« Hold Off On Proposing Solutions | Main | "Can't Say No" Spending »

October 17, 2007


What would stop health care companies from vetoing highly expensive life-saving treatments for the chronically ill/disabled? They might make more money if those patients just died than to continue to pay for treatment. Profit maximizing strategies would be to take "cheap" sick patients with quickly curable conditions, cure them, and reap the benefits of having the new healthy patient. This would not be the case with many chronically ill patients, who will continue to require very expensive care and may never come very high into the quality of life category. What safeguard would you propose? Partial veto power? Who would decide then?

Also, how would this "large discretion to veto treatment" be implemented? Needing to file forms to get permission every time you need to see physician or get a medication, will be a logistical nightmare, which is why most plans tell you what they will cover before the fact- ie, not personally tailored. If you're talking about filing for reimbursement after the fact, this is also problematic, because the patients are unable to pay for it themselves, hence the medicare, and the doctors/hospitals will end up eating the cost. This is *already* a huge problem, esp. in NYC, where the majority of hospitals are deep in the red, in large part due to government failure to reimburse for many treatments.

Laura, we want health plans to veto very expensive treatments. That is the whole point; we don't want all possible medicine, just cost effective medicine. Since there is an auction for plans, it is not profit-maximizing to just take "cheap" sick patients; other plans will bid high for those patients as well. I had in mind something like health plans now; unless you want to pay out of your own pocket, you go to the plan approved doctors, etc.

Laura: If something causes a logistical nightmare for patients that will cause them to not seek treatments that could improve their quality of life and thus the payments to their plan providers, thus plan providers would make less money and would have an incentive to avoid the logistical nightmare in question.

Evaluating quality of life would be a substantial problem. So would the sheer size of the asset class that would be produced in this proposal (in terms of the value of the life-contracts for all the patients). I think that we have big macroeconomics questions and corporate governance questions to answer in evaluating this proposal, but in any event it seems implausible that an experiment this dramatic could be tried without many much smaller earlier experiments acquainting the public with this sort of mechanism.

Michael, yes a large new asset class would be created. But even with a crude evaluation of quality of life this should give a better incentive for quality than in current systems.

I don't think I'm clever enough to work out whether your proposal is a good idea, but I do think I'm human enough to suggest why the idea is not gaining traction. I recognise it as a tricky exercise in "mechanism design"; it makes my head hurt and I stop thinking about it least my brain explode.

Is it an insurance plan or a saving plan? A traditional view of medical insurance is that a cohort,
imagine ten people, all pay premiums. One of them is unlucky
and needs expensive life saving treatment that consumes most
of every-ones premiums. Then they grow old, and
traditionally there is no treatment for old age, so they die
without incurring great further medical expenses.

I'm missing the pooling mechanism. If the health care
company refuses to treat the unlucky member of the pool they
only lose 10% of the income, so they only fund one tenth as
much medical treatment as they would under a traditional
insurance regime.

Hmm, maybe it is the life insurance aspect that provides the
pooling. If only one person in ten will die young, absent
treatment, then $1000 will buy $10000 of life cover, and the
health care company will spend $9000 on treatment to dodge a
$10000 pay out.

What happens it the health care company decides that the
case is hopeless? It justs sits on its hands, refusing to
provide medical treatment in order to limit its costs to
just the $10000 payout. This is where my brain starts to
hurt. This looks intuitively like a flaw in the scheme. The
premiums were intended to pay for medical treatment, but
they have been transmuted into pure life insurance.

Thinking harder this looks like it is the point and strength
of the scheme. The obvious accounting treatment of failed
medical treatments is gross cost = lots, gross benefit =
zero, therefore net benefit = minus lots. This starts to
bring the idea of spending to show that you care to the
fore. Clarity about this is going to make a lot of people
feel very uncomfortable.

Then there is the difference between life insurance and life
assurance. When we consider end of life care we no longer
need to consider pooling of risk, we are all going to get
sick and seek treatment for our final illness. The usual
idea is that we have some money saved for medical expenses
and our doctor keeps spending it until its gone or we've
gone. Suppose that instead of putting money into a saving
scheme, ear-marked for medical expenses, we sign up with a
health care company. The company calculates and
recalculates. Will this treatment add a month to life
expectancy? Yes, just barefly. But it will cost two months
worth of health-payments, so it is not worth it. Obviously
if we had two months worth of payments in the bank we would
use it to purchase the treatment. Or we could offer to
double our monthly payment.

