Asymmetrical Information

I read the article in the NY Times Magazine and agree with your description of it. I suspect that it is a salvo in her "non-campaign" for vice-president and nothing more.

Some of her ideas are interesting, to say the least:

"Why rely solely on the doctor's brain to store that information? Computers could crunch the variables on a particular patient's medical history, constantly update the algorithms with the latest scientific evidence and put that information at the clinician's fingertips at the point of care.

Americans may not be getting the care they should 45 percent of the time, but the tools exist to narrow that gap. Research shows that when physicians receive computerized reminders, statistics improve exponentially. Reminders can take the form of an alert in the electronic health record that the hospitalized patient has not had a pneumonia vaccine or as computerized questions to remind a doctor of the conditions that must be fulfilled before surgery is considered appropriate."

If I were to distort, I would say that she's advocating that computers would be choosing (or at least) guiding doctors' decisions. Obviously, they do already, although not to the extent that she would like. (Would cost-cutting accountants be in charge of the software?)

This may have served as a campaign ad for Sen. Clinton; it may also be a difficult albatross to hang around her neck come 2006. ('Hillary wants to take medical decisions out of your hands and out of the hands of trusted doctors and give them to computers. Will you let her?')

Not quite "no one",
Sydney Smith has some comments on it.

I have not read the article. Both Klug and Ms. Smith mention that Ms. Clinton discusses computers; there is some evidence that the use of expert systems (in conjuction with a skilled clinician) can greatly improve diagnostics. Of course, who knows what the good senator had in mind.

"Will her Republican opponent, in the Senate or the next election, be given a similarly large bloc of space to make his arguments?"

When did Asymmetrical Information turn into a humor column?

You write as though the New York Times explicitly donating editorial space to Hillary Clinton's campaign is somehow surprising.

Also you seem somewhat surprised that her diatribe is illogical totalitarian drivel.

And you expected something different? Hillary spontaneously transmogrifying into a giant winged monkey, flying off into space, and entering orbit around the moon, perhaps?

This may have served as a campaign ad for Sen. Clinton; it may also be a difficult albatross to hang around her neck come 2006. ('Hillary wants to take medical decisions out of your hands and out of the hands of trusted doctors and give them to computers. Will you let her?')

Sure I guess it could but how convincing would it be? Doctor's descisions are already 2nd guessed by HMO's looking to save a buck. Who wouldn't want their doctor to have a computer program that would warn him if two drugs don't mix well? In fact, many doctors would probably be in favor of such a thing since it could cut down on their malpractice insurance!

Anyway, in regards to Jane's question; I'm sure the NY Times Magazine would run a piece on healthcare by her next Senatorial opponant if that person meet any one of the following conditions:

1. Was a former first lady.
2. Was known to have been a major figure in health care reform proposals.
3. Has widespread recognition & was widely regarded as a politician of some importance.

TO help illustrate this point; Would Jane have bothered to carp about 'free airtime' if the author of the article was the Representative from kalamazoo?

Boonton: if it was as bad as this piece was? Yes. And as far as I know, allowing a sitting politician this much magazine space is unprecedented -- never mind that the article is awful.

I guess I'll have to read it to see if it was as bad as you claim. The part about using computers to improve healthcare wasn't too bad (not novel but not bad).

Jane: I, too, found the long essay pretty uninteresting, though I took exception to her use of the questionable "43.6 million Americans are uninsured" figure.

A modest proposal: Allow consumers of health care to contractually waive their right to sue for malpractice before care is provided in return for a reduction in fees equivalent to the cost savings the provider realizes by not having to practice "defensive medicine" and insure against malpractice claims by those waiving their right to sue. Obviously enabling legislation will be necessary [my local Oregon State Senator is intrigued] and the legislation will have to compel health insurers to also pass through savings realized if their insureds enter into waivers. But doesn't this empower consumers [minors and those not competent to contract obviously won't be able to waive], and give them a [libertarian] choice to "trade" an inchoate right to sue if something goes wrong for a real, immediate cost saving?

