In a comment, Tim Worstall wrote,
Assurance is a means of, possibly tax privileged, saving for high probability events. Insurance is a method of, well, not saving but providing for, low probability and expensive events. Fire insurance for a house is insurance, burial insurance, given the low probability of being lost at sea, is actually burial assurance.
Health care is both. That appalling cancer that the new drug is $475k for, the scraping up off the road and 12 months in ICU as a result are both low probability events. Insurance is the right model here – and we’ve got it, catastrophic insurance.
What ails US health care more than anything else is that things which are assurance – annual blood tests, contraception for the majority of a woman’s fertile life, shots for the kids, are dealt with through an insurance, not assurance, model.
I don’t say this is all of it, but splitting out the two would help. Catastrophic, even government run (as Brad Delong has suggested) insurance plus those medical savings accounts for the assurance. Will never happen of course but it would help.
I believe that is the Singapore model, but sadly its prospects here are quite dim.
Much catastrophic care is a mixture of what your calling insurance and assurance. If I need an expensive medicine, and I need it continually, then its both a large expense and predictable. For the population the probability may be low, but for the pre-existing condition person its high. You can’t insure against something with a high probability.
In a week millions of people will type their drugs into Medicare Plan Finder and pick a plan not based on insurance risk but based on their known current health costs, some of which may already be known to be tens of thousands of dollars a month.
Singapore treats predictable low cost expenses based on the incentive model (copays/coinsurance), but the predictable high expense stuff is paid for and regulated by the government. There is no copay you can invent people with if its only going to be a couple % of the actual cost. Since the amounts involved are too high for normal people to pay, a third party payer must be involved, and that third party payer must impose price and quality discipline on that spending (which Singapore does). Singapore does have “insurance” for actual one off health events, but these are small compared to chronic conditions.
It doesn’t help that most of us never get to see “real” numbers. The hospital we go to charges over $1000 for routine bloodwork – the kind of thing that comes under “assurance.” Do a couple of these a year – and for multiple family members – and it adds up significantly. If we lived in an urban or suburban area, there would likely be labs that compete with high cost hospitals on this. But living in the boonies makes that difficult to find.
Some 70% of costs are catastrophic. The problem is paying that much for something that unlikely is not attractive, especially when you can declare bankruptcy and throw yourself on charity or government. Lesser costs are just loss leaders to persuade people to pay for the former and help reduce catastrophic costs. They are not something people should economize on and is why a division can be counterproductive.
Citation on that 70% stat?
I’ve often made short comments on how healthcare is really many markets that we’ve smashed together. Since Arnold is going to have a post on the topic, I’ll spend a few minutes to lay my thinking out more explicitly.
On short notice, I can come up with at least 9 distinct markets. Each of these is yet more markets and even those break apart (stroke care is not cancer care and breast cancer is not pancreatic cancer).
That said, here they are. Notation is: Market (examples) [typical age of onset] [est cost, one-time or annual (episodic/chronic)] [whether it’s controllable]
(1) Late stage chronic (CHF, renal failure, Alzheimer’s) [elderly] [$25k+/yr] [often but not always due to personal decisions, severity sometimes controllable]
(2) Unexpected catastrophic episodic (trauma, neonate ICU) [any age] [$100k+] [hard to control]
(3) Expected catastrophic episodic (cancer, stroke) [late middle age or elderly] [$100k+] [generally no, but hits most people eventually]
(4) Early chronic (diabetes, obesity, high blood pressure) [early middle age or later] [$1k+] [yes]
(5) Serious episodic (ACL repair, normal birth) [young adult or later] [$5k-$25k] [often can be influenced]
(6) Early onset chronic (lower back pain, IBD) [adult or later] [varies, procedures can be $10k+] [maybe?]
