The Puzzling Consensus in Favor of the Genetic Information Nondiscrimination Act

The Genetic Information Nondiscrimination Act, which bans certain types of genetic discrimination by employers and insurers, passed the House by a vote of 414 to one, and the Senate by a vote of 95 to zero. That means it's a good idea, right? Wrong.

Suppose an insurance company wants to offer a policy capped at $10,000 over a year. It has two types of potential clients: high-risk types who have a 0.05 risk of suffering a $10,000 injury and low-risk types who have a 0.01 risk of suffering a $10,000 injury.  In expected terms, the high-risk types cost the insurer $500 each, and the low-risk types cost the insurer $100 each.

Consider the following question. If the insurance company can distinguish potential clients on the basis of easily visible markers (such as age), do you think it should be able to offer an expensive policy for high-risk types ($500) and a cheap policy for low-risk types ($100)?

If the insurer can do this, then everyone gets insurance who wants it. If the insurer can't do this, then fewer people will. The insurer can't offer policies for $100, for then the high-risk types will snatch them up, and the insurer, receiving $100 and paying $500 per person, will go out of business. If it offers policies for, say, $250—the average cost for the two types—the business will still probably not be sustainable. All or nearly all the high-risk types will buy the policy, while many low-risk types will be reluctant to pay so much for insurance against a low risk. The average cost will thus rise above $250, while receipts will continue to be $250 per person. Perhaps the insurer will offer only $500 policies, in which case half the population—the low-risk types—must go without insurance that they desire.

Most people agree that insurers should be able to discriminate on the basis of risk. We don't expect a 25-year-old to pay the same premium for life insurance that a 90-year-old must pay.

Suppose, then, that the types cannot be distinguished on the basis of a visible marker, but a simple checkup with a doctor will determine which type a person belongs to, perhaps based on a blood test that determines whether the person currently has a dangerous disease. Should the insurance company be permitted to offer the cheap $100 insurance policy only to people who obtain a doctor's certification that they belong to the low-risk group? If you think that insurance companies should be able to discriminate on the basis of visible markers such as age, you ought to think that they should be able to discriminate on the basis of doctors' certifications. If the insurance company should be able to deny insurance to a person visibly dying from a disease, then it should be able to deny insurance to a nonvisibly dying person on the basis of a blood test. There is no morally relevant distinction between looking at the person's outer shell and looking at his blood under a microscope.

Suppose, now, that a person's risk type is based not on a simple blood test that determines whether he is infected with a particular disease, but on a genetic test that determines whether he has a greater than normal susceptibility to a particular disease. Should the insurance company be permitted to offer the cheap insurance policy only to people who obtain a doctor's certification that a genetic test shows that they belong to the low-risk group? If you think that insurers should be able to discriminate on the basis of visible markers and on the basis of simple doctors' tests for the presence of dangerous diseases, then you should think they should be able to discriminate on the basis of genetic tests. There is no morally relevant distinction between looking at a person's blood for the evidence of infection and looking at his DNA for evidence of susceptibility to a disease.

Or, at least, none that I can find.  The only argument in favor of banning genetic discrimination is that employers and others "misuse" genetic information. If this is true, then misuse of genetic information should be banned, not the proper use of genetic information for the purpose of assigning people to different risk pools. And if genetic information can help determine whether a person is suitable for a particular job, perhaps one that is dangerous for some types of people but not others, then it should not be considered misuse for employers to make hiring and job-assignment decisions on the basis of that information—no more than taking into account that person's visible physical abilities such as strength.

Another concern is that Americans are refusing to take genetic tests because they fear that their test results will be used against them by insurance companies and employers. But this is like saying that we shouldn't let insurers condition insurance on a visit to the doctor's office because then Americans would refuse to see the doctor, lest health information be used against them. The opposite is more likely. As genetic tests improve, insurers would require customers to take the tests if they want to purchase the cheap, low-risk-type policies. People would have to undergo genetic tests, just as today they have to visit the doctor if they want insurance.

The only explanation for the enthusiasm for GINA is that there is an inchoate feeling among people that there is something wrong with the way the insurance market operates. After all, as long as insurance is permitted, insurers will offer cheaper rates to lower-risk people, which seems unfair to higher-risk people, especially those who are high-risk because of bad luck in the genetic lottery rather than because of a choice to pursue high-risk activities like motorcycle riding.

But this is like saying that it is unfair for employers to offer higher salaries to people who are talented, and whose talent can be traced, as it almost always can, to a lucky outcome in the genetic lottery. After GINA, employers can still discriminate against a person whose genes have bestowed him with a bad smell, awkward social skills, or a weak grasp of arithmetic. GINA does not ban discrimination on the basis of genetic information. GINA bans only discrimination on the basis of genetic information that has not yet manifested itself in observable characteristics or behaviors but that is likely to in the future. There is no sense in this distinction.

Cross-posted here.

Comments

[Evan]

The problem with Professor Posner's argument is that it assumes that biology=destiny.