The fact that you can't get insurance if you have a pre-diagnosed disease may not be a failure of the insurance market but it sure is a failure of the health system.

But much of that will become moot over time. Given advances in genetic testing and privacy concerns it's a question of time before insurance companies are driven out of the healthcare market altogether, at least for major illnesses. Clearly only the government, directly or through regulation, will be able to provide healthcare under these conditions.

Why don't we try address that now?

I am just an undergraduate economics student, but I confess to be throughly baffled by the above comment. Specifically, I don't understand how having MORE information would somehow make insurance unavailable. My understanding of the "moral hazard" problem (i.e. differences in information available) is that the difference in the information available to buyers and sellers causes sellers to only offer lower-quality items, since buyers have no way of differentiated and thus can't offer the correct (market clearing) price.

So, shouldn't more information available lead to a price tailored to the individual, rather than the current system where people who live healthy lives have a large incentive not to buy insurance? Since the premiums are more "community based" in the current system, there is less incentive to live healthier (since you don't get rewarded financially) AND healthier people have to overpay for premiums which are based on the overall community, driving prices still higher.

If I have misunderstood moral hazard, please feel free to correct me; I'm no expert. But I have never heard of a market being harmed by less asymmetrical information. (Note: I'm not discussing whether it would be a good thing to have premiums based on genetic risks, just the confusing statement that somehow the market would disappear. I don't think it would; rather, we would no longer have the redistributional system we have now, away from the "genetically advantaged.")

Because the information will only be available to the individuals. If I can get a genetic screening that tells me I'm not at risk for any major diseases and you find you have a high risk of some catastrophic illness who do you think will chose to get insurance right away? Both life and health insurance.

That's why already insurance companies are fighting to have the right to prescreen.

The information would be available only to individuals? Isn't that roughly the situation we have now? I mean, I as an individual know full well what I consume, how much I drink, how much I exercise, whether I take a mutivitamin, and a thousand other little factors which affect my health costs. My insurer (I'm a student, so I haven't actaully been in the market yet) cannot measure these variables, only crude proxies of them.

Again, the market failure problem is a result of NOT allowing pre-screening (by the insurers), not of scientific development. If pre-screening were available, then the market should be more clearing than it is now.

Again, whether or not this is a good thing is debatable. But I don't think one can rightly fault this as a situation of market failure.

It's not available to individuals today because the tests I mention still don't exist.

Yes, if you allowed the insurance companies to pre-screen it would eliminate the problem they face but create many, many more. This is a nation that rejects a national ID card, where millions vehemently oppose registering their guns, an inanimate object. You think they will accept giving insurance companies the right to run DNA tests on them when testing will let the insurance company know so much about them?

And even if they did, what would you do with the millions that find they are highly susceptible to catastrophic illnesses? They are the ones that most need the coverage and, as Jane points out, there is no price at which they can be insured (particularly if the tests are good enough). You'd have a situation where the people that the insurance companies want to insure will not be that interested and those that most desperately want to be covered are shunned.

That's why the end result is clear. We WILL have some sort of national coverage for major illneses. The soner we accept this and start working towrads this the better.

I'm confused. I distinctly remember a federal law being passed a few years ago that prohibits health insurers from refusing coverage because of a pre-existing health condition. In other words, one can move from one job to another, and consequently move from one insurer to another, without worrying about losing coverage for a condition that began under the first insurer. This used not to be the case, and was of personal concern to me, so I greeted that particular law with relief.

Rex,

I think that only applies to group insurance and not to individual insurance.

And even for group insurance they can make you wait for up to 18 months under certain conditions.

GT-

I'm not sure mandatory national health insurance for catastophic illnesses is unwarranted (though I really have trouble envisioning why this would change based on genetic testing--if people have the right to be taken care of in the event of a catastrophe, I don't think technology would change this).