At this point I'm really confused about the when's. When we
pay a health care company are we building credit against
future expense? If not, do we also need a savings plan
against rising premiums if we want to meet our health care

I suspect that the answers are in http://hanson.gmu.edu/buyhealth.html if I look hard enough,
but this is a fraught area and will need a gentle
popularisation before the idea can hope to gain traction.

Quality of life insurance sounds less sensible than your earlier suggestion of relying more on traditional life and disability insurance. Pain seems too hard to measure, and it isn't clear that medical care's effect on pain is broken. The RAND experiment's evidence about dentistry sounds like the best evidence we have on pain management, and it suggests relatively good results.

It's not obvious that "we want health plans to veto very expensive treatments", but current incentives are poor enough that if the life insurance part of your plan is high enough it will improve the incentives.

Alan, Robin's original paper suggested $10 million in life insurance. Your intuition that $10000 won't work very well is true but not very relevant to his original proposal. The dollar amounts he mentions in this post are smaller than he proposed before, and I suspect they are too low.

Presumably, I have the right to create a health plan and to bid on myself as a patient. If what I want is freedom to choose my own health care, that is exactly what I will do.

Of course, if I have inside information about what it will cost to keep me healthy, I may let myself be outbid by some other health plan.

In other words, at first blush this sounds to me like a system that can easily be gamed.

Alan and Peter, I never uttered the word "insurance" in my post, and don't think that is a useful way to think about it.

Alan, yes, some are uncomfortable with any health system that makes it clear that everything possible will not always be done.

Arnold, I focused in my post on a Medicare reform version, but alluded at the the end to private versions. For private versions, you would of course have the option to not auction for a health plan, and just cover yourself. I agree asymmetric info could have effects, but I don't see that as a serious problem.

"we want health plans to veto very expensive treatments"

No- I assumed we wanted to avoid unnecessary treatments, not merely expensive ones. The 65 year old woman with artheriosclerosis and left coronary dominance needs that triple bipass surgery, no matter how much a company will lose from providing it. Ok- maybe you'd argue it's more cost effective just to let her die and use the saved money on helping other patients who can be more easily helped. This line of reasoning is very dangerous for society, because once a person is sick enough or old enough, it will not be "worth" treating them anymore. Endgame. Why would anyone take care of the elderly or dying? It might be worth giving the bypass to a 45 year old, likely to "pay off" for another 20 years, but not for the 65 year old, who'll pay off for only another 5. Why provide government sponsored healthcare at all if you are going to overlook the people that will need it most? I think giving financial incentives to improve quality of life is a good idea, but I don't think absolute veto power over medical treatment to maximize profits could work out for the better.

Yes, I also assumed we wanted to elininate waste. An extremely expensive procedure that saves a life cannot be defined as waste unless we put a price tag on that life. Whereas a dirt cheap pill that does nothing to you is waste.

Laura, unless you are willing to endorse always doing any treatment with any positive benefit, no matter how high the cost, you must somehow compare costs and benefits when choosing whether to treat. My proposed mechanism allows the government to choose any value of life it is willing to pay for, but for any value of life chosen there will be some treatments that are too expensive. This is just as it should be, is it not?

But why make difficult choices and sacrifices before we have even eliminated all waste?

"The government" would pay the various amounts?
And where are they going to get the money?
Ultimately, it comes from us.
I think perhaps I could afford $30,000 "in great pain" plan, but it would greatly stretch my budget.

"My proposed mechanism allows the government to choose any value of life it is willing to pay for..."

No- by this plan the government is not paying directly for the healthcare itself, but rather paying a set fee and contracting the actual payment and deciscion to make payment out to companies who only care about profits. There may be costs that are too high to be reasonable, but this upper level should not be determined by whether or not the continued existence of a person will profit a company. If the company contracted out and received money to take care of that person, they should be responsible for all reasonable treatment whether or not it will profit them.

Laura, the whole point is to find a way to decide what treatments are "reasonable." If you have some other good way to do that, then you don't need my mechanism. My proposal is that treatments are reasonable when their cost is less than the benefit.

Evil, you are forgetting revenue goes in from the auction, and then goes out again for the annual payments. The net outgo equals the cost of medical treatment. For private versions, you would of course choose a value of life level you thought best given your budget and priorities.

