If hospitals were run like startups, they'd advertise free advanced medical procedures to draw people in, then the moment they'd find a buyer for the hospital they'd shut the power down and told dying people in mid-operation that they thank them for joining and being part of this great success.
Seriously, the tech startup model isn't always the best one. Here are some more points:
11. If hospitals were run like startups, 10% of emergency patients would be turned away ('not the target audience'), and any medical staff who complained would be labeled "insufficiently pragmatic".
12. If hospitals were run like startups, 90% of them would close their doors within the first year of opening.
13. If hospitals were run like startups, patients' medical information would be used to sell them ads. The databases storing this information would also be leaked within the first 10 years of opening, since details like security are rarely a priority in a startup.
14. If hospitals were run like startups, junior medical staff would be overworked, underpaid, and burnt out. Actually, this is already the norm (at least in Canada).
15. If hospitals were run like startups, patients who went in to have their appendix removed would find that their tonsils and wisdom teeth had also been removed, and an RFID chip with their patient ID had been implanted "to serve them better". This would be OK because it would be covered by the blanket consent form referred to by the sign at the entrance that reads "by using this facility, you agree to its terms and conditions (available upon request; may change at any time)". Patients who had read it would know that they could opt out.
It's also an unfunded mandate. The government declared that all patients must be seen, but did not provide a mechanism to pay physicians, nurses, and hospitals for their services to care for those patients. They also did not shield the clinicians from liability. Hell, in order to perform a "medical screening exam" that EMTALA mandates, you have to do everything required to determine there is no true emergency. If you determine there is no emergency and discharge the patient (with no compensation), then the patient has something happen in the parking lot, you are 100% liable. As a result, people get these frighteningly expensive workups in the ED out of fear for litigation. People should get turned away if they don't have an emergency, but the fact is, no one turns anyone away, because there is no incentive to.
Now you can trace a direct link between passage of EMTALA to the over utilization of the Emergency Department.
15. happens more and more often too. I know at least two different docs who "overdiagnosed" or were force to overdiagnose in order to rush people to unnessary surgeries, i.e. fill the op calendar.
Wow. Talk about missing the point. He's obviously not talking about how the whole hospital is run. He's talking about using selective technology to make certain things run better.
16. If hospitals were run like startups, ER patients would be treated by college dropouts who taught themselves surgery by watching videos of mom's stitching their kid's dolls together.
I have worked IT in hospitals. I agree with the actual point the original article is trying to make - hospitals are not doing technology well at all. I don't want to bash the article too much, because that point is correct.
But there is an extremely important factor here - patient health. No technological improvement is worth risking the health of a patient.
This is a core business principle at work - the point of most businesses is not to have the best technology. Technology is a tool that often helps them achieve their actual goals better. Just a tool. Technology for its own sake is usually a mistake, and very much a mistake if it risks the health or life of any person.
I think the best point the article makes is regarding the patient records systems. It sounds like the "best" systems out there were developed with the same sensibilities that leads to Enterprise and QA tracking systems with more than 100 fields per screen, 90% of which are useless in any given installation.
Basically there's a ton of room for improvement, and the reason the improvement never happens is that the people who evaluate the systems are never the same people as the users. Moreover, the people who evaluate the system have no idea how to do anyone else's job, and so they have no idea how much inefficiency they're saddling their underlings with.
> No technological improvement is worth risking the health of a patient.
What about the time spent on 350 clicks? Do that 20 times a day. How many hours of sleep did that intern lose to clicks?
As a physician who writes code, and is on the hospital's informatics board, his main point is a massive drag on medicine. And the problem is very much on the IT side. The hundreds of clicks are the problem.
People who write this code really should be made to write that code, at least sometimes, maybe one day a week, on the floors. Just sit at the nurse's station and try to do abstract, thoughtful work in a near panic. Oh, sounds terrible? Yes, yes it is.
The people who buy this software, quite frankly, I have no idea what to do about them. It makes me sick.