Rather, I think the problem is on the day-to-day treatment costs. Do you think it warranted that a person with a healthier immune system should be FORCED to pay for the same, say, prescription drug coverage as someone with a weaker one--or would you advocate removing "ordinary" medical expenses from the insurance system altogether?

I don't know. We'll have to figure out the details as we go along. But given the general demand for more and more health coverage (see the new Medicare drug benefits) I suspect more, not less, will be covered.

We are going in that direction. Already the US spends more in public healthcare than most european nations and yet our basic coverage services are much worse than theirs. Add to that the impact of genetic testing and I think it's time we start looking at this seriously.

Already the US spends more in public healthcare than most european nations and yet our basic coverage services are much worse than theirs.

I wish this particular canard would die out.

GT, as someone who has both studied this as an interested observer with a medical background, and been a resident of both Canada and Germany, I can tell you it's simply not true. Just not. I can tell you lots of examples, but in fact it isn't true.

I believe GT is on to something here. It is entirely possible that advanced genetics will make it that all of us are basically uninsurable in that an individual application will always ask if we know anything at all about our own health. If we don't share what we know it is fraud.

I submit to you the recent discovery of the "heart attack gene" by the good folks at the Cleveland Clinic. It is now identifiable. So if a baby is born with this gene is it appropriate to deny coverage? What if the child grows up and does everything possible to take care of himself - proper diet, exercise, no smoking - a perfect discipline. She has a heart attack at age 42 anyway - predestined. What then? She was more than happy to pay premiums for years just so she would have coverage. Should she have been refused coverage - or had it "ridered out"?

I think this type of scenario is definitely part of our future. I don't come to the same conclusion as GT in terms of a national health care solution - but I think his point is quite valid.

Chris, consider what would happen if the insurance company had perfect information. That is, it knew exactly what medical conditions you would develop and what treatments would be required. Then they would set the price at what they expect to pay out, plus their overhead costs and profit. You would always be better off paying the bills directly and avoiding the insurance company's cut (tax considerations aside). I don't think we'll ever be there, but the better testing is available to insurance companies, the closer we'll get. If you legislate that people can keep their genetic and other medical-predictive test results secret from insurers, then the insurers are bound to assume that those that want to buy insurance KNOW they are about to run up big bills, and will price accordingly.

What's confusing about medical insurance is that already much of it isn't really insurance so much as a tax-deductible bill-paying service. Major medical coverage makes sense as insurance - you pay a small amount to be insured against an unlikely, unpredictable, and very expensive possibility. E.g., in my almost-bankrupt employer's new medical plan, the important part isn't that it pays about 50% of the first $10,000, but that it pays 100% above $10K - if my wife needs a major operation costing $300K, I might have to take out little home equity loan, but I don't have to sell the house and qualify for welfare for the other 150K.

But most of what med "insurance" does now is pay routine and rather predictable bills, with the cost plus at least 50% (for overhead and profits) passed directly on to employers and employees. It would not make financial sense at all to buy insurance with a deductible less than your average annual medical expenditures except that the insurance payments are deductible from social-security and medicate taxes as well as income tax. If you paid the bills directly, you might be able to deduct them from income tax (if they are big enough to be worth losing the standard deduction and filling out mountains of forms), but not from SS/Medicare. That's a massive subsidy granted by Congress to one particularly inefficient method of bill payment.

It is not possible for genetic testing to make everybody uninsurable. Since the average rate must remain the same, simple arithmetic shows for every person whose rate goes up by a factor of 10, nine will have their rate drop to zero or more realistically ninety will have their rate drop by 10%. I suspect rates would go down for 80-90% of the population and would only go up a lot for a small fraction perhaps less than 1%.
Even these people would not be uninsurable (everybody is insurable at some rate) but for some the insurance might well be unaffordable. A solution to this problem is to do the tests prenatally and abort those babies.

markm, insurance has advantages besides taxes. The insurance company can and does negotiate better rates with health care providers than an individual can.