The Professor claims there is no moral difference between discrimination based health problems revealed by a blood test or by a medical examination, and discrimination based on genetic susceptibility to disease. I believe this is clearly erroneous. Illnesses that can be revealed by an examination or by a blood test are a matter of certainty. If the blood test reveals the presence of infection in the blood, then the insurance applicant certainly has that illness. Furthermore, if a doctors examination reveals with certainty the symptoms of a known disease or at least the certain indication of health problems.

I believe it's easy to draw a moral difference comes down to the fact that by allowing genetic discrimination you are in fact condemning whole families for generations to come to higher insurance rates, (if they can get insurance at all). For example Bill applies for a policy in a no GINA world. Bill carries the gene for marfan's syndrome so the insurance company declares him uninsurable. Bill's daughter inherits the gene from Bill and is also declared uninsurable, and so on. The fact is no matter how healthy a lifestyle Bill and his family lives they will always be uninsurable.

And therein lies the moral objection to genetic testing. No matter how you live your life you cannot change your genes. If genetic testing becomes the decision making factor in insurability you will not only overcharge individuals who have bad genes but ultra-healthy lifestyles but also encourage bad behavior by individuals with good genes since they would now be able to pursue unhealthy life styles.

Finally I think the Professor's prejudice in favor of genetic determination is revealed by the last paragraph. To conclude that talent is "almost always" be traced genetics simply flies in the face of reason an experience. Perhaps Professor Posner's position in life is a result of his genetics, but I know that in my life hard work has been far more important than my genetic makeup.

(In the interest of full-disclosure I freely admit to having a vested interest in this subject as I have really bad genes.)

[anon]

"GINA bans only discrimination on the basis of genetic information that has not yet manifested itself in observable characteristics or behaviors but that is likely to in the future. There is no sense in this distinction."

Well, there is political sense in the distinction. You suppose that everyone thinks insurers should be able to discriminate on the basis of risk, but that is clearly incorrect. (Witness, for example, calls for universal health care.)

Many people believe insurers should not be able to discriminate on the basis of risk. For these people, GINA represents a compromise between that which they hope for, and that which they can accomplish politically.

[Ed]

It's true that genetic testing allows more accurate insurance measurements; there's not much in dispute there. But what is the real purpose of insurance? Consider what would happen if there were perfect information about future health. People would be categorized according to the exact cost of their future medical expenses. Maybe people in a similar class would structure payment so that people who had health problems in the present would agree to pay for someone else in the future in exchange for financial diffusion, but more likely, insurance wouldn't exist. Health insurance relies on healthy people. In order to prevent abuse, we'll draw the brightline at voluntary action and acknowledge that it's an inefficient system, but one that prevents larger disutility to society. While this may not be the legislative purpose of preventing genetic discrimination, it's a satisfactory impact.

[Jeroen]

What about the following argument. People who seek health insurance don’t know which risk group they belong too, especially on the genetic level (as opposed to e.g., age or how many burgers they eat). They have no private information about their genetic makeup. Forcing them to find out seems to defeat the purpose for seeking insurance in the first place. Ex ante we are happy to trade the risk of “overpaying” in insurance premiums in a good state of the world against the benefit of “underpaying” in a bad state of the world. It’s not as if there is much moral hazard in the choice of your genes. Maybe we would want to agree ex ante not to find out before we get coverage. (Ex post we could then find out in order to maximize precaution and efficient care.)

On a related note: doesn’t your argument lead straight to universal health care? The day we have a full health risk profile before getting insurance, the point of insurance (pooling risk) sort of vanishes. The only way to recapture the benefits of risk pooling, assuming we want them, is to mandate pooling as of the day of birth.

I admit this is all very simplistic, but it seems your analysis is too.

[jimbino]

As a person who chooses to go "bare" in all situations, I applaud the recent availability of genetic testing, for it will enable folks who test free of risk to cancel all insurance as I have done, which will have the effect of raising rates for all the rest, some of whom will then have greater incentive to get tested and drop out as well, leading to destruction of the stupid regime of insurance.

The situation Posner describes is akin to the silly policy of prohibiting employers from asking a young woman about pregnancy or her "family plans." The law does not prevent a young, single woman who has no "family plans" from offering the prospective employer that information or even presenting him a doctor's certificate that states that she has been sterilized. In a sensible world, no company would be offering health insurance to begin with, of course, and no sensible young man or sterilized young woman would work for a company that offered health insurance, since they are all required by law to cover pregnancy.

[km]

The author assumes that those who are at higher risk of something or other can't do anything about it, meaning they will cost insurance companies more. Ultimately, the ability to predict ahead of time that someone is at risk of something or other should mean you might avoid or at least lessen negative outcomes that would otherwise occur, thereby lowering medical costs for that person in the future. That should ultimately save health insurance companies money, not cost them more. But for that to happen, people have to feel confident enough that they will get the tests or participate in research that can make the necessary connections in the first place.

