After FDA Objections, 23andMe Won't Offer Health Information

sciencehabit writes "The company 23andMe will no longer provide health information to people who purchase its DNA testing kit, it announced last night.The change was 'to comply with the U.S. Food and Drug Administration's directive to discontinue new consumer access during our regulatory review process,' the statement said. While current customers will still have access to a 23andMe online database noting the health issues associated with their particular DNA, the company will not update that information, and customers who purchased its Personal Genome Service (PGS) on or after 22 November will receive only information about their ancestry and their raw genetic data without interpretation." It would be great to see a secondary market in this kind of analysis emerge.

Comments

There already is a secondary market already

[PapayaSF]

From snpedia.com:

The easiest way to make a report is to visit promethease.com. This takes about 10 minutes and costs $5. [...] Promethease is a tool to build a personal DNA report based on SNPedia and a person's genotype (DNA) data. Customers of DNA testing services (23andMe, FamilyTreeDNA, Ancestry.com, Complete Genomics, ...) can use it to learn more about themselves completely independently of whichever company produced their data.

good riddance

[ljw1004]

Good.

My family all had our genomes sequenced by 23andMe. The only area we have expertise in is Alzheimers, and (1) their Alzheimers explanations were misleading, (2) they made it REALLY hard to learn the raw data about what they found in our genomes, i.e. which SNPs they tested and what they found: instead they only boiled it down into a useless "you have 20% chance of getting Alzheimers" which was scientifically incorrect, lacked confidence levels, lacked context.

I would love to get the raw data from their results, and I'd love to have someone better than them provide the tools to analyze & understand it.

Re:good riddance

[PCM2]

Perhaps this is why the FDA put the kabosh on it

The FDA was very clear about why they stopped it. It wasn't necessarily that the information was misleading, but that it would lead patients to make decisions about their own care without necessarily consulting a doctor, which the FDA thinks is not a good idea -- and I totally see their point, frankly.

For example, one of the things that 23andMe can tell you is how well you might respond to one drug versus another, because of your specific genetic makeup. If you take that advice and change the dosage of your medication or switch to a different medication without discussing the issue with your doctor, you could cause yourself serious harm.

On the far end of the scale, "false positives" for some diseases could be catastrophic -- say, if a woman was erroneously told she had a high chance of contracting a certain type of breast cancer and decided to have a double mastectomy, like Angelina Jolie had done.

23andMe claimed that all it was doing was giving people information. But really, the way the information was structured and presented to the customer made it clear that it was designed to be diagnostically relevant and that they should use it to make decisions about how to proceed with health care. Any service that performs that function clearly falls under the jurisdiction of the FDA, IMHO.

Re:good riddance

[PCM2]

How could that possibly be within any legitimate government's domain? Using the same rational they could shut down wikipedia or rxlist.

They sure would shut down Wikipedia or RXList if those services allowed you to make an appointment to consult them for medical advice. Even campus health nurses have to be licensed.

What Wikipedia offers now is pretty much the same thing as reading information out of a book. You can't stop people from doing that, and there's no law against it.

What 23andMe does is market a product that you use to extract unique information about your own body, which is then presented to you in the form of suggestions about what health measures you should take -- in other words, medical advice. Very different.

Re:good riddance

[E 99]

What 23andMe does is market a product that you use to extract unique information about your own body, which is then presented to you in the form of suggestions about what health measures you should take -- in other words, medical advice. Very different.

Whoa, what? They have never been in the business of medical advice! What they did is to say, "you have genetic marker X which according to studies A and B are indicative of a 20% increased susceptibility to disease Y or and 50% increased likelihood to have an adverse reaction to drug Z." That is not medical advice! That is mere information, filtered by your genetic markers.

Re:good riddance

[tlambert]

The FDA was very clear about why they stopped it. It wasn't necessarily that the information was misleading, but that it would lead patients to make decisions about their own care without necessarily consulting a doctor, which the FDA thinks is not a good idea -- and I totally see their point, frankly.

The FDA made them stop because doctors dislike being cut out of the loop, and insurance companies like being cut out of the loop even less than the doctors, and they would prefer to have you get the data through a disclosure mechanism which gives your insurance company better actuarial information. "Having a Dr. explain the information to the patients one on one" is just a place to hang that hat.

For example, one of the things that 23andMe can tell you is how well you might respond to one drug versus another, because of your specific genetic makeup. If you take that advice and change the dosage of your medication or switch to a different medication without discussing the issue with your doctor, you could cause yourself serious harm.