Those are some fair objections to my original point, I concede, and some of them I will have to think about. One of them I disagree with is that people "won't allow" insurance companies to have access to their genetic information. Perhaps I am unusual, but I really don't consider that to be much more private than the massive amounts of data that would already be gathered by an insurance company, as it is today. In any case, I don't understand precisely why an individual with a "clean genetic bill of health" wouldn't be willing to use this information. Again, you have this problem of redistributional effects, in that you create a situation where the genetically blessed are no longer bearing the costs of the less fortunate, but the net effect of allowing both sides access to the data (which is what would have to happen, otherwise you really would have the end of insurance) would be a lowering of the market failure problems. I don't see how this isn't as likely an outcome as government monopolization of the health care insurance industry. But if I could, say, surrender a little bit of privacy to lower my insurance costs, frankly I think that is a trade-off that is mine to make.

On the topic of medical informatics, I have long advocated a serious and dedicated approach to implementing decision support systems for clinical use. The algorithms are already there (eg. bayesian belief networks, case-based reasoning) all that is needed is the encoding of the knowledge.

Such systems would allow less skilled people to diagnose/treat a large number of common illnesses and thus expand the overall supply of health care and (hopefully) reduce the cost of much of the routine health maintenance.

Charlie, read clearly what I wrote. Those are facts, not opinions.

Chris, you will onoly be willing to 'share' infdormation if you know in adavance that you have a clean bill of health. If it turns out you have a near-certainty of some major disease in the future you will not want to share that with anyone.

Not only will you become uninsurable (or, what is the same thing your insurance premium will reflect that risk and become unaffordable to you) but you run the serious risk of becoming unemployable. Who would want to hire you if you had 90% chance disability in 10 years, for example. You think you will be able to keep that information just with the insurer?

I wonder if there's an incorrect unstated assumption here someplace.

Genetic tests can only reveal a pre-disposition to most diseases, not a certainty that they will manifest themselves. So discovering that someone carries "the heart-attack gene" may cause their insurance costs to go up after a certain age, with possible discounts for a healthy lifestyle, regular blood-pressure and cholesterol screenings, etc. Costs will be made more group-specific, just as auto-insurance costs are specific to the driver's age, location, prior driving record, etc. etc.

GT, it is not true that people are only willing to share information if they know in advance that the information is good. High value life insurance policies require a medical exam which may reveal serious problems the applicant was unaware of. People still buy them. For example that is how Magic Johnson found out he was HIV positive. Furthermore someone who knows he is at higher risk also knows he has a greater need for insurance. Where do more people buy earthquake insurance California or Kansas? Tornado insurance?

Yes, of course I would want to only share "good" genetic information with my insurer. But, if I had 3 DUI convictions, I wouldn't want that information to be provided to my auto insurance provider, but I don't get to make that choice. The point is that given the choice between the current system and a system where BOTH the insurance company and the individual get to see, I see no a priori reason to assume that I would be worse off.

Lads, you're getting yourself twisted up here because you need to remember that insurance is a market in which timing matters.

For example, Jane wrote:

a real "market failure" would be if there were companies willing to write policies for risks with a 100% chance of occurring, for any premium less than the cost of paying for the treatments out of pocket.

Death is a risk with 100% chance of occuring, and yet it gets insured. The reason is that the insurance company receives the premium up front and is able to make an investment return on it in the time between getting the premium and paying out on the claim. This is a fundamental part of the economics of the insurance industry.

This ought to give you a clue that there are two sources of risk which an insurer has to manage:

1) The risk that a claim will happen
2) The risk about when the claim will happen.

There's been decent actuarial work which suggests that, even if I had perfect information about all the issues relevant to the health of my customers, it would still probably only reduce the volatility of my insurance liability by about 20%. For big risk pools, the uncertainty is predominantly in the timing of claims.