No technological improvement is worth risking the health of a patient.
I disagree, least of all because the health of the patients should be put ahead of the health of a patient.
The fact is that this unreasonable fear of harming patients has caused everyone to be completely blind to the routine fuckups and inefficiencies which, when taken together, doom the industry and patients as a whole.
Let's look at EMRs. Epic, for example. Integrating with Epic is pretty hard--unless the client really twists their arm, they basically won't say anything to help you with their software or access their piles of shitty databases (mumps lol). They get paid tens or hundreds of millions of dollars a year (per hospital!) to handle the task of not losing patient records, a task at which they semi-succeed. They have no incentive to change. Oh, and one more thing...hospitals are deathly afraid of having data duplication anywhere, so they freak out if you start looking like you're a sibling system to their EMR instead of just blindly consuming it.
Moreover, they--like many smaller vendors--do extreme customization work that basically guarantees a new learning curve everytime somebody moves to a new system. And the software looks like garbage, and is hard to use, but at least lives aren't on the line and nobody was risked being harmed by some ~crazy new startup EMR~. Hah, oh wait, they are on the line, and people die due to bad medication decisions that aren't flagged or are flagged and ignored by the system.
And this whole system? Most of the mess is geared towards solving the billing problem. It isn't geared towards the patients records problem, towards the "what did this patient's body do at this time" problem. It isn't geared towards the "hey, that's odd, all my lab messages have the same jank-ass free-text formatting instead of an interface from the 90s".
The fact is that the vendors and hospitals have conspired in their fear and greed to basically raise the barrier to entry so high that iterating on new designs to help is quite hard, and instead just a good way of giving yourself an aneurysm. And you don't want that, because by that time you'll want to step nowhere near a hospital.
The state of ERP and MES systems is very similar to what you describe. I had a brief chance to work with a few of them, and between the bugs, unimplemented features and the UI design straight from Necronomicon, I have no idea how factories even work, much less are able to produce and deliver quality goods reliably.
While I agree with many of the ideas in principle, I'm wondering if startups are really the best benchmark to compare with. Much of what the article states might be closer to the ideals of Lean-driven firms such as Toyota. In fact, many successful attempts have already been made to bring these principles into healthcare. Without questioning the effectiveness of startups, I doubt that all of their organizational principles are well suited for a context where shortcuts and experiments can lead to real harm.
Maybe there's a good startup in all this, Medical CTO Outsourcing to modernize or at least fix UX/UI's in healthcare, and also fix performance and optimize workflows.
If hospitals were run like startups, they would ship a 1.0 product as soon as possible, tolerating the embarrassment it causes and not worrying about the people they killed.
A better model is aviation, i.e. an industry where failure is often catastrophic. Look at the success of jet engine manufacturers in producing product that does not fail. Trans-oceanic flights on just two engines is something that would have been considered folly 30 years ago. It's all well and good for Southwest Airlines to have super-perky flight attendants that innovate by singing songs during post-landing announcements, but the real core of their business is safety, just like the real core of medicine is treatment outcomes.
Interestingly enough, in aviation you do see checklists with dozens or even hundreds of items that need to be "clicked" every time, and incredibly complicated UIs that take a huge amount of training to be able to use effectively.
Pretty disappointed by the derisive comments in this thread. For a bunch of startup enthusiasts and innovators, we sure do seem to like status quo. Or maybe it's just a contrarian signalling thing? Sorry for going meta.
Contrarianism is a bit strong. I'd say bullshit detector. It's a natural response to the original author blindly praising one method while ignoring all its flaws.
He's not all wrong. I work in healthcare and can absolutely relate to most of his complaints.
Paging systems are absolutely archaic and I've had important pages simply never be delivered many times. Bed tracking software at my current hospital has multiple full time employees whose job is 90% looking over things and sending support staff to their next task via a page. Callback numbers that take priority over any other task but then they won't pick up which just wastes time for short concise directives. Manual charting that's made incredibly tedious by poor UIs.
