Live data from Hacker News

A petabyte of health insurance prices per month

blog.turquoise.health

241–250 of 275 posts

Re: A petabyte of health insurance prices per month

#241
This is interesting: could we see it as the most exhaustive set of plausible/probable human events of a harmful nature ?

Sure it is for insurance purposes, but I’m wondering what we could use it to label human experience in general—and what applications that could help.

“Hey Chat-gpt, I’m thinking of taking a walk near that pond, what should I expect…”

Re: A petabyte of health insurance prices per month

#242

Because our (US) system is at Monty Python levels of absurdity. We waste more money with this "system" than most other countries spend as a total, and with worse outcomes on average [1]. Fundamentally, there are two factors, IMO, that have been conclusively proven (repeatedly, and for decades - much more strongly, recently) to be incompatible with delivering quality healthcare with some level of efficiency: 1) For-pr…

> 1) For-profit entities (particularly, publicly-traded / with "fiduciary responsibility" to shareholders)

Then why is laser eye surgery so efficient?

Answer: it's not covered by insurance, so there's an actual market.

The problem isn't for-profit entities, it's for-profit entities operating in the absence of markets. Rent-seeking, in short.

Re: A petabyte of health insurance prices per month

#243
post #77

Earlier quoted context omitted.

I work with petabytes of data. A petabyte of text is an insane amount of text. You need a billion X multiplication of basic data. That’s not just combinatrionics, it’s deliberate obfuscation.

4 tables, each with 1k rows and one of them with a 1K text column in it, joined together to give a cartesion product, will give you a petabyte of data. So it doesn't sound impossible (just absurd).

even taking your contrived example, you have to store it in the stupidest way possible to get that level of storage - even assuming your 1K of text is incompressible, you only have 1K unique values. The most basic compression algorithms will make easy work out of compressing the crap out of it, leave alone the advanced compression algorithms used for any big data storage formats.

Re: A petabyte of health insurance prices per month

#244

Because our (US) system is at Monty Python levels of absurdity. We waste more money with this "system" than most other countries spend as a total, and with worse outcomes on average [1]. Fundamentally, there are two factors, IMO, that have been conclusively proven (repeatedly, and for decades - much more strongly, recently) to be incompatible with delivering quality healthcare with some level of efficiency: 1) For-pr…

> 1) For-profit entities (particularly, publicly-traded / with "fiduciary responsibility" to shareholders) Then why is laser eye surgery so efficient? Answer: it's not covered by insurance, so there's an actual market. The problem isn't for-profit entities, it's for-profit entities operating in the absence of markets. Rent-seeking, in short.

What fraction of the people who could benefit from LASIK get it? Would you be okay if cancer worker like that?

Re: A petabyte of health insurance prices per month

#245
post #49

There is so much data because of insurance. Insurers are in the business of declining claims. They have discovered that the best way to decline claims is to create byzantine rules while offloading all the work to comply with those rules to healthcare providers. The only way to fix it at this point is to nationalize the system or to exclude all routine care from the insurance system. Something like a legal ban on dedu…

In Switzerland, healthcare is somewhat private. The legal framework is enacted by law, and the Confederation decides the types of claims, prices, conditions, etc..

Every hospital (public or private) must follow the prices decided by the Confederation, and every health insurance must reimburse following the same prices, not more and not less.

And it works quite well I'd say

Re: A petabyte of health insurance prices per month

#246

For reference, here is the Israeli Ministry of Health price list as of July 1st 2023: https://www.gov.il/BlobFolder/dynamiccollectorresultitem/moh... Edit: Direct link to an Excel spreadsheet.

More or less the same in Switzerland with regulated, mandatory, private healthcare: https://browser.tartools.ch/#/tarmed_kvg/data/L/39.6040

Everyone must conform to these prices, rules.