And here's the issue; there's a problem of composition here. Any individual "genetically inferior" person might be a money-losing proposition at a given price, because the insurer will end up paying out more than the premium received plus the investment return on that premium. However, the class of genetically inferior individuals, taken as a group, might be a value adding proposition at the same price, because a) they expand the risk pool, and big pools are more efficient than small ones (pooling risks makes them smaller), and b) the correlation of their risk with the other risks the insurer is running is likely to be less than perfect, providing a gain from diversification.

The wrinkle is that the insurer can't write a contract with the entire class of genetically inferior individuals, so he has to make the underwriting decisions case by case. And case by case, they're money-losing.

Hence, a situation in which more information can destroy a market. There are a number of cases one can construct in which insurance markets don't have equilibria (Chris: your college library might have a copy of "Current Issues in Microeconomics" by John Hey, which has a nice model of how an insurance market can get into a situation where neither the pooling nor the separating equilibrium is sustainable).

I need to clarify here, D2; I was talking about people in the article who had already gotten some expensive disease, who were looking for insurers to insure them against getting that disease. An insurer who did that would be lunatic.

Nonetheless, I agree that GT's example doesn't quite capture it. First of all, as someone else pointed out, we can't all be uninsurable; for every group of people whose insurance gets more expensive, some other group would have to see theirs get cheaper.

But also, genes don't tell the whole story and never will. Most of the human condition is, as good liberals have been telling us for a while now, a combination of genes and environment. Identical twins don't all die of the same thing at the same time; siblings have an even lower connection. Moreover, when you think of how many things there are to die from, and how many genes, it will be a long, long time before all but the most deadly genetic diseases, which afflict a very small fraction of the population, will be identified. The things they're coming up with now have been selected because they're highly unusual, and therefore easy to identify -- things like Huntingdon's or Brca1, where entire families die of the same thing around the same age. How many families do you know like that? Besides which, the information technology that allows us to discover these things could also, in not-too-distant futures, provide the cure, making this whole discussion moot. Other advances in medical technology might do the same. And finally, at lower risk levels, genetic screening could give individuals the tools to avoid these expensive health risks -- those with high lung cancer risks can avoid smoking, those with high heart attack risks can modify their diet and excercise. So I don't think these issues are nearly as clear cut as the favourite genetic testing horror stories of the press. Plus, as D2 points out, as long as the timing is in doubt, many conditions may still be insurable.

Remember that my example is based on a specific premise, that someday we will have the equivalent of what you see in the movie Gattaca. And of course, it depends, as Jane points out, on there not being an easy cure so that you end up identifying the risk with high certainty and not beiing able to solve it (or only able to do so at great cost). Also that it won't be for just a couple of diseases but for a whole gamut of them.

If I am wrong about the medical outcome this scenario will not happen. But talks with people in the industry have led me to believe they see this as potentially a big problem in the future.

http://www.insurancejournal.com/news/international/2000/01/03/12638.htm

Plus, even for those cases where an individual shows an unacceptable (from an insurance standpoint) genetic tendency for a particular disease, that does not mean they would be uninsurable. They would likely just have an exclusion added to their policy for that particular disease. If it someday becomes possible to determine that I have an 80% chance of developing cancer in the next 5 years I probably would not be able to be insured for medical costs related to cancer treatment. But there is no reason I could not get insurance to guard against heart disease, appendicitis, car accidents, etc.

We would need to come up with some method of providing coverage for the uninsurable risks, but I can't see the medical insurance market collapsing because of this.

Sean,

But what if you also find that you have almost no risk of heart attack while someone else finds the opposite (that he is prone to heart problems but not to cancer).

Remember that is the premise of my argument.

In that case you would have little interest in getting health coverage for heart problems and the other guy the same for cancer. As you add people up this could easily make parts of the health market simply uneconomical for insurance companies.