To bad it would be too corrupting to offer Medicare recitients the money that their care should cost so that they can spend it where they please. Like you had a pretty severe heart attack here is $50,000. You had a small heart attack here is $10,000. etc.

"Evil, you are forgetting revenue goes in from the auction, and then goes out again for the annual payments. The net outgo equals the cost of medical treatment. For private versions, you would of course choose a value of life level you thought best given your budget and priorities."

Wouldn't that kind of loop have a net profit of $0.00 ??

I really like that you're thinking about it, but after having given a great amount of thought to this myself, I honestly think that a pure free market approach would be far better than anything government run. Seriously, when was the last time the government did anything efficiently? Or even on time and under budget?

"These evaluations of disability and pain might be based on random auditor visits."

Robin, can you elaborate on how you would quantify the health gains or quality of life in this mechanism?

Would people be auctioned 1-at-a-time, or in cohorts? I am not clear why any plan would bid anything for a chronically-ill person in a lot of pain.

Maybe I'm picking up on the wrong part of the post, but

> I remain puzzled to see no interest in this idea.

attracts most of my attention. Radical ideas need a following, not least to have many eyes to get the bugs out. So what are the weak points of the exposition, viewed as an attempt to attract interest? My reason for posting the following criticisms is that I would like to see a more attractive expostion leading to a wider discussion of this important idea.

> Thus bad medical choices would hurt them just as such choices hurt you.

I first read this as

> Thus bad medical choices would hurt them just as much as such choices hurt you.

and wondered how that could be. If the choices are to hurt then as much as they hurt me don't I have to hand over the full value of my good health? Scary!

The article doesn't quite say that. Perhaps I hand over less and the choices hurt them less than the hurt me, leaving the incentives weaker than ideal. If I'm serious about analysing the posts proposal I have to regard it as a parameterised proposal and analyse the logic with mathematical formula involving a hurt-leverage ratio.

> Since payments to plans would usually far exceed medical expenses,...

I think this is where the post loses 80% of its serious audience. The obvious meaing here is lumpiness: most years you are well, so pay a hefty fee that far exceeds your medical expenses that year. However the post is about something else. Everytime a healthy young person is fatally splattered in a motorcycle accident, the health care provider is seen to close a hugely profitable account. In essence the stream of payments constitute an annuity with the health care provider as the beneficiary and the patient as the annuitant. The annuity is valuable, so the health care provider should pay for it. For a preliminary analysis of the health care provider's incentives, intended to drum up interest in the idea, we should simply assume a fair price. An auction is a pricing mechanism, but we don't need to discuss pricing mechanisms at this stage; that is just a distraction.

> Yes, the government would acquire a financial incentive to hurt your health,...

Once we recognise that the health care provider has purchased an annuity, we also see that the government has sold an annuity. Selling annuities is routine. Pensions companies don't murder their pensioners. Pensioners don't decline their pensions out of fear of assasination. The objection being raised is outlandish and distracting.

You can boil the idea down to one sentence: get your health care provider to purchase an annuity based on your life with itself as the beneficiary.

A follow up question is: Are third party annuities underpriced? Life assurance companies take adverse selection into account. They price their polices to reflect a bias towards long life amongst the actual purchasers. Do they take active management into account? That is, do they allow for beneficiaries paying for life extending treatments to prolong the life of the annuitant and thus extract more money from the policy provider? If not, why not?

Evil, I'm describing a mechanism that can be used by government or private parties. I just illustratred it first for government.

Biomed, "Hello Mrs. Winslow, sorry to disturb you but today turns out to be your random health audit day. How are you feeling? Can you get to the store? Are you in any pain? ..."

Botogol, I had in mind one-at-a-time auctions, but am open to variations.

Alan, you and your plan are both hurt; there is no need for you to be hurt less if they are hurt more.

Extending the hospice system would palliate the medical cost problem.

Arguing about which people should get less medical care than they need will not get anywhere for the foreseeable future. It will just make people think you have Asperger's and no emotional intelligence and that kind of thing.

The argument is how to cut costs by refusing treatment to people. This is not going to be politically popular. It just isn't.

The time to introduce these ideas is when we're in an extended war and our national existence is threatend. When 70% of the doctors are serving with the military, and hospitals have electricity 10 hours a day, and everybody's making big sacrifices for the war effort, *then* is a good time to point out that we have to cut medical costs for people who can't hope to contribute more to the war effort than their treatment will cost.

We only make hard choices in wartime.

Maybe you can find a way to fudge it, so people don't notice that the choice is deciding who should die.

When I was in Alabama I heard about some chain letters. One of them worked very well -- instead of just sending money to strangers and hoping other strangers would send you money, they had a deal where you recruit 2 people and get each of them to give you $50 and you watch them mail $50 to the guy at the top of the list. You tell them that when they recruit 2 people they get their whole investment back and all the money they get from strangers is just gravy. Lots of people bought into that. I told a med tech why it was bogus. "What's to keep you from putting your mother's name at the top of the list and recruiting as many as you can? Every one gets you $50 and your mother $50." "But that would be cheating!" "Who gave the guy who started this more right than you have to start your own?" Three days later the newspapers denounced it as a scam and it all fell apart. An undergraduate working in the lab thanked me. "I made $350 off that chain letter and so did my mom." And I didn't make a cent. The point is, fuzz out the decisions. Make it look safe. People will volunteer to get all their medical benefits cut off if you can make it look like they win.

Wow. I've said it before, but I thought I was saying it somewhat metaphorically. But the comments here show it to be quite literal: many people are horrified by any health care system that could clearly fail to deliver a medical treatment, no matter how expensive, that might plausibly offer some health benefit, no matter how small. Since every health system must in fact do this, then such people can only not be horrified by unclear health systems.

Robin, yes! If a health care system fails to anticipate some needs and so some people lose out by accident, people accept that. Bad things happen sometimes, nobody has enough foresight, etc.

But when somebody has the job of choosing who lives and who dies, that person has too much power to suit us. So we have to hide how it happens. Sometimes the choice is made by a woman with a high-school education working in a windowless office building in chicago, who's supposed to apply an insurance company's guidelines and who has a degree of flexibility within the general guidelines that spending has to stay in reasonable bounds. She decides individual cases based on a few lines of coded medical information and a company database and her quota. Will your treatment be disallowed? Nobody knows until she pushes the keys. If you'll die without treatment you can sue and hope you don't die before the court resolves the issue.

We certainly don't want to have a known individual who can make that choice and have it be final. If he cuts off the care you need to survive, what would you offer him to restore it? What would your wife and daughter offer him? Don't go there.

Possible solutions include -- project the costs of new treatments before you finish developing them, and defund the ones that will be too expensive. Keep careful statistics on alternative treatments and promote the ones that are cheaper but work just as well. Perhaps there could be subtle encouragement for that sort of thing -- like, MDs who discover cheaper treatments that work just as well get a fraction of the savings for their further research, but MDs who discover better treatments that are more expensive get nothing but praise. Fund public health approaches to improve the public's health. Promote hospices for people who've already enjoyed as much medical care as they can stand. Etc.

Admitting that we are telling people to go off and die, we don't think they're worth keeping alive, is mostly workable only during popular wars. We could have told old people they needed to take their chances just like the boys at the front during WWII. We can't be so blunt today.

Robin, I'm surprised you haven't commented on this.

Robin, two things
1) when I worked in a hospital we used to do a treatment for certain people that cost nothing. As far as I know, that treatment worked 999,999 times out of a million. The current treatment cost $5,000 (includes MRI...) and I have never seen evidence that it works any better.
So one problem is that expensive new technologies have to be used because the machines cost alot.
2) What would the health care system look like if we rid the system of malpractice insurance?

Evil Mike: Why don't you compare government health insurance to private health insurance?

You might be surprised; the government can do many things more efficiently than industry ever can, because the government is not required to seek profit - so it doesn't have to include the inefficiencies whose only goal is profit.

Of course, it needs other reasons to avoid inefficiency - which typically happens when it's forced to be open and transparent. (Again, something industry never does ...)

Hi- I tried to find this in the existing comments, but couldn't.

One person did say "pain" (or quality of life) may be hard to measure. Not only is it entirely subjective, it leaves the plan entirely defenseless against hypochondria. And not only full-blown hypochondria, but any variant thereof.

People respond differently to stress. A car accident may put one person in the hospital for a night, and another in a wheelchair (or with a cane) for 30 years.

So we need them to feel their pain, too. We don't want to penalize them for being healthy (pay for health). I agree that rewarding the healthcare system when people are sick is also dangerous, so perhaps a blended approach is warranted.


Carl, pain isn't that hard to measure; it is done all the time. The patient would not be paid to be in pain, so it is not obvious why they would want to lie about their pain.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30