unnecessary reasons if a simple checklist prevents 37% of surgical deaths. If that's the case, the only way I could imagine classifying those deaths would be "gross negligence" and therefore not subject to the protections of the legal agreements you sign before surgery.
I can talk a little bit about England.
There are two main ALBs (arms length bodies) that will be involved: NHS Resoulation (the organisation that handles legal cases) and NHS Improvement (the organisation that handles QI work). NHS England and NHS Improvement are merging and I don't know what the new name will be. NHS Resolution used to be called NHS Litigation Authority.
The information that NHSi has about "Just Culture" is here: https://improvement.nhs.uk/resources/just-culture-guide/
If you have a look at this flow-chart you can see that they're trying to find out if an incident that caused harm was deliberate, grossly negligent, caused by wider system failings, etc. https://improvement.nhs.uk/documents/2490/NHS_0690_IC_A5_web...
If you have a look at NHS Resolution's page about learning from harm you can see that they're keen for healthcare professionals to 1) Say sorry, 2) explain in full what went wrong 3) Explain how that's going to be prevented in future. https://resolution.nhs.uk/services/safety-and-learning/
This expands upon a legal duty of HCPs and NHS Trusts in England: the Duty of Candour.
Here's the advice for doctors: https://www.gmc-uk.org/ethical-guidance/ethical-guidance-for...
And nurses: https://www.nmc.org.uk/standards/guidance/the-professional-d...
And other registered healthcare professionals: https://www.hcpc-uk.org/assets/documents/10003f72enc07-dutyo...
And organisations: https://www.cqc.org.uk/guidance-providers/regulations-enforc...
If a patient does decide to sue I think they can only recoup their actual losses. I think we don't have punitive damages in England.
As a software engineer, I'd hope this was done in software, where it could be trivially filled in by multiple people, checked across the organisation, not go missing or get drinks spilled on, etc.
Knowing the NHS south of the border, I assume it's not.
For example when you are flying a cesnna 172 from the nineteen sixties, you are basically flying a glorified lawn mower. It has procedures for starting up. Not following those will lead to disasters like your engine quitting, or worse self destroying. Hence a lot of checklists in aviation tend to be about engine management. These are important, they keep you alive and save you from high maintenance cost and accidents.
Cars used to be similarly complicated but these days you simply turn them on and drive. Likewise, checklists for modern airplanes that have elaborate engine management software tend to be a lot shorter and also focus more on other things (check the amount of fuel).
Checklist are easy to automate for a lot of use cases and doing so increases safety because computers are a lot better at checking lots of things than we are. After checklists become part of your process and increase your effectiveness, automating them can increase your efficiency.
The analogy in software engineering is replacing release processes with CI and CD. It used to be that people were enduring such things as commit freezes, meetings about blocking issues, go/no-go decisions, etc. This could last for weeks or months. Also the process of actually creating release artifacts and publishing those was a lot of work. These days, some systems release whenever a change is merged to a particular branch. All the checks around this are automated (tests pass). The entire process of releasing can be automated.
I'm surprised how little checklists are used in software development. In our application, checklists track even simple things. Suppose you're adding some new feature we have a checklist that tracks related activities such as:
- Does it need a toolbar button?
- Does it need a keyboard/mouse input biding?
- If so did you add a keyboard binding to the default configuration list?
- If so did you add the binding information to the user help file.
- Should it be exposed in Mobile Native App?
- Should it be exposed in Mobile HTML App?
- Should it be exposed in HTML App?
- Was the button label localized?
- Was the identifier added to the toolbar mask list?
- Does it need to be added to the context menu list?
- Does the behaviour need to work with Reset?
- Does the behaviour needs to work in a collaborative session?
- Does the button need a feature flag?
It's a quick list that we can get through during feature kick-off before unpacking commercial requirements and starting a sprint and during final validation to make sure al the i's are dotted and t's crossed.
I use one for travel (did I pack a neck pillow?), night-time (did I pack my gym clothes for tomorrow?), and recently for the gym (what workouts am I doing today?)
Any surgical staff around to tell us more?
As a software developer, while not having checklists as printable documents, I highly appreciate the checklist approach. This can take several shapes. When doing a software release, I perform the necessary steps always in the same order. This makes it easier to correctly perform everything needed. And of course I check the result of every step for the expected result.
Checklist-like approaches can be even very helpful when debugging issues. I often enough get to resolve "the software doesn't work" like issues. And of course, there is no obvious reason, like a meaningful error message or stack trace. To debug this kind of issues, I start with simple and trivial things. While not a static list - the cases are too varying for that - I tend to start with asking and checking very basic questions, like: is there enough free disk space? Repeating the steps which, according to the bug reporter, lead to the issue. At each action taken, verify that any observeable result matches the expectation, even if those actions are not close to the bug. This approach has usually two outcomes. Either a very trivial explanation for the problem is found, because some seemingly trivial step fails, or at least, I know, that all the things tested are note related to the problem. Excluding a wide range of possible problems can be surprisingly helpful in narrowing down the actual problem.
* have the unit tests been built and run?
* have I reviewed all changes and removed commented out code, tidied up?
Edit: oh, you meant apps. I read your comment as asking for good items for checklists.
I'm surprised that the researchers didn't use better data or variation (e.g., hospital-level mortality or the timing of the rollout) to strengthen the case. So much can be lurking beneath annual country averages...
and other languages
I find the approach of things like rail and air accident investigations to be really useful. The goal for the investigation of any accident should not be about apportioning blame, but understanding why accidents happen and how their occurrence can be effectively reduced or eliminated.
