Does 23 and Me make more of its income from selling individuals ancestry and health records or from selling data to insurance companies, research entities, and the like?
Considering that they don’t sell your DNA information in any identifiable form (they DO sell aggregate data that’s stripped of any identifying information to researchers), I’d say it makes more off of individuals paying for their tests.
Even anonymized data could be of significant value to insurance companies, though. If, for instance, it can be found that innocuous, easy-to-treat disease ABC is strongly correlated with expensive disease XYZ, then an insurance company would want to jack up the rates or drop anyone who seeks treatment for ABC, even though ABC itself isn’t a big deal.
And with enough data, you can de-anonymize it: For instance, 23 and Me has data on how many relatives their customers have, and who’s related to whom how. Couple that with other databases that also include that data, but with names attached, and you might be able to figure out that this person who has one sister (with two daughters and a son) and no brothers, with a mother who has 7 sisters and 3 brothers (with various counts of descendants each), and a father with one brother and two sisters (one of which sisters has two daughters), is the same person as the person in that other database with that same set of relatives.
They don’t need 23andme data to do that. Simple health care records, which the insurance companies already have, and need to have, will be enough. “Look at this, 30% of the people that received treatment for a stubbed toe in their 20s, developed toe cancer in their 50s.”
Now, I do think calling it anonymous genetic data is pretty silly. Here are several million SNPs on a subject. Nobody knows the subject’s name, but no other person in the world is going to have the exact same combination of millions of SNPs, but the subject’s relatives are going to have similar combinations. I don’t think the idea of separating a name from the subject’s data is silly, I just think it should be called what it is, which is de-identification, not anonymization.
The genetic data can easily be matched to the subject or the subject’s relatives in different databases, and that is in fact something I do occasionally. Researchers often combine data from many genetic studies, and we will go through and eliminate relatives. We only know that two subjects are relatives because their genetic data is similar, not because we know who they are. This is done because having subjects too closely related can bias the result of some analyses.
There are also laws prohibiting health insurance discrimination based on genetic data. So, if, for example, expecting parents get a genetic screening for a variety of conditions, as they often do, their health insurance cannot be raised based on those results. This prohibition does not apply to other forms of insurance. So, if there is a huge breakthrough, and we discover a set of genes which strongly predispose carriers to alcoholism, their car insurance (and life insurance) companies could raise their rates.
Finally, in answer to the OP, I don’t think health insurance companies in the US are spending too much buying 23andme (or ancestry.com, etc.) databases to raise rates. They may be getting access to the data for basic research purposes. If 23andme are making money from selling the data it is to pharmaceutical companies who want access to huge subject pools for drug discovery purposes.
And in healthcare, it’s worth noting that these problems are only of any concern in a system based around private for-profit insurance where the concept of a “pre-existing condition” is relevant. Under such a system, genetic data are just one aspect of the overall economic incentive to deny healthcare to high-risk individuals.
AKA, killing sick children and grandmothers to improve your bottom line.
As an aside, I’ve always wondered how people in the US who work for health insurance companies live with themselves, given the fact that their companies entire existence and continued profitability literally depends on killing sick children and grandmothers, and other folk at their sickest and most in need of medical treatment. It’s a notch or three worse than working for a for-profit prison company or being a professional executioner or hitman, and yet they blend in seamlessly in polite society for some reason.
Well, supposedly the idea is to provide a service, and make a profit while doing it. In order to do that all the insurance company has to do is charge higher total premiums than the total payout. They can do that and still cover all legitimate claims. It requires lots of data and statistics to guess what the premiums now should be to cover future payouts. It also requires having lots of customers with small payouts, to cover those who have large payouts. There can be done in a way that treats customers compassionately. There is no need to deny legitimate coverage.
However, if maximizing profit at any cost becomes the managing principle, then denying coverage is an easy way to decrease payouts, and increase profit. Defining “legitimate” coverage in such a way that it not only denies lots of fraudulent claims, but also a few real ones? Hey, less payout.
In summary, for profit insurance is a stupid model for routine and on-going health care. (It might work for catastrophic care.)
As long as the government does not provide an insurance-like service (single payer system), then private insurance is needed. If they did not exist, then costly procedures (major surgery) would effectively be impossible for 90% of people.
As others have said, I don’t think insurance companies are the place they are going with this.
More likely that profiteers would be pharmaceutical companies and medical diagnostic companies. I see two main uses
-
By collecting all these samples they can test new mapping and analysis techniques to improve the accuracy of their system, and identify rare genomic variations. they can then market their improved system to medical researchers or companies who do genetic screening and have a whole lot of data to back up their accuracy in a number of different situations.
-
If you agree to participate they ask follow-up medical questions to those who try their system. By having such a large database, they may be able to identify new markers for different diseases that will allow for better diagnosis and possibly even novel treatments, which they can sell to pharmaceutical companies.
Although there is a profit motive behind, I think that this is on balance a good thing, as it overall promotes the advancement of science and medicine.
The first part does not in any way demonstrate the second part.
(all numbers made up) If they were paid $100 by 10,000 people to for their private data, and then sold aggregate data for 10,000 people that had been stripped of (non-inherent) identifying information to researchers for $10 million, then they would be making more from the aggregate data sale portion of their business model than from the individual private sales portion.
The proportions of how much they are paid for which parts of their business model do not, in themselves, mean any wrongdoing or unethical behavior, and I do think it’s an interesting question as stated.
So, I don’t think the OP’s GQ has been addressed, everyone seems to be answering a question about how the data is handled that was not asked.
Yeah, I think I read the OP as “Do DNA companies sell people’s data to insurance companies?” rather than “Do DNA companies make MORE from selling their client’s data than they do from the clients?”
Moderator Warning
Textual Innuendo, you’ve been around long enough to know that political and professional jabs are not permitted in General Questions. This is severe enough that I am going to make it an official warning. Do not do this again.
Colibri
General Questions Moderator