Friday, July 25, 2008

Ezra Klein writes about smart cost sharing.

He wants a committee to decide reimbursement rates.

Oddly, I had another idea about smart cost sharing. Make the doctors pay for the care and pay the doctors based on outcomes.

update: I should have mentioned that I am working on an idea I got from Mark Thoma

...preventative care ... ought to be encouraged, and one way to help with this is ... to forge an unbreakable lifetime relationship between the insurance company and the consumer so that expected lifetime costs are important to the insurance carrier.


I strongly suspect (with no evidence) that Thoma's thoughts were influenced by
an empirical result that very small financial incentives to doctors based on their patients' blood glucose caused big changes in those outcomes (pdf warning).
which will save huge amounts of money for medicare but small amounts of money for the HMO's that introduced the incentives.

One politically unfeasible approach to this would be to assign people randomly to HMO's and pay the HMO's based on their health but have the HMO's pay for their health care. Then the HMO decides incentives. You have to decide how much a life is worth (and eyesight and all that) but it doesn't depend on individual income and the decisions are made by an orgnaisation with tons of data.

Now no way are Americans going to give up all choice (even in Italy I got to choose my GP). So there would be a huge huge cherry picking problem. One could try to deal with it by charging the managed competitive insurance plans (that's not English it's Magazinerish) based on costs per patient minus predicted costs given region and patient characteristics and rewarding based on outcomes minus predicted outcomes.

Obviously it would never work.

I think the best we can do is to charge medical costs not just to the current insurance plan but also, in part, to the one that covered the patients in the past (to give the an incentive to keep their clients healthy). That has the effect of partially funding medicare with a tax on health insurers which would be OK since it is insolvent. Plus paying insurance plans based on documented improvement in well the 3 blood things say.

If insurance companies saw obese people with horrible eating habits who watch TV all day as a profit opportunity, the USA would be a healthier place.

Just think, sleazy insurance agent tells his boss (hey I just found someone with an LDL level of 300, we got to move fast before our competitors sign him).


To try to explain better

My plan is the Edwards plan plus insurance companies pay for care of former clients based on alpha(cost of the treatment)*(years with that company)/(age at time of care) where alpha is well below one and for the care of current clients minus the part paid by former insurers. They get paid a constant which depends only on the region where they are located times the same alpha factors.

Thus they have an incentive to keep their clients healthy (which they can pass on to doctors).

Plus they get paid based on progress on preventive measures (patients who quit smoking, got blood pressure from x down to y, lost weight from obese to not obese etch)funded with a tax on insurance companies per patient so on average they get zero.

This means they would be more willing to sign fat lazy smokers as there is lots of room for improvement compared to things as they are.

Update: Mark Thoma linked to an older version of this post. Following the link back to him, I was reminded that I am talking about how to implement his idea. Also, since some people are actually reading this post, I looked up the cite on the effectiveness of incentives.

Now having stolen Thoma's idea I will steal his comment thread too.


More Expensive says...


Taxing insurance is probably the wrong way to go. More expensive insurance means fewer people/employers will buy it. Finding a way to lower insurance costs is the way to go if we want more people covered. Of course, lowering insurance cost means lowering health care cost in general..

Cost all comes down to can we trust paying customers (patients) to make the right choice. If not, the free market approach won't work. Customer choice means no more competition stifling licenses or prescriptions. The customer must be in complete charge, free to choose an unlicensed Wal Mart trained doctor if she wants to, and buy her meds OTC from Amazon.

All other advanced nations have decided that the customer can't be trusted, so they use a paternalistic state run system. Our system is dysfunctional because we don't empower the customer with true choice (competition), but still expect market forces to work normally.

Posted by: More Expensive | Link to comment | July 25, 2008 at 03:21 AM


My proposal began with "start with the Edwards plan" which mandates coverage.



save_the_rustbelt says...

"Oddly, I had another idea about smart cost sharing. Make the doctors pay for the care and pay the doctors based on outcomes."

That sounds fine, until you meet a "non-compliant" patient.

The diabetic who loves cupcakes, patients allergic to exercise, those who cannot quit smoking, etc.

Next idea?

Posted by: save_the_rustbelt | Link to comment | July 25, 2008 at 04:55 AM
ken melvin says...


This is a really interesting comment. I wrote hundreds of words in reply back at economistsview then I surfed away without posting them (Robert Waldmann's other idea make sure that Robert Waldmann has nothing to do with managing health care).

I know that there are non-compliant patients. My sister treats the homeless. My mother treats HIV positive people without insurance (by now most are intravenous drug addicts the nice polite gay men have learned to use condoms). They tell me about their patients using nicknames to preserve confidentiality. I have heard a lot about "Eeyore" (a very gloomy patient such that I recognised that my mom was talkinga bout Eeyore when she just said "a patient") and "Rage" (primary complaint on first contact was "rage" he said it first).