Hospitals shouldn't be run like startups, but they desperately need better integration with technology.
Haha, I am a good friend of Omar's and the referenced physician-coder in the story. While he made it sound as though I had avoided healthcare altogether since graduating, I really did put in the effort! Part of me felt like I was obligated to put in some time given I spent half a decade in my 20's learning medicine. I worked for Listrunner in the past year and had a lot of fun combining my software development talents and medical knowledge.
Well it's a good analogy on the information technology front, highlighting the deficiencies. But a hospital doesn't have the same solitary focus that a start up has.
The main point is that hospitals are political organisations. And like political organisations, large scale change only comes after deepening crisis, or leadership overhauls. Those are the only two ways. Efficiency and speed are not even on the list of things a hospital is trying to optimize.
Hospitals need a "SpaceX" - some company that comes in that's fairly well funded and capable, starts out small and proves that there's a lot to gain from using a different process than the typical ones currently employed. This would probably mean instead of selling software to a hospital, you'd make your own hospital - not exactly a startup project.
The main problem in healthcare IT, imho, is the barrier to entry for anyone unfamiliar with the existing systems and regulations. Writing software for the industry is so daunting that even some of the big boys won't touch it.
For instance, in the US, if someone hacks your user's credit card data there will be some penalties involved, but life will continue. But if someone hacks your patient's records in your healthcare system and you didn't follow the rules regarding security? Up to $25k per infraction. If you did it willingly you're in criminal territory and the fines get much larger.
And that's just the regulations in the US. Every other country has it's own regulations you must follow.
Fortunately things are changing. Politicians are clarifying and updating laws to accommodate new technology. The tools are getting better. And startups like Dr. Chrono are slowly chipping away at the entrenched enterprise. But it's going to be a slow process. After all, it's a very serious business.
I'm going to shamelessly plug http://verbal.care/ here. They started out building a tool for aphasia patients to communicate with hospital staff and have pivoted to the more general mission of improving the UX of intra-hospital communication.
Here are my 0.02USD on the situation. I lead an engineering team at a health IT startup doing realtime physiological monitoring and operations analytics for clinical use. I have seen our software through FDA certification in record time. I have glimpsed the madness.
dlitz and others are playing the Waldorf-style heckling from the sidelines "doh ho ho, hospitals-as-startups would shave ads into your scalp". That's real cute. The fact is, hospitals as they've evolved are super gnarly behemoths. They have this insanely conservative culture at all levels...nobody wants to lose their job, and especially in IT, the successful ones have basically made it their life's mission to prevent physicians (who are, generally, pretty smart and freakishly stubborn people) from being overzealous and doing stupid things like spewing patient data over cell phones. They still do this.
Any change whatsoever in their operating environment takes months, because they are completely optimized to prevent systematic change from occurring. There is paperwork and lawyers and billing the likes of which you've never seen. Sales cycles are over 18 months. 18. Months. You could conceivably have two full children at a hospital before they have approved the project you'd use to pay for same.
The technology in hospitals is a bad joke. They cannot reliably tell you the status of a patient, what department they're in, or where in the hospital they've been. Frequently, updates to patient details will come out saying "Oh, whoopsie...this patient? They really are this other patient...wait, shit, those are the same patient." Nurses and docs will occasionally just horde equipment in their offices to guarantee that their department has what it needs, and the hell with the rest. It's absurd. They thought, for a long time, that no nurses would ever be interested in texting results around--and now, guess what, in the handful of places rolling that out, it's been like the best thing ever.
The people? Hoo boy. The nurses are awesome, but turnover and burnout tends to be large, least of all because of how shitty the alarm management situation is in certain ICUs. The docs? Some pretty great people, but getting product feedback from them is basically like calling rand() over and over. They also basically are in fire-fighting mode all the time. So, they'll do nothing to help you until they need something yesterday, and then magically their bitching will open up administrative doors that previously you didn't even know existed.
