I hear this sort of argument a lot in different fields. Usually it's because the IT guy doesn't really understand the business they are trying to automate or where the true pinch points or time savings are.
Could you provide some examples of fields where practitioners control both supply and standard of practice where automation is also shunned, perpetuating high costs? Also, note, the largest source of bankruptcy in the US is medical costshttps://www.cnbc.com/2019/02/11/this-is-the-real-reason-most...
"They dont understand the business" is a great excuse for maintaining status quo. I'm an Engineer, a quant, and a computer scientist by training and I refuse to accept defeat w/o sound reason. I will if I'm given a good reason, but "go away you guys, you dont understand our business" is defeatist. If we all accepted such answers society would never progress. I'm sure horse carriages said the same thing when people tried to invent motor vehicles.
I’ll give you a concrete example in the legal field. Big firms might have reasons to avoid labor-saving automation, because they bill by the hour. But a large fraction of legal work isn’t billed by the hour, it’s contingency work (where the firm gets a certain fraction of a recovery) or fixed fee work. If you’re getting paid 1/3 of the amount you recover (a typical contingency fee) you have enormous incentives to do as little work to get a good result as you can. But those firms don’t use a lot of legal technology either, because it’s just not very good and not very useful.
The bulk of legal practice is about dealing with case-specific facts and legal wrinkles. And machine learning tends not to be useful for that, at least in current forms.
So, machine learning does get used quite a bit in the legal industry, at least outside of small practice. But it tends to be much more successful when it's used as a force multiplier for humans rather than a replacement for humans.
For example, the idea of using document classification to reduce review costs has been around for a long time. But it took a long time to get any traction. Some of that was about familiarity, but a lot of it was about the original systems being designed to solve the wrong problem. The first products were designed to treat the job as a fairly straightforward binary classification problem. They generally accomplished that task very well. The problem was you had to have a serious case of techie tunnel vision to ever think that legal document classification was just a straightforward binary classification problem in the first place.
Nowadays there are newer versions of the technology that were designed by people with a more intimate understanding of the full business context of large-scale litigation, and consequently are solving a radically reframed version of the problem. They are seeing much more traction.
The coordination problems in creating a system designed from the beginning to be human in the loop is a challenge.
There are a lot of great ML algorithms, even if you limit yourself to 10-20 year old ones, that aren't leveraged anywhere like how they could be because very few know how to build such a system by turning business problems into ML problems and training users to work effectively alongside the algorithm.
CRUD application development projects blow past deadlines and budgets frequently enough. ML projects have even greater risks.
Edit: I hope the people making the successful legal document management system you mentioned write about their experience.
FWIW, my experience has been that, if you're trying to build a system that works in tight coordination with humans, you're better off sticking to algorithms that are 40-80 years old. Save some energy for dealing with the part that's actually hard.
That WP article doesn’t support your claim. It’s about the number of bankruptcies, not the leading cause. Nonetheless it does cite a survey that found medical bills contributed to 60+% of bankruptcies, and that it doesn’t really make sense to talk about a single cause.
It's a stat that requires a lot of contextualization. To your point, you're absolutely correct that the number of bankruptcies is important, because over the last couple decades, 1) bankruptcies in general have been falling, 2) medical bankruptcies have also been falling in absolute terms; but because the denominator has dramatically fallen relative to the numerator, the numerator looks larger than it actually is.
Why not simplify the medical bankruptcy discussion?
Fact is Americans have high personal cost and risk exposure relative to nearly all of the rest of the world.
Second, our system has making money as the priority, again in contrast to much of the world.
Finally, most of the world recognizes the inherent conflict of interest between for profit and sick/hurt people and both regulate that conflict to marginalize it, and make it so people have options that make sense.
My take, having been chewed up by our toxic healthcare system twice now (having a family does matter, lol), is the temporary dampening on cost and risk escalation starting the ACA brought to us is fading now, and issues are exceberated by the pandemic (demand for care crashing into variable supply), and shifted somewhat as large numbers of people fall into subsidy medicaid type programs due to job loss.
The honeymoon period is long over now, and the drive to "make the number" is going to be front and center and escalating from here.