There is a huge literature on optimal incentive schemes when outcomes are not completely under the control of the agent. If agents (here doctors) are risk averse it is not optimal to pass all costs on to them, since then they demand high expected payments to make up for the risk. That is why I proposed incentives at the level of the insurance company. According to standard economic theory they, as rational profit maximizing agents, will decide what incentives to pass on to the doctors. They can just take the hit for poor patient outcomes (note the fraction of non-compliant patients for a whole company has very low variance. It's like the point of insurance). Or they can pass them on to doctors *and* raise expected compensation so doctors accept their patients. According to standard theory, they will design the optimal incentive scheme (and CEO compensation is optimal for shareholders). I admit that I don't care at all what standard theory says as I don't believe in it.

I think part of the issue is that we have a strong sense that we must never punish the innocent. This means that making a doctor pay for having a non-compliant patient is unfair and unjust. If this logic were applied consistently, we would the market system which provides incentives which depend in part on individual choices and in part on luck (I'm not sure Mr Rustbelt would disapprove of such an approach). Somehow it's different if the government is involved. I once heard someone angrily opposing a gas tax (in the 70s) arguing that some people need to drive to get to work -- so it would punish people who weren't doing anything wrong. Look I'm not talking about throwing people into jail. I'm talking about an MD getting $50 less in a year, because he didn't manage to convince a non-compliant patient (note MDs can subcontract the nagging to nurses).



First, how best provide healthcare. I see no reason to believe insurance companies will ever provide the answer. If they were somehow the answer, responsibility would have to be continuous.

Posted by: ken melvin | Link to comment | July 25, 2008 at 06:34 AM


Nor do I. I was playing by the rule "you can't get single payer through the Senate," which is standard in the wonkosphere.


Ninja Zombie says...

"If insurance companies saw obese people with horrible eating habits who watch TV all day..."

Therein lies the problem with american health care. But what can insurance companies, socialized medicine or any other *medical care* payment scheme do about this? Send out "Dear Fatass, please exercise to save us money" letters?

The only thing that has any reasonable chance of success here is creating incentives for diet and exercise. One possibility: everyone buys their own insurance, and Chubby McLazy pays higher premiums...

Public ridicule might also work.

Posted by: Ninja Zombie | Link to comment | July 25, 2008 at 06:53 AM


Your hypothesis has been tested and rejected by the data (see big pdf). To fantasize, the insurance companies can pass incentives on to the MDs who pass them on to patients. If the doc gets $50 when the obese patient becomes non-obese, he or she can offer a prize to the patient who loses the most weight (biggest loser). That works (just watch it on TV). In any case, patient's behavior can be influenced. It can be done. It has been done.

GK says...

I guess that I have to bring this discussion back to reality.

My insurance company loved me when I was healthy, after my battle with cancer they just keep raising my rates every year, even though I am considered "cured" from the type of cancer I had.

Insurance companies are "For Profit" corporations that hate sick people.

Fuck the insurance companies, let the Feds take over Health Care like Medicare.

The rest of this discussion is just ivory tower babble


Absolutely true. The departure from reality was explicit when I wrote "Start with the Edward's plan". The idea is to completely change incentives to insurance companies so that they make profits by keeping people healthy. It is theoretically possible. I know it aint gonna happen.

Posted by: GK | Link to comment | July 25, 2008 at 07:54 AM
Holly W. says...

How do you have competing insurance companies and an "unbreakable" lifetime relationship between the insurer and insured?

Posted by: Holly W. | Link to comment | July 25, 2008 at 08:04 AM


Good question (which almost hints that I added something to Thoma's idea)
.
That's what the alpha is for. Costs and income are shared between the current and the past insurer. An insurance company still gains by getting a client from another (alpha is less than one). The old one gains if the patient stays healthy (alpha is greater than zero). Look friendship can last a lifetime, but people still try to "win friends and influence people". There are numbers between 0 and 1.

save_the_rustbelt says...

I jsut received an e-mail newsletter from my health care insurance company, with a nice column written by a doc about why eating healthy is good for me. I guess I head for the gym and give the M&Ms to my wife.


People can be influenced. The claim that it can be done is proven by the fact that it has been done. Click the (warning pdf) link which I irresponsibly didn't have in my first post. Also see (ivory tower babble)^2 about MDs offering prizes to patients.

Gk: sorry about your experience, but Medicare IS an insurer, and doesn't cover everything

Posted by: save_the_rustbelt | Link to comment | July 25, 2008 at 08:15 AM

It is way better than private insurance companies. This is mainly because seniors are politically powerful and congress throws public money around (see prescription drug benefit no bargaining over price $1.8 billion/yr or so plus medicare advantage $6 billion per year or so.

However, it is also true that medicaid is stuck with its clients and itself saves nine for a stitch in time. It's generosity would support my arguments if it were a profit making company (and simultaneously my grandmother with wheels).

ken melvin says...

Guide dogs assist the blind. Dogma makes one blind.


did you ever notice that "Dogma" spelled backwards is "A.M. God" ?

No comments: