" For those thinking that AI will reduce healthcare costs by improving efficiency, I point you to the 'electronic medical records' hype of 15 years ago. Are per capita healthcare costs any lower now? "
Can you elaborate?
In countries like Denmark or Estonia electronic medical records are well used among doctors and patients.
I think the point is that one of the big hype phrases is that it will decrease costs of healthcare and lower people's bills, but even if you lower the cost, there is no actual incentive to lower patient bills, so the extra value simply goes to owners and investors rather than the end consumer.
That's assuming the extremely broken and frankly incomprehensible American model.
Over in the EU, where most of the healthcare is single-payer (sometimes with optional add-ons/extra payers/private care), cost of care has been growing too. There are simply many more treatments which are available now than before, more tests possible to discover issues, etc. as well as of course the higher life expectancy and higher average age across the board. That doesn't mean improvements shouldn't be undertaken. Such as electronic medical files (which are the norm in multiple EU countries but sadly not everywhere, and not on an EU scale - as born in one EU country and now living in another, it would have been extremely useful to be able to automatically share my childhood records), or maybe ML to do predictions and stuff.
I agree 100%. In America the capitalist dream of competition and technological innovation actually driving costs lower for the consumer has mostly just died in almost every category. Any savings are now passed onto a cabal of private equity firms and legislatively-entrenched insurance corporations.
Goodhart's Law is in full force, and the measure of business health, profit, has become the target, rather than consumer welfare and rights. Boy have we optimized for profit at the expense of anything else.
Would you rather be treated in a modern hospital or one from 50 years ago? Would you choose the historical hospital if it charged you the same fraction of median income as it did back then?
It's not that black and white. There's more at play than just charges. Things can be better and still shockingly corrupt, with extreme amounts of unjustified overhead. Things are minimally better for any individual patient, while being extremely more expensive, and the insurance overhead has skyrocketed.
The question is a little bit like responding to a complaint about industrialized agriculture's effects on the environment by saying "Would you rather eat sausage from 100 years ago or now?" It's not actually that relevant to the initial complaint, and it reads like it's trying to just terminate the conversation on a loophole. Things can be better and still also be filled with many problems, particularly modern problems that weren't relevant in the time period you're comparing to.
To actually answer the question honestly, it depends on my condition. For the kind of thing I would typically see my GP for, including run-of-the-mill checkups, definitely one from 50 years ago. For any serious medical issues, modern (as long as I can convince the doctors that it's actually serious so they spend more than 30 seconds on investigation).
As far as charges go, per-capita healthcare expenditure is now over 5 times as high as it was in 1980. Are we getting a 5-fold improvement in medical care? When I go to the doctor for anything more than a checkup, 90% of my time is spent waiting, and I spend twice as much time with RNs who repeatedly ask me the same questions, before a frazzled doctor comes in, asks a couple questions, then rushes out, maybe handing me a prescription for antibiotics regardless of what my concern is.
The reason I wrote anything was that your original comment sounded very black and white to me.
I think we tend to overlook how much things improve in the long term. Knowing that they do hints at potential to do better still and should encourage us to take action and improve the situation.
Pointing out the current problems is necessary for that to happen. However, your comment came across as denying that healthcare can be improved, or at least that improvements can benefit patients. This is obviously false and I'm sure you didn't mean to claim it, so it's great we're able to add a bit more nuance.
Fair enough, and I appreciate the extra perspective.
I wouldn't deny that many of these things have improved. Very little actually gets worse in a literal sense (environmental issues and other negative externalities aside, though they are very relevant as well). The concern is that they tend to get marginally better when they should by all means be getting significantly better, because the lion's share of the benefits feed the ever-widening wealth gap.
Ceteris paribus is doing a lot of heavy lifting here.
If the 50 year old hospital has an available operating room and the modern hospital does not, I'm going there. I won't have a choice if the treatment needed is urgent.
The idea was that with all this healthcare data, patients would realize better outcomes. Like, your super-Doctor is going to realize that 6 months ago, you went to urgent care of XYZ condition, and that's going to shine some kind of light on your situation now, so he's going to be better prepared to treat you.
The reality is, the vast majority of that data is useless noise, nobody has the time to make any kind of analysis on it, and it's just another healthcare cost center providing zero value to the patient.
A major contributing factor to the lack of realized benefits with electronic health records in the United States is that the interoperability specification, CCDA, is only loosely complied with under the "meaningful use" rule from Health and Human Services. Certification is done on site at the EHR manufacturer, in an environment controlled by the manufacturer. The demonstration of compliance does not have to be done in a default "out of the box" environment, but merely that the system is _capable_ of being configured in a way that allows for emitting standards-compliant CCDA.
In practice, this means that most manufacturers do not build systems that are standards compliant by default, and definitely do not install systems that are standards compliant in their initial configuration. Then, individual healthcare practices add customization on top of that, which further complicates data integration.
Ultimately I believe electronic records are a huge factor contributing to the assembly line like experience of modern health care. I pray that I don't die in a hospital.
In the US, the cost of health care has not significantly decreased since the introduction of rules, regulations and government money pushing healthcare organizations to implement and use EHR systems.
I would say all of the intervention by the US federal government has made everything much worse. The EHR vendors (I'm looking at you, Epic) managed to consume the money thrown at the problem and yet have only moderately improved. In my opinion it also pushed smaller hospitals into expensive contracts with EHR vendors, eroding their already tenuous financial health and leading to even more consolidation.
Precisely, and for this reason, AI will be the snake oil panacea, just as computerization of records were. Both will be important, but the size of the contract will be based on the "transformative" messaging around the pitch. The real value as you've said will accrue to the consultancies and vendors as it always has, so even without the enormous compute bills for AI tools, prices will keep going up.
I knew someone who worked at a company that was contracted to be a support provider; they had at least one nervous breakdown.
Also interesting to note that the wikipedia article for the company mentions that in some European countries, moving to Epic was a huge pain that caused doctors stress.
OTOH apparently their actual corporate campus is some whimsical Alice-in-wonderland world with no expense spared.
>I would say all of the intervention by the US federal government has made everything much worse
This is assuming without intervention things would've gotten better. It's entirely possible things would be just as bad as they are, or worse, without said intervention
Lack of timely access to accurate information is and was a huge problem, that can be solved by electronic health records.
Whether or not the currently implementation is worth the cost/benefit ratio is debatable.
I envision a scenario though, where all EHR vendors integrate with Apple/Google, and someone’s phone can give access to their healthcare data, maybe even involuntarily in the event of an emergency.