Re: A petabyte of health insurance prices per month

#247
post #245
post #49

There is so much data because of insurance. Insurers are in the business of declining claims. They have discovered that the best way to decline claims is to create byzantine rules while offloading all the work to comply with those rules to healthcare providers. The only way to fix it at this point is to nationalize the system or to exclude all routine care from the insurance system. Something like a legal ban on dedu…

In Switzerland, healthcare is somewhat private. The legal framework is enacted by law, and the Confederation decides the types of claims, prices, conditions, etc.. Every hospital (public or private) must follow the prices decided by the Confederation, and every health insurance must reimburse following the same prices, not more and not less. And it works quite well I'd say

I wonder how you deal with the usual problems of price fixing. What services aren't done because they pay too little? How do you lure doctors into fields where the pay is terrible? What about rationing of procedures that are costly but pay too little - is the scheduling delay so long it impacts your health?

Re: A petabyte of health insurance prices per month

#248

Earlier quoted context omitted.

Here's the full list of codes for "Bitten by ____, _____ encounter" https://www.icd10data.com/ICD10CM/Codes/V00-Y99/W50-W64/W61- The hierarchy looks like this: W61.6 Contact with duck = W61.61 Bitten by duck == W61.61XA …… initial encounter == W61.61XD …… subsequent encounter == W61.61XS …… sequela = W61.62 Struck by duck == W61.62XA …… initial encounter == W61.62XD …… subsequent encounter == W61.62XS …… sequela = W6…

Thankfully, W61 is a non-billable code family… hope it stays that way! I do wonder what all the billable code families are, maybe I should find out.

The Z codes (ICD to be specific) are similar. Although those two are not exhaustive -- it can be conditional so can be confusing.

You have different rules for chronic conditions, so for example something could be rolled up into a billable DRG one visit, but should not be included in a follow up visit.

The DRG example on this website is interesting -- so I think these are valuable services, but I look at claims data everyday for my job and I would have a hard time knowing the exact codes before I go for a visit. (Also obviously in an emergency I am not shopping around!)

And then there are service modifiers for "how long" for CPT codes for just outpatient visits -- you wouldn't be able to know 15/30/60 minutes before hand even if you did know the specific CPT code. So for that you may want to know the proportion the doctor bills, not compare the prices on the same CPT.

Re: A petabyte of health insurance prices per month

#249

Earlier quoted context omitted.

You don’t need arbitrary codes to get statistics for something that can be typed in a description field.

I've spent a lot time around people doing statistics for environmental science, especially regulatory compliance around accidentally killing birds and bats. The unfortunate reality for the statisticians is that every tech thinks they're Henry David Thoreau or something. Like, nobody cares if it was a particular large exemplar of a given species, or that it had magnificent plummage, just write down the species of the…

> The billing codes exist explicitly to limit the amount of independent thought that can be applied to something that is adequately summarized as "bit by a duck, third visit".

None of this explains why these codes are needed for billing though. We simply don't need to bill differently for these pieces of information. The visit might have all these tags for informational purposes, but the bill should be for "visit to GP" and that's it.

Re: A petabyte of health insurance prices per month

#250
post #88

Earlier quoted context omitted.

But why? I could understand from a treatment perspective the type of animal mattering (but even then initial vs subsequent encounters?). From a billing perspective, wouldn't it make sense to have animal bite exam fee, with possible additional charges for bandaging, stitches, rabies shot, etc based on doctor's discretion?

Because these are world health organisation 'international disease classification' codes that are designed for population health study and happen to be used for insurance billing in the US. There are requirements for some data to be reported to federal and international organisations. For instance, these codes are also used on death certificates I think, so they can track whether a lot of people are dieing of flu (pr…

> Because these are world health organisation 'international disease classification' codes that are designed for population health study and happen to be used for insurance billing in the US.

This is the root of the problem. These codes should not be used for billing. That should be much more coarse. They bill you for visiting the GP. They bill you for a surgery. Adding the codes to the records is fine and good, but really should just be an internal detail that isn't that relevant to the patient.

Post reply on HN