I am not saying that ALL healthcare insurance would be in trouble. After all there are many diseases and health problems that are not genetic at all (including the risk of an accident) and people will want to cover for that.

But if (a big if but not unreasonable given medical advances IMO) genetic and other testing advances to a "Gattaca" stage then the role of government will have to increase for these diseases.

If so I think government (through taxation) will take care of the major diseases that can be detected but not easily cured or prevented and private insurance will be there for the rest.

Yes, of course I would want to only share "good" genetic information with my insurer. But, if I had 3 DUI convictions, I wouldn't want that information to be provided to my auto insurance provider, but I don't get to make that choice.

Bad example, IMO. DUI convictions are the fault of your own previous irresponsible behavior. Genetic information is nothing you control in and of itself; it's just there. Perhaps both are similar risk factors from the insurer's point of view, but to treat them that way raises different sorts of moral dillemmas.

" Besides which, the information technology that allows us to discover these things could also, in not-too-distant futures, provide the cure, making this whole discussion moot. Other advances in medical technology might do the same. And finally, at lower risk levels, genetic screening could give individuals the tools to avoid these expensive health risks -- those with high lung cancer risks can avoid smoking, those with high heart attack risks can modify their diet and excercise."

Exactly. Which is why we should move from the present system to one of a combination of Medical Savings Accts and catastrophic insurance only.

A clarification if you please. My understanding of the heart attack gene discovery ( Business 2.0 magazine article a few months ago I think) is that it will indeed cause the person to have an attack. No question - as the article put it if they aren't hit by a bus or they don't get cancer they will for certain have a heart attack.

Whether the same medical science and genetic research will solve the problem as JG suggests is certainly quite possible if not probable. But that in know way changes the dynamic of insurance underwriting - and the fact under current law and custom that if you have the gene and know you have it and withhold that information from the insurance company you have committed fraud.

"In that case you would have little interest in getting health coverage for heart problems and the other guy the same for cancer. As you add people up this could easily make parts of the health market simply uneconomical for insurance companies."

I think that assumes that the gene is like an on/off switch - either you'll get cancer (with a high probability) or you won't. Isn't it more likely that for most conditions risk levels would fall along a wide range, with some people being high risk, some low-to-no risk but most of us falling in the middle? The high risks would be uninsurable, perhaps requiring some sort of government intervention, the low risks wouldn't care but for most of us it's business as usual.

I'm sure there would be a few diseases where genetics is the only factor, but I would think these would be the exception.

If you are saying that genetic testing may mean that in the future government intervention may be needed to supplement private health insurance in some areas, OK. But your initital post made it sound inevitable that the US would be forced to move to nationalized health care someday, so why not just get it over with.

Any, agree with GT's premise or not, he did do a helluva job getting the topic off of Hillary. ;)

For starters I don't know that National Health Care or Single Payer Plans have any real future. Sorry if I gave that impression.

That said, if I understand genetic research (big if) we could very well find out for each and every one of us the exact medical problems we will inevitably have. If this is so, it changes the fundamental precepts of insurance as we now know it to be. Health insurance and life insurance are quite different - the difference between mortality and morbidity is profound. Just because I am going to have a heart attack does not mean I am going to die as an immediate result (especially since I can prepare for it) but it does mean I am going to incur some large medical expenses.

I also think it is most likely that we may indeed discover the pharmachological (sp?) solutions to any genetic problems - EVENTUALLY.

The issue is IMHO, that this time lag (which might be quite long if we continue to demonize big pharma) will be the catalyst that will drive us to our inevitable solution: 1)forget about insurance altogether and adopt a national health care plan or; 2)rethink insurance underwriting and eliminate medical qualifications - but retain "lifestyle guidelines" such as smoking, obesity, etc.

Personally, I am all for the insurance solution but I am quite clueless as to how this will really play out.

Comments are Closed.