And that brings us to the administration. Some folks "get it". Mostly, though, especially in IT, they don't. They started their careers as machine operators, and with the awesome application of seniority, they rise through to their maximum level of incompetence. Things like billing 10K for 10gb of network storage on a VM. They only buy big name software, and that's how the deathgrip of Epic happens.
And if you could just bribe these folks to get shit done that'd be one thing--but hospitals (correctly) tend to have these super draconian gift policies, to the extent of not even being able to buy like a ham sandwich for a client. The hospital systems sometimes employ technically incompetent people, but always also close the backdoors we'd use to work around them.
Let's assume you find some docs and nurses on your side. Let's assume you find some IT folks that know how to Get Shit Done, or at least stay out of your way (you know, instead of taking two months to open a fucking port to let something work). Well, now, it's off to the fucking races with legal and billing.
Why? Because legal will have somebody look at a contract, go over it, send it back, and then have somebody else look at it, back and forth and back and forth, take vacations, sometimes just fucking lose the contracts, and then if anybody anywhere made a revision, the process resets. We've had some SLAs and MLSAs held up nearly years because of this. A...
Absolutely nailed it, and a great piece of insight. The bureaucracy and technical incompetence in hospitals is just agonizing to the point where I'm not sure you can even fix them.
I'm increasingly convinced that healthcare needs a SpaceX style makeover. Someone with a few billion dollars that can fund the top to bottom construction of a new state of the art facility and fund the software and medical equipment development to get things running how they ought to.
If hospitals were run like Silicon Valley startups... I would quit my job and become a plaintiff's attorney, and I would get rich as fuck suing these motherfuckers.
If hospitals run like startups they would not focus on software. That is not what they are selling and if they are using money wisely they would get the cheapest possible software and the meets the requirements and instead focus on providing better health care. Which is what they do.
> Each “progress note,” of which physicians can write two dozen or more each day, takes about 250 mouse clicks to navigate (not including the textual content of the note!).
Man, that is quite impressive, must be some kind of a world record? I would really like to see that app :)
Hospitals involve capital investment, human resources, and exposure to liability unlike any startup can imagine. Thinking they will ever run like startups is pretty much an impossibility. Even startups that get big don't operate like startups anymore. If they have problems with IT operations, the solutions are not going to be found by telling them to operate like startups.
50 comments
[ 2.7 ms ] story [ 99.6 ms ] thread11. If hospitals were run like startups, 10% of emergency patients would be turned away ('not the target audience'), and any medical staff who complained would be labeled "insufficiently pragmatic".
12. If hospitals were run like startups, 90% of them would close their doors within the first year of opening.
13. If hospitals were run like startups, patients' medical information would be used to sell them ads. The databases storing this information would also be leaked within the first 10 years of opening, since details like security are rarely a priority in a startup.
14. If hospitals were run like startups, junior medical staff would be overworked, underpaid, and burnt out. Actually, this is already the norm (at least in Canada).
15. If hospitals were run like startups, patients who went in to have their appendix removed would find that their tonsils and wisdom teeth had also been removed, and an RFID chip with their patient ID had been implanted "to serve them better". This would be OK because it would be covered by the blanket consent form referred to by the sign at the entrance that reads "by using this facility, you agree to its terms and conditions (available upon request; may change at any time)". Patients who had read it would know that they could opt out.
It is illegal because it was a frequent occurrence, and, while less frequent, it (patient dumping) remains a serious problem.
It is a major problem (both independently, and as a symptom of other major problems) today.
Its less of a major problem today than it was before it was outlawed.
Now you can trace a direct link between passage of EMTALA to the over utilization of the Emergency Department.
http://www.dispatch.com/content/stories/local/2014/12/30/ohi...
Sad, but true.
Forced by who? The management? Or did they do it themselves for their own benefit?