(7) Elderly lifestyle (joint replacement) [late middle age or later] [$25k+] [decision to treat: yes, but impairs quality of life if not treated]
(8) Childhood episodic (croup, POTS) [thru 18yrs] [varies, hospitalization can be $10-$30k] [more no than yes]
(9) Childhood onset chronic (type 1 diabetes, genetic disorders, autism(?)) [varies, can be $200k+/yr for some meds] [no]
I would address the markets as follows:
* Impacts everyone in later life: 1, 3, 7. Transition to 401k-like golden-years medical savings accounts with expectation that people will save for their own end of life care (if you won’t save for it why should anyone else?). Possibly with social matching/backstop
* Impacts a small percent of people, but is a major issue if it does: 2, 5. This is the sweet spot for hospitalization insurance. Maybe provide #2 like police or fire.
* Impacts some people more than others and some earlier than others. More quality of life than mortality. Treatment costs would probably some down substantially in a real market (e.g., outpatient joint replacement) but raises issues of equality. If socially supported, we have the current, not-responsive-to-incentives model.
* Impacts many kids to some degree. Reasonable case for social support of low-income kids
* Impacts a small percent of kids, but is a major issue if it does. Good case for social support of low-income kids, though can be gamed/abused (e.g., the high % of school budgets spent on ‘special needs’, variously defined)
Markets where behavioral factors play a role (like BMI or sky-diving), I think you allow insurers to price deferentially.
I think I would outlaw use of genetic data in pricing (to blend the risk pool more). Though it might impact propensity to purchase…so I’d have to think about that.
I lean toward ending EMTALA and similar rules that require providers to treat all comers as an incentive to purchase insurance. Though again, I’d have to think that through.
Varying degrees of regulation. Most for #2. Some for #1, 3, 5. Probably #8 and 9 too.
Arguably less for the others, but probably with some sort of standardized quality outcomes capture and reporting.
5% of the population drives 50% of the cost and 25% of the population drives 75-85% of the cost.
#4 and #5 aren’t the big hitters. #1 and #3 are big. #2 and #6 probably more for commercial plans (younger populations). #7 is material.
If we moved to a fractured model like this we would need an adjudication mechanism for the edge cases. Maybe a dedicated court system like for patent or military law?
That might also help with malpractice, which drives up costs in areas like obstetrics.
If I ever start a blog, this is one of the things I’d try to flesh out. So I’d be interested in reactions.
Worstall does not understand health care spending. The kind of expenses he cites make up a tiny percentage of health care spending. Fielding covers this some. The other important stat to remember is that 50% of people account for 3% of our health care spending. Shots for the kids come under this 3%.
Medical savings accounts can work well for those are well off. Not so much for everyone else. Read Mark Pauly on this topic.
Steve
50% of people generating 3% of cost is why I’m very skeptical of the economics of wellness plans and broad patient centered medical homes (PCMH). There’s just not enough spend there to justify any reasonable program cost. At least in the short term (3-5 years, the amount of time someone is typically on a given health plan).
PCMH for CHF patients, well that’s another matter…
I generally agree. Insurance should not be first dollar. But that’s what ppl want.
I currently have a $7200 deductible bronze hsa plan. I’m ok with this model but others aren’t. It would need to be a welfare type set up for those who can’t afford routine care. Then maybe market forces can work for things like blood work for cholesterol checks etc.
Who knew health care was so complicated??
I’d love to get insurance out of the small-dollar value routine stuff, but, as others have noted, there’s not much reason to think there’s a whole lot of savings there. The real healthcare money is going to the expensive cases. Getting real savings in the system means getting those prices down or delivering fewer services to those folks.
It is the Singapore model and prospects in the US are dim. But many other countries run “lite” versions of the same. You do pay, out of pocket, some nominal fee to see a doctor etc. Even in what many would call socialised medicine places, Norway certainly, Sweden I think.
The idea of combining catastrophic insurance with an HSA assumes that the catastrophic insurance will be cheap.
The only this happens is if we let the insurers do underwriting.
And this will happen if we are willing to pay taxes so that people who are already sick can go onto Medicare.