TL;DR: We are not improving on this front at all. We need to.
I could go on at length about high student debt and it's impact on these discussions too.
The radiology control over labor, preserving income for it's members is totally real, and fron their point of view, necessary. They ask the legitimate question in the US: How can I afford to practice.
Most of the world does not put their medical people in positions to ask that question, with some exceptions, those being far more rare and easily discussed than most of the topic is here.
"American healthcare in short:
~60% (in good employer plans, generous state Medicaid, or M.Adv/Medigap) have the best healthcare in the world.
~30% have insurance with gaps/risk of big bills.
~10% uninsured must rely on uncompensated care, go without treatment, or risk bankruptcy
The strength of M4A proposals is that they begin with an understanding that the 40% exist and need things fixed.
Their weakness is that they pretend that the 60% don't, and threaten to take away what they have."
The fact of the matter is that the majority of Americans have excellent, world class health coverage. The problem is that there exists a small percentage of Americans that are totally screwed, and this is a higher percentage than most other comparable countries. There are a couple reasons why, which brings me to...
> Second, our system has making money as the priority, again in contrast to much of the world.
First of all, this is false insofar as not all health insurance in America is for-profit. Blue Cross Blue Shield, for example, are predominately 501 non-profits (with a few notable exceptions).
Second of all, while you're right that much of the world has public insurance companies that don't seek to "make money", there are a number of countries with world class healthcare that do have profit seeking insurance, many of them with purely private profit driven insurance companies: including Switzerland and the Netherlands. Some have a hybrid of public/private, including Germany (public/private mix), Singapore (public/private mix), etc. In fact, while many countries have a public insurance system, it is extraordinarily rare for countries to outright ban private insurance options.
Third of all, in America, health insurance is one of the most regulated industries in the country. After ACA was passed, there's a strict cap on profit margins that health insurers can enjoy. It's not too dissimilar from how private health insurance is regulated in Switzerland and the Netherlands, both of which have some of the best healthcare on the planet.
> Finally, most of the world recognizes the inherent conflict of interest between for profit and sick/hurt people and both regulate that conflict to marginalize it, and make it so people have options that make sense.
Again, as I mentioned above, this is not only not true, it's debatable if such an "inherent" conflict of interest even exists. By this logic, there should be an inherent conflict of interest between for profit food providers and "hungry/starving" people. The profit motive alone can't explain America's health outcomes, because there exists countries with fantastic healthcare systems (Switzerland, Netherlands) which are driven purely by private health insurance.
America actually has a pretty good apples-to-apples experiment of "profit seeking" vs "not profit seeking" insurance, ironically in Medicare Advantage. When you turn 65, you have the option to enroll either in "Original Medicare", which is what we usually think of when we talk about "single payer healthcare in America", or you can enroll in Medicare Advantage (aka Medicare "Part C"), where the premiums that would go to the CMS instead go to private insurers like Humana, United, Oscar Health, Aetna, Clover, etc. These plans replace Original Medicare, also cover Part D prescription drug benefits, and often include supplemental benefits that Original Medicare doesn't already cover. There are some interesting findings so far:
- 39% of Medicare beneficiaries are on private Medicare Advantage plans instead of the public "Original Medicare". Because everyone is entitled to "Original Medicare", this is purely voluntary. This number has been growing so rapidly, that we expect by 2025, more seniors to be on a private plan than the public one. There's also great variance by State. In Florida, Pennsylvania, Wisconsin, Michigan, Minnesota, Oregon, Alabama, Hawaii, and Connecticut — nearly 50% of beneficiaries are on Medicare Advantage. By 2022, we expect more seniors in those States to be on a private plan than a public one. https://www.kff.org/medicare/issue-brief/a-dozen-facts-about...
- In Urban areas, Medicare Advantage costs less per capita to administer than Medicare — and that's not including the extra Medicare Part D insurance that you would have to buy if you're on the Original Medicare plan. https://www.commonwealthfund.org/publications/issue-briefs/2...
So the reality is really more complicated than you're making it out to be.