I have worked IT in hospitals. I agree with the actual point the original article is trying to make - hospitals are not doing technology well at all. I don't want to bash the article too much, because that point is correct.
But there is an extremely important factor here - patient health. No technological improvement is worth risking the health of a patient.
This is a core business principle at work - the point of most businesses is not to have the best technology. Technology is a tool that often helps them achieve their actual goals better. Just a tool. Technology for its own sake is usually a mistake, and very much a mistake if it risks the health or life of any person.
Basically there's a ton of room for improvement, and the reason the improvement never happens is that the people who evaluate the systems are never the same people as the users. Moreover, the people who evaluate the system have no idea how to do anyone else's job, and so they have no idea how much inefficiency they're saddling their underlings with.
They probably also don't care.
What about the time spent on 350 clicks? Do that 20 times a day. How many hours of sleep did that intern lose to clicks?
As a physician who writes code, and is on the hospital's informatics board, his main point is a massive drag on medicine. And the problem is very much on the IT side. The hundreds of clicks are the problem.
People who write this code really should be made to write that code, at least sometimes, maybe one day a week, on the floors. Just sit at the nurse's station and try to do abstract, thoughtful work in a near panic. Oh, sounds terrible? Yes, yes it is.
The people who buy this software, quite frankly, I have no idea what to do about them. It makes me sick.
I disagree, least of all because the health of the patients should be put ahead of the health of a patient.
The fact is that this unreasonable fear of harming patients has caused everyone to be completely blind to the routine fuckups and inefficiencies which, when taken together, doom the industry and patients as a whole.
Let's look at EMRs. Epic, for example. Integrating with Epic is pretty hard--unless the client really twists their arm, they basically won't say anything to help you with their software or access their piles of shitty databases (mumps lol). They get paid tens or hundreds of millions of dollars a year (per hospital!) to handle the task of not losing patient records, a task at which they semi-succeed. They have no incentive to change. Oh, and one more thing...hospitals are deathly afraid of having data duplication anywhere, so they freak out if you start looking like you're a sibling system to their EMR instead of just blindly consuming it.
Moreover, they--like many smaller vendors--do extreme customization work that basically guarantees a new learning curve everytime somebody moves to a new system. And the software looks like garbage, and is hard to use, but at least lives aren't on the line and nobody was risked being harmed by some ~crazy new startup EMR~. Hah, oh wait, they are on the line, and people die due to bad medication decisions that aren't flagged or are flagged and ignored by the system.
And this whole system? Most of the mess is geared towards solving the billing problem. It isn't geared towards the patients records problem, towards the "what did this patient's body do at this time" problem. It isn't geared towards the "hey, that's odd, all my lab messages have the same jank-ass free-text formatting instead of an interface from the 90s".
The fact is that the vendors and hospitals have conspired in their fear and greed to basically raise the barrier to entry so high that iterating on new designs to help is quite hard, and instead just a good way of giving yourself an aneurysm. And you don't want that, because by that time you'll want to step nowhere near a hospital.
A better model is aviation, i.e. an industry where failure is often catastrophic. Look at the success of jet engine manufacturers in producing product that does not fail. Trans-oceanic flights on just two engines is something that would have been considered folly 30 years ago. It's all well and good for Southwest Airlines to have super-perky flight attendants that innovate by singing songs during post-landing announcements, but the real core of their business is safety, just like the real core of medicine is treatment outcomes.
Paging systems are absolutely archaic and I've had important pages simply never be delivered many times. Bed tracking software at my current hospital has multiple full time employees whose job is 90% looking over things and sending support staff to their next task via a page. Callback numbers that take priority over any other task but then they won't pick up which just wastes time for short concise directives. Manual charting that's made incredibly tedious by poor UIs.
Hospitals shouldn't be run like startups, but they desperately need better integration with technology.
That is a perfectly logic statement that I'm sure most folks would agree with.
However, a blog title like that wouldn't get many clicks, would it?