For 2016/2017 the summary of "never events" is here: https://improvement.nhs.uk/documents/2347/Never_Events_1_Apr...
There were 189 wrong site surgeries, and 114 retained foreign object post surgery.
That's not good enough, but it's not terrible.
The month by month data is here: https://improvement.nhs.uk/resources/never-events-data/
I have a comment about the regulatory framework (in England) here: https://news.ycombinator.com/item?id=19683366
To catch errors, I started double-checking math problems, by generating a checklist on the spot.
But I also made fewer errors, and developed more intuition. I think checking itself was a deeper practice; and also gave an overview of the task (less cluttered by detail) that helped me make connections at a higher level.
Checklists: insightful and tedious!
Referring to things in multiple ways is supposed to have a more thorough confirmation, although not a good idea to point whilst wielding a scalpel.
A checklist is a practical tool to document processes.
This news is that we've now run research across an entire country (Scotland) for the surgery carried out between 2000 to 2012 to find out how much harm has been prevented.
The problem with relying on tests written after the implementation is that the checklist might be incomplete. If your checklist is incomplete, then you can't rely on it. Of course, even a highly disciplined practitioner of TDD cannot produce a test suite that is guaranteed to be complete and correct, but there is an enormous difference between a checklist that is adequate 50 % of the time and one that is adequate 95 % of the time.
Look for ways to avoid simply "going through the motions". Checklists (including TDD) are about avoiding making mistakes, not merely about noticing when you do make mistakes.
I heard of the experiences from countless others and, sadly, encountered it myself. As the patient, even though you are paying for the service through high taxes, you are not seen as the customer. This leads to a certain sense, when you do need a medical service, that they are doing you a favor, that you are somehow the recipient of charity and should be grateful for what you get.
Decisions about what medicines or treatments are available are often political, with certain high-profile conditions sucking up scarce resources at the expense of others. The focus is very much on managing public opinion.
There are many situations in which cost considerations have a horrific impact on lives. For instance, if there is a medicine which can prevent you losing your sight, but it is expensive, you will be offered it only for your second eye after you have lost sight in your first - the reasoning being that it is only worth spending that much money to prevent total blindness, but sight in one eye is enough.
If you think that your high taxes mean that your healthcare needs are covered, think again.
There is also a deep-rooted coverup culture that circles the wagons around bad doctors and poor processes. In my case, a ridiculous misdiagnosis had a real impact on my life for over a year. The other healthcare professional only came clean about it after the lead doctor had retired.
Again, you are not seen as the customer, as the one paying all their wages, so, you should just shut up and be grateful for what you get.
I often laugh when I hear inexperienced American talk about how much better the health system is in the UK. Sure, health insurance is expensive, but the actual healthcare is leagues ahead of anything available via any sort of national health service. Being recognized as the customer, with real rights, is of pivotal importance in receiving the care you need, when you need it.
In fact, you often come across UK citizens with a rose-tinted view of the National Health Service, but such opinions tend to change rapidly once you actually need something more than an occasional General Practitioners appointment. The whole thing is a cruel joke.
Except if your poor though right? Isn't that really the case?
In the UK we don't tend to ignore kids broken bones if they have poor parents. No one goes bankrupt and ends up homeless for contracting an illness, or having an accident at work.
We also spend less on our taxes towards the NHS than Americans spend on their Medicare - and then you have to pay for private 'health care' insurance on top, including all of the 'co pays' and whatever. It's a system that's rigged against you. For the rich, by the rich, to make the rich richer.
Even if the top 1% of private healthcare is better in the USA, you're ignoring the 99% of healthcare that isn't.
Most Americans just can't see the simple fact - many many other countries are better at this than you are. This is a solved problem in many other developed countries.
Universal healthcare simply benefits everyone in society, and does so purley for the common good.
Kids with broken bones are at the easy end of the scale. Most people, before they have experienced serious, complicated, expensive health issues, have a series of relatively positive experiences with the NHS in which they present with a minor issue that is dealt with by in a satisfactory manner. The system is optimized for that. This maintains the illusion that, through your taxes, your health requirements are "covered".
This falls apart once you present with something more complicated. You might be surprised at the extent to which people in the UK are pretty much abandoned when they hit a certain point. I know of many cases in which people had to find the money, somewhere, to buy vital treatment that the NHS was not willing to allocate. This is no less brutal than what happens to the uninsured in the US ... but ... the big difference is that UK residents are under the delusion that they do not need to worry.
Again, my argument is not that the NHS is bad, or evil, or that it does not sometimes do a good job. What I am saying is that, over your lifetime as a whole, it costs more than good health insurance would while letting you down badly when, inevitably, you encounter more serious problems.
The best way to help the poor is to improve the economy. A massive, complicated tax system makes it particularly hard to hire people at the low-end of the skill range. It is simply not worth it, and this will become increasingly evident as workers continue to be replaced by technology.
The UK has ended up with a multi-generational unemployed class who are under the illusion that everything will be taken care of "from cradle to grave". Their personal agency has been almost entirely removed. Alcoholism, obesity, diabetes and depression are at epidemic levels among the unemployed. We have created the perfect storm and it is the poor who will bear the brunt of that.
Exactly the same private healthcare is available to you in Scotland, if you chose to pay for it.
My argument is that people are _already_ paying but in a mandatory, indirect way, that removes their power as consumers and creates a situation in which the health services' prerogatives often differ from those of the patients.
"... some of the largest population-wide reductions in surgical deaths ..."
_reductions_. but how does it compare using absolute rates? it's usually easy to achieve good numbers when the baseline is sub-par (diminishing returns).