[jimbino]

KM:

Quite apart from your quaint phrasing of "predict ahead of time," I don't think you grasp the insurance problem. Of course, genetic testing, in the absence of the insurance religion, would be a great idea, akin to knowledge, information and education. Such testing would save or improve a vast number of lives.

But genetic testing would spell the death knell of insurance, since the idea of insurance is to hide the ball, deceive the healthy and insured, and to rob from the healthy and dumb in order to subsidize the life of the sick but smart.

Once the healthy insured become able, through genetic testin, to attach a number to their stupidity, insurance will suffer a welcome and rapid demise.

It is interesting that, in the absence of universal healthcare, the healthy, young and smart will gain from genetic testing, while the insurance industry and the sick will lose. The easy solution to this problem for the insurance companies, who now see the genetic writing on the wall, is to back universal healthcare. Once the union of government, socialists, health insurers and healthcare providers starts backing universal healthcare, our fate is sealed, except for those of us who have the foresight to maintain our freedom by simple renunciation of our Amerikan citizenship.

[Sajid Zaidi]

I agree that insurance companies should be able to consider all risk factors when pricing their policies. Otherwise, we are unfairly forcing healthy people to subsidize unhealthy people, rather than fulfilling the purpose of insurance which is to spread the risk among a group of people who are equally likely to suffer some catastrophe. It makes sense for me and 9 friends to get together and pledge a certain amount each in order to cover any catastrophes that may befall one of us as long as we all have the same health risk profile. But if one of our group is significantly more unhealthy, this is an unfair situation and they should have to pay more.

The simple solution is that if we as a society value the health of others, then we should be subsidizing the health care costs of those who are unfortunate enough to have genetic defects. The entire society should bear the cost of their health care, not just those unlucky enough to subscribe to the same health insurance company.

[Vineeth]

I agree with Professor Posner that insurance companies should be able to evaluate all available risk factors, but when it comes to health insurance the potential for the emergence of a disease or disorder is not enough. I think when the professor states that there is no distinction between presenting observable characteristics and the likely emergence of observable characteristics, he is wrong, especially when that likelihood is determined using the existing technology to evaluate genes.

I may be misunderstanding the issue but when a disease presents itself as an observable trait whether it affects the person's outward appearance or the person's organs or blood, the person effectively has the disease. They possess the characteristics of the disease and are subject to all of the adverse effects. But when someone has the genetic code to suggest that the emergence of that disease is "likely" that person does not have the disease. The idea is that, however unlikely, something can happen that would prevent this person from being subject to harm by the disease. I think this distinction is real and important. If someone is not subject to anything that would affect their health, they should be considered as healthy as anyone else.

More importantly however, the technology has not developed enough to understand how, when and why diseases present themselves from genetic code. Even if a person is predisposed to a certain disease we simply do not know what other aspects of the genome counteract or suppress that predisposition. I think until we have a complete understanding of how these processes work, we cannot consider genetic predisposition real risk. In fact, until we consider these genetic markers as a signal of the inevitability of the disease, we have to ban the use of that information.

The Director of the National Human Genome Research Institute said, “there are no perfect specimens at the DNA level; each one of us carries numerous gene variants that increase our risk of developing one disease or another.” I think at this point, our understanding of the genome is too muddled to have it be reliably used by insurance companies to determine to whom they may need to pay out claims.

[Chicago 2L]

Prof Posner,

Your example of life insurance is misleading: GINA does not pertain to life insurance or long-term care insurance coverage; rather, it only prohibits the use of genetic information in making determinations of health insurance. Health insurance tends to be heavily pooled via employers, and is thus doesn't allow the kind of opting out (and screwing up of the risk-sharing) that your examples use.

[Canada life insurance]

I don't believe some law or GINA can "secure" affected people. Insurance companies will just search for some other ways - they will have to. I am selling life insurance and usually you get some bonus, when you prove your good health status. It can be the same with health insurance - you will not be "punished" for bad test, you will be "rewarded" for good one. But the final result will be the same...
Lorn

[Eric Rasmusen]

Something rarely noted is that individuals often know they are high risk even without genetic testing, but insurance companies need an easy way to find out. Individuals know lots of things about themselves-- that they were sickly as children, or that their family is generally unhealthy-- that can't be well quantified. Thus, genetic tests may end up just equalizing information.

I think a big problem is that ordinary people don't understand insurance well. They think of it as a way to help unfortunate people, not as insurance. They do not make the connection between giving more money to unfortunate people and having higher premiums. They think that it is people with a certainty of accident or death who are the ones who most need and who most should be given insurance.

[drpatrick]

These regulations are a good idea, but there still seems to be some holes in the new laws. For instance, health information is not the only thing you can get from your DNA:

http://dnatestingguides.com/2008/05/what-is-dna-testing/

For example with the wholesale collection of DNA the government, or any company that so wished, could compile a DNA fingerprint database or trace your ancestry.
It's not clear how the new regulations would apply to this. Perhaps the best way to stop companies/governments abusing genetic information, is to not only protect the information, but not let them have it in the first place.