The main point is that hospitals are political organisations. And like political organisations, large scale change only comes after deepening crisis, or leadership overhauls. Those are the only two ways. Efficiency and speed are not even on the list of things a hospital is trying to optimize.
Not sure of a fix. The best I can think of is rankings.
For instance, in the US, if someone hacks your user's credit card data there will be some penalties involved, but life will continue. But if someone hacks your patient's records in your healthcare system and you didn't follow the rules regarding security? Up to $25k per infraction. If you did it willingly you're in criminal territory and the fines get much larger.
And that's just the regulations in the US. Every other country has it's own regulations you must follow.
Fortunately things are changing. Politicians are clarifying and updating laws to accommodate new technology. The tools are getting better. And startups like Dr. Chrono are slowly chipping away at the entrenched enterprise. But it's going to be a slow process. After all, it's a very serious business.
dlitz and others are playing the Waldorf-style heckling from the sidelines "doh ho ho, hospitals-as-startups would shave ads into your scalp". That's real cute. The fact is, hospitals as they've evolved are super gnarly behemoths. They have this insanely conservative culture at all levels...nobody wants to lose their job, and especially in IT, the successful ones have basically made it their life's mission to prevent physicians (who are, generally, pretty smart and freakishly stubborn people) from being overzealous and doing stupid things like spewing patient data over cell phones. They still do this.
Any change whatsoever in their operating environment takes months, because they are completely optimized to prevent systematic change from occurring. There is paperwork and lawyers and billing the likes of which you've never seen. Sales cycles are over 18 months. 18. Months. You could conceivably have two full children at a hospital before they have approved the project you'd use to pay for same.
The technology in hospitals is a bad joke. They cannot reliably tell you the status of a patient, what department they're in, or where in the hospital they've been. Frequently, updates to patient details will come out saying "Oh, whoopsie...this patient? They really are this other patient...wait, shit, those are the same patient." Nurses and docs will occasionally just horde equipment in their offices to guarantee that their department has what it needs, and the hell with the rest. It's absurd. They thought, for a long time, that no nurses would ever be interested in texting results around--and now, guess what, in the handful of places rolling that out, it's been like the best thing ever.
The people? Hoo boy. The nurses are awesome, but turnover and burnout tends to be large, least of all because of how shitty the alarm management situation is in certain ICUs. The docs? Some pretty great people, but getting product feedback from them is basically like calling rand() over and over. They also basically are in fire-fighting mode all the time. So, they'll do nothing to help you until they need something yesterday, and then magically their bitching will open up administrative doors that previously you didn't even know existed.
And that brings us to the administration. Some folks "get it". Mostly, though, especially in IT, they don't. They started their careers as machine operators, and with the awesome application of seniority, they rise through to their maximum level of incompetence. Things like billing 10K for 10gb of network storage on a VM. They only buy big name software, and that's how the deathgrip of Epic happens.
And if you could just bribe these folks to get shit done that'd be one thing--but hospitals (correctly) tend to have these super draconian gift policies, to the extent of not even being able to buy like a ham sandwich for a client. The hospital systems sometimes employ technically incompetent people, but always also close the backdoors we'd use to work around them.
Let's assume you find some docs and nurses on your side. Let's assume you find some IT folks that know how to Get Shit Done, or at least stay out of your way (you know, instead of taking two months to open a fucking port to let something work). Well, now, it's off to the fucking races with legal and billing.
Why? Because legal will have somebody look at a contract, go over it, send it back, and then have somebody else look at it, back and forth and back and forth, take vacations, sometimes just fucking lose the contracts, and then if anybody anywhere made a revision, the process resets. We've had some SLAs and MLSAs held up nearly years because of this. A...
I'm increasingly convinced that healthcare needs a SpaceX style makeover. Someone with a few billion dollars that can fund the top to bottom construction of a new state of the art facility and fund the software and medical equipment development to get things running how they ought to.
Thread over.
Man, that is quite impressive, must be some kind of a world record? I would really like to see that app :)