From where I sit, the one thing that sets apart America from the rest of the world is not that health insurance can be profit driven (so do the Swiss and the Dutch, for example), it's that health insurance is coupled with employment. There's really no other peer nation for which this is the case, and a lot of the economics of health insurance look the way that they do because big employers buy most of the health insurance in today's market, and that has resulted in market distortions that hurt those that are unemployed. What we're seeing in healthcare costs is analogous to what you might see happen to airline ticket costs if we all got our air tickets through our employers: the vast majority of us would fly business class, while the unemployed would be simply unable to pay for business class fares out of pocket. Employers (especially medium-to-large businesses) have a much higher purchasing power (and hence, willingness to pay) than individuals.
As a question, why haven't any of these techniques made waves outside the US? Other countries don't have the same monopoly/monopsony powers in the medical industries that are prevalent in the US.
US is exactly the place where those techniques would make waves because of what the US is paying for radiology; in countries where radiologists don't have the same monopoly/monopsony powers it's not nearly as lucrative to replace them.
For example, I'm distantly involved in a project with non-US-radiologists about ML support for automating radiology note dictation (which is a much simpler and much "politically cleaner" issue than actual radiology automation), and IMHO they and their organization would be happy to integrate some image analyis ML tools in their workflow to automate part of their work. However, the current methods really aren't ready, and the local market isn't sufficiently large to make the jump and make a startup to make them ready, that would have to wait until further improvements, most likely done by someone trying to get the US radiologists' money.
There's not really a way to disambiguate the two though - the fact that there are lots of medical technology startups and new drugs coming out of the US is because of the costs involved and how much can be harvested by being a little better. This creates new technologies that the US can't really protect against proliferation - so all of the money has to be harvested from the US market.
This isn't necessarily a bad thing - I for one happen to think it's great that our expensive medical system is financing all kinds of wonderful new technologies that benefit the world overall. However, the major problem here is that things that would be useful for other places simply don't have the market to support it, so most medical innovation exists in the context of the US medical system and it's problems - some of which are widespread, some of which are not. I do wish there were some other testbed healthcare systems out there for companies to try to disrupt, but I don't think it is (by itself) a call for medical reform.
My preferred medical reform is to "legalize insurance markets" (ie: repeal laws that state that insurance companies operating in state Y cannot sell insurance to people in state X because state Y policies are not legally compatible) and try to break the monopoly that doctors and nurses enjoy....somehow. Telehealth? Maybe?
But, is it? Almost half of the funding for Healthcare innovation is governmental even in the US, and a competent public health system already has a strong incentive to reduce costs. So if a technology has the potential to reduce costs, a more efficient healthcare system would also pay for it - and if one doesn't there are dozens that can too - and if there is no path to it providing value overall in such a system, then it's going to be on the balance less efficient anyways.
To note, a big issue in public innovation is that rich western countries, led by the US, HATE governments competing against the private sector. So if a government comes up with an innovative solution, they are generally disallowed from selling it, which hurts everyone except private companies.
This happened in my city a few years ago, which had a very early innovation in bike sharing, much before any VC funded bike sharing company, and other cities had expressed interest in paying my city for implementing this service locally.
But because of laws banning public endeavors from engaging in commercial activities, this was struck down, hurting all the taxpayers in my city, citizens of my city that would have benefited from better service from the experience, and millions of citizens from interested cities which would have received better service and who would have saved money.
Or another case in my province was the invention by the electricity utility of the electric hub motor - which is now over 40 years later in widespread use due to its efficiency and low cost - but instead of exploiting that patent and selling those things, it had to partner with a private company which got exclusivity and mostly squandered it. Again, hurting almost everyone which might have benefited from lower cost electric transportation as well as the taxpayer here.
This is actually part of why China is eating everyone's lunch, they used their size and never really agreed to these rules, leaving themselves the opportunity to profitably invent at the state level, leading into more efficient state owned enterprises that can often profitably outcompete the private sector.
> I for one happen to think it's great that our expensive medical system is financing all kinds of wonderful new technologies that benefit the world overall.
Does it factoring in situation of people unable to pay medical bills?
If the entire rest of the world isn’t a big enough market to be worth developing for then maybe we don’t need ml radiology we just need medical reform.
