I would even argue they should take all of those bits of text and replace them with an on-hover tooltip element, with the icons indicating the degree to which those things are stored externally. It's a bit more alarmist than it needs to be as is.
This is such a great example of the kind of disruptive impact we can expect from LLMs.
It exposes part of how the health insurance industry works:
"Out of roughly 40 denials, she won more than 90% of her appeals, she estimates"
Most people don't have the stamina to dig in when this happens, so my guess is the entire insurance industry is designed around the assumption that a lot of valid claims won't be paid out.
It's obviously disgusting that the industry has incorrect denials baked into their business model.
If we fix that with LLMs, what happens? It's going to have knock-on effects, since it could eliminate the profit margin these companies have right now.
They could all merge into one giant insurance agency, perhaps a National Health Insurance.
Once they eliminate all competition the next target for optimisation is between the company and hospitals. They realise its cheaper, with better margins, to run primary care facilities rather than hospitals. The government see that they are a monopoly and force them to provide coverage for a minimal monthly cost. It seems like such a good idea that eventually the government agrees to give them a set amount each month for universal coverage of citizens.
They rename themselves as the National Health Service.
I see what's going on here.
It's another example of externalization of costs.
Doctors have no incentive to fight the denial of insurance. They should, but it's not important yet. If they have to sue you eventually for the billed cost, they will; or at least they'll sell the debt to a collection agency at some number of cents on the dollar.
The Insurance company most definitely doesn't have an incentive. Once you're denied, they're off the hook. They kick it back to the provider, and ultimately to you through some debt collection agency.
The only person that has an incentive is you. And now you're stuck with the cost; either in dollars or time, but what's the difference really? It's going to cost you either way.
Maybe Holden Karau's approach is good and valid, but should it be converted to a paying service we're just back at square one where it's either time or money. Maybe there'll be a discount which makes it cheaper net/net, but who knows?
The little person here is the patient. All the others, doctors and insurers, are the big persons. The little person absorbs the cost, whether they want to or not.
Single payer or not but I think everybody should agree that dealing with health insurance needs to be easier and more predictable. When I read the stories of how people won against them, it's always mind boggling to see the giant effort they had to put into it.
Hospitals and insurances should be fined if they have a high ratio of claims that get rejected first and reversed after appeal. Patients simply shouldn't have to go through this. The whole system is set up to profit from wearing down patients by attrition.
Most of the time it's not too bad to appeal a claim for smaller amounts. People go through the hoops for the larger denials because they are so motivated.
I got a $150k bill for emergency care for my daughter once; it wasn't too bad because they denied it as "uncovered elective procedure" which was trivial to show it wasn't, but it was still a tense month for me waiting. The punchline, of course, is that the insurance company and the provider agreed upon $30k.
Doctors' offices can't tell you what a procedure will cost ahead of time, and the amount they bill you afterwards is completely fictitious (because if you don't have health insurance you probably can't pay 5x what the insurance companies pay, so will be bankrupted anyways).
“Most of the time” in my case, it has been bad. Because it wasn’t one big bill. It was dozens of bills for $200-500. Each submission could fail in unpredictable ways. “Lost” claims forms, database “errors”, “incomplete information”, etc. And trying to deal with this burden while being sick, was too much for me. I scraped back a few dollars, but had to give up and move forward with my life.
Too often, we throw our hands up and declare legislation impossible "because companies will just get around it in ways, X, Y, and Z". The solution is to write better laws that also remove X, Y, and Z as options. And also X0, X1, Y0, Y1, Y2, Z0, Z1, Z2, Z3, and Z4. The world is complex and regulation can be complex, too. You'll never fix things by giving up and saying "We can't write the perfect law--oh well, let's just let companies do whatever they want!"
This is what I was thinking reading TFA. Doesn't it become an AI arms race, where insurance companies use AI to deny claims, and customers use AI to fight the denial ad infinitum?
Readit News
https://fighthealthinsurance.com/scan
Specifically laying out what and how everything is stored. Hahah. Probably overwhelming to a non-dev, but Kudos either way!
However, it would mess with mobile?
It exposes part of how the health insurance industry works:
"Out of roughly 40 denials, she won more than 90% of her appeals, she estimates"
Most people don't have the stamina to dig in when this happens, so my guess is the entire insurance industry is designed around the assumption that a lot of valid claims won't be paid out.
It's obviously disgusting that the industry has incorrect denials baked into their business model.
If we fix that with LLMs, what happens? It's going to have knock-on effects, since it could eliminate the profit margin these companies have right now.
Once they eliminate all competition the next target for optimisation is between the company and hospitals. They realise its cheaper, with better margins, to run primary care facilities rather than hospitals. The government see that they are a monopoly and force them to provide coverage for a minimal monthly cost. It seems like such a good idea that eventually the government agrees to give them a set amount each month for universal coverage of citizens.
They rename themselves as the National Health Service.
Hospitals and insurances should be fined if they have a high ratio of claims that get rejected first and reversed after appeal. Patients simply shouldn't have to go through this. The whole system is set up to profit from wearing down patients by attrition.
I got a $150k bill for emergency care for my daughter once; it wasn't too bad because they denied it as "uncovered elective procedure" which was trivial to show it wasn't, but it was still a tense month for me waiting. The punchline, of course, is that the insurance company and the provider agreed upon $30k.
Doctors' offices can't tell you what a procedure will cost ahead of time, and the amount they bill you afterwards is completely fictitious (because if you don't have health insurance you probably can't pay 5x what the insurance companies pay, so will be bankrupted anyways).
Or it could go the other way depending on arbitrary factors. Nobody knows.
They'd just invest more in not having to reverse on appeal. Handling these kinds of incentives with rules and strictures gets messy fast.
Too often, we throw our hands up and declare legislation impossible "because companies will just get around it in ways, X, Y, and Z". The solution is to write better laws that also remove X, Y, and Z as options. And also X0, X1, Y0, Y1, Y2, Z0, Z1, Z2, Z3, and Z4. The world is complex and regulation can be complex, too. You'll never fix things by giving up and saying "We can't write the perfect law--oh well, let's just let companies do whatever they want!"
That would be the goal. Get it right first time around.
Deleted Comment