1. Genetic disorder. Tay–Sachs, etc. This stuff is priceless, because in the past, you had to pay thousands of dollars to get these same tests tone. These are accurate and very worthwhile ... but the majority of people will come back with all negative results, which is actually a good thing.
2. Your "increased risk" to certain diseases, like heart disease, etc. IMHO, these aren't worth much. A substantial portion of the risk for these types of diseases is related to lifestyle, so take it with a grain of salt.
3. Fun info, like telling you your eye color, where your ancestors came from, etc.
> why are their tests not being subsidized by insurance companies.
Insurance companies really only want #1. And guess what? That's info that I really don't want them to have.
This is just to say that that sword has two edges.
EDIT: Made it abundantly clear that this is not my own personal fatalistic perspective.
Just because you're more inclined to do something, doesn't mean you're likely to do something. See the other comment about a 50% increased likelihood moving someone from 1% to 1.5% risk of succumbing to alcoholism.
That said, a lot of people may make the same mistake and succumb to alcoholism erroneously based on the tests.
Let us assume that the absolute risk is non-negligible, and that the predisposition is also non-negligible. Even then, the way that an individual responds to knowledge of genetic risk is idiosyncratic. Some will fight the predisposition even harder, others will succumb even quicker, still others won't know or care.
Because we are talking about the psychology of the individual, and because people do not understand statistics, much less the difference between absolute and relative risk, the fact that the absolute risk for virtually every disease is very small will not matter when an individual considers the issue psychologically.
What does that mean?
If the probability of alcoholism is the general population is 1% ... well, that means yours is 1.5% ... or still ridiculously small.
Obviously those are just numbers I'm using to illustrate a point, but you see where I'm going with this.
In the case of cardiovascular disease specifically, one might not want to do that.
You still have to pay thousands of dollars if you need the results for actual medical purposes. 23andme only tests for a few of the common SNPs that can cause each disease, but the testing isn't as comprehensive as what you'd get from a doctor.
4. You're (nearly) immune to an infectious disease. It's good to know when a friend is puking their guts out that I can help take care of them since I'm one of the lucky 20% of white people with wonky fucosyltransferase that won't act as a receptor for the virus.
And knowing (2) lets channel potential anxiety about salt consumption into anxiety about fat consumption, letting me stay healthier on a given anxiety budget.
[1]: http://www.nytimes.com/2012/11/25/business/seeking-cheaper-i...
"Sorry sir your infant child is not eligible for that heart operation due to the genetic marker we discovered in your profile."
The premise of insurance is that you are insuring against the unknown. If you know that something is the matter with you, then it's no longer insurance. It's getting someone to pay for a treatment that you know you will need. Whether other people should be paying for that is another discussion.
I do think a nontrivial amount of the mess in the health care system is the confusion between the two. People want to create health care plans, but they are trying to make it out of insurance companies, which is going to be problematic at best. Further, people want health care plans but don't want to think about it from an actuarial perspective... because they don't like the answers that come out. That doesn't prevent the answer from coming out, but it does mean we end up trying to build glorious systems that seal themselves away from reality from the get go, and that also can only be problematic at best, total failures at worst.
And as of 2014, regardless of the nature of the evidence, insurance companies will no longer be allowed to exclude customers based on predisposition to illness or preexisting conditions.
In fact, HIPAA might have already forbidden that. The law wasn't clear but it's entirely moot now.
Because of money.
"`(A) IN GENERAL- For purposes of this section, a group health plan, and a health insurance issuer offering group health insurance coverage in connection with a group health plan, may not adjust premium or contribution amounts for the group covered under such plan on the basis of genetic information."
Don't think that's legal.
> What if employers require their employees to do this much like drug / background checks?
Also don't think that's legal.
We already have employers requesting "voluntary" disclosure of health information (e.g., cholesterol check) as a condition of continuing group insurance coverage. Since it's illegal to require this information, they make it voluntary - but anyone who opts-out also opts-out of a hefty premium "discount". In reality, the "discounted" rate is closer to what employers would expect to pay for comparable group plans, while the non-discounted premium is high enough that 99% of employees line right up to have their blood drawn by the visiting nurse (conveniently stationed in the conference room down the hall).
However, medical knowledge isn't at a high enough level for these accurate genetic results to translate into truly meaningful medical results except for a handful of conditions. What does a 10% increased risk of getting diabetes mean to the average person? In practice, it means nothing, because you would act the same with or without this risk.
Second, 23andMe is already walking the fine line of not needing FDA approval. If insurance companies were to pay for it, there's a good chance it would require this. There are several conditions that can be self diagnosed through 23andMe that require genetic tests that cost more than $99 through your doctor, but each of those tests has gone through a medical approval process.