The entire rest of the world isn't a market, it's many separate markets that need to be entered separately by overcoming different moats. Market fragmentation matters, especially in regulated industries like medicine.
But yes, medical reform is definitely something that might be helpful - technological solutions almost always aren't the best way for solving social/political problems.
> the largest source of bankruptcy in the US is medical costs
That's not what the article says.
"Two-thirds of people who file for bankruptcy cite medical issues as a key contributor to their financial downfall."
Those issues can absolutely include direct costs, but they also include things like not being able to work, needing a lot of day to day help, and other things that increase costs and reduce income even if the actual medical costs were largely covered.
I don't really know of one.. I don't think automation is ever shunned as long as it is useful and known to be useful.
Everyone likes things that save time.
There is an essentially an unrestricted demand for healthcare across the world.. they will use the time to either talk to their patients more (or start to if they don't already).. or they will move into other medical fields.. or increase the volume of screening.. (may be harmful, but that's another matter). They probably don't want to do it as it won't really save them much time. OR it will save them time and they have been burnt before. For example, early voice recognition was very poor and over promised. Stopped me using it for ages after it became fairly good. It's still not actually better than typing, but it is closer now. Let's all focus on voice recognition that works before moving on to grander plans....
The job of a GP is actually probably one of the top hardest to automate, because the GP's main (and often only) job is to extract information. And that _does not_ consist in performing plenty of tests, but in speaking to and most importantly listening to the patient.
> You don't need a ton of training to get the above from a patient and into a computer, and to ship the samples.
Great! And you know what good that would do to improve diagnostic accuracy? Zilch. Zero. There's a saying that '90% of diagnoses are done on history'. Now tell me why that would be different for an algorithm given identical information? If there was a simple answer to that, we'd already be running statistical models over patient labs all day long, which we're not.
> are you either a) a defeatist, or b) a GP?
I'm an epidemiologist and also a practicing anesthesiologist, which is why the statistical theories of people who have never set foot in the clinics to see what's the job really about make me want to jump off a bridge.
- Doctor says "Say ahhh" and looks in my throat with the thingy
- Doctor looks in my ear canals with another thingy
- On other occasions, my other vitals are taken, maybe some vials of blood, etc. Again, a student can do this.
I'm asked a few general questions, with some follow-up questions based on my answers.
Then the doctor puts this information - along with my patient history - into the decision tree in their head and comes up with a result. If the doctor is stumped, I'm sent to a specialist.
The above can be automated, plain and simple. It would also be an improvement over my experience of the health system - in Canada. I have never seen my GP pull up a multi-year graph of my blood pressure, weight, or whatever. What I am describing is a system for creating regular data points of the kind currently used in diagnosis. What I fail to understand is how you cannot see that there must necessarily be predictive value in such a database.
Even if only 80% of the job can be automated, public health would improve immensely if the global population can do regular checkups like the above cheaply.
That's not really the hard or useful part part. According to a radiologist and machine learning researcher[1]:
"It turns out that deep learning is a very good match for some of the most time consuming (and therefore costly) parts of medicine: the perceptual tasks.
We also saw that many decisions simply fall out of the perceptual process; once you have identified what you are seeing or hearing, there is no more “thinking” work to do."
The thing about Health Care is most efforts to automate it have failed. Arguably that's because no one "understands" the field, in the sense that no one can give, codified summary of the way they operate; each professional who's part of a health care pipeline takes into account twenty different common variabilities in human body/health/behavior/etc.
It's similar to the situation of self-driving cars, where the ability to do the ordinary task is overwhelmed by the existence of many, many corner cases that can't be easily trained-for. Except in health care, corner cases are much more common. Just seeking health care is an exceptional relative to something in ordinary life.
"They dont understand the business" is a great excuse for maintaining status quo. I'm an Engineer, a quant, and a computer scientist by training and I refuse to accept defeat w/o sound reason. I will if I'm given a good reason, but "go away you guys, you dont understand our business" is defeatist. If we all accepted such answers society would never progress. I'm sure horse carriages said the same thing when people tried to invent motor vehicles.