Yeah, in case you wondered, people can not self prescribe non-over the counter medications. So that excuse doesn't fly, either, since your doctor will be involved in writing the script for the new medication, and your insurance company will be paying for it, and like mine did, probably try to give you a cheaper generic version of a similar drug in place of the one your doctor actually wrote the script for, and then called it "equivalent". I've had that pulled on me, and been given "generic allergy medication" containing a cornstarch binder in place of the other one - when corn products were why I taking the damn stuff in the first place.

Re:good riddance

[Mashdar]

Does anyone else find it upsetting that the local CVS is packed with whole aisles of homeopathic "remedies" with no proven efficacy (or worse, disproven), but some company can't tell you what your genes might mean? Apparently the FDA is about protecting what goes in your eyes than what goes in you mouth?

Re:smart

[Albanach]

There is a secondary market. Most large hospitals have genetic counselors who are trained and qualified to interpret and present the results of genetic testing.

Re:smart

[pepty]

Those are the results of individual FDA approved genetic tests.

A "secondary market" for running thousands of tests at once would run into the same problem that 23andMe did: if you are going to sell diagnostic services in the USA then you will need to get FDA approval. The options are to only provide raw data and let someone else generate the report for free (see Promethease, mentioned below) or move the whole company someplace where law enforcement won't bother it. I could see 23andMe spinning off an independent foundation that would generate free reports, thus allowing their core business (building up a database of peoples DNA and personal/family medical histories that they can rent out for medical research free from the normal regulatory hurdles) to proceed unhindered. They would just have to be very careful to make certain there was no linkage between the for profit and the free companies: the non-profit would have to generate reports from data of any source, not just SNPs from 23andMe, it couldn't share any board members or employees with 23andMe, Google, etc.

Next legal challenge for 23andMe: doing medical research on people (gathering their DNA and medical histories, analyzing, allowing 3rd parties to analyze, etc) without getting IRB approvals first.

Too bad it's taken down...

[sldunn]

As long as people understand that the results are only based on research/papers, 23andme is really awesome. It's best to discuss things with your doctor, and get things double checked, rather than to treat 23andme as some alternate to proper healthcare and checkups.

I mean, really, who the hell would spit into a tube, pay $100 bucks and start a potentially harmful treatment regimen without seeing a doctor?

Speaking from experience, 23andme did identify that one of her genes leaves her susceptible to having bad side effects of one of the medications she was taking (and she was suffering from this side effect). Taking the 23andme health report to her doctor let her move onto an alternate treatment, which is working *much* better.

I hope that a revisited health report/traits thing comes back soon. Or maybe put it behind a test wall, and make sure people to understand exactly what they are getting.

The science is open..

[xtal]

Just need to link the science (published) and the genotype. It's all open.

Next up is banning people from sequencing their own genomes without a MD.

The real story here is who's the loser - it's not you; your DNA is your DNA, and the sequences are there or they aren't. The insurance industry are the ones who are actually worried about these tests - all of a sudden you have data they don't, and they can't apply their actuarial models anymore. Hilarity ensures.

May we all live in interesting times.

From a 2010 Cell paper

[sat1308]

One company, for example, offers 166 tests in one of its testing packages where approximately 60% of the tests (99) are categorized as âpreliminary researchâ(TM) because the genetic-association data have not yet been replicated (www.23andme.com/health/all/). These tests are given 1, 2, or 3 stars based on the size of the study that supports the genetic association for which they test. Information for each of these tests cites references for the original ïnding of the genetic association, including the journal where it was published and the study size. It also provides the number of attempted replications and the number of contrary studies that have been published. Although transparent, examination of the scientiïc evidence provided for many of the genetic associations in this category raises the question of whether these tests should even be included in a genetic-testing package. Two of the ïve genetic tests with 1-star status (those for âavoidance of errorâ(TM) and âobsessive compulsive disorderâ(TM)) are based on single studies with fewer than 100 participants (https://www.23andme.com/you/health/). In both cases the variants map to the D2 dopamine receptor, a gene that has repeatedly been associated with human behavioral traits and attracted newspaper headlines, only to have the associations refuted in later studies [8]. Eight of the 37 (22%) available 2-star-rated genetic associations (originating from a single study with less than 750 participants) have a âcontrary studyâ(TM) indicated. Two different 3-star tests, one for Lou Gehrigâ(TM)s Disease (ALS) and another for obesity, utilize variants that have been positively associated with disease in one or two studies, respectively. However, both these variants have failed replication in four additional studies (https://www.23andme.com/you/health/). Although, the company boasts of its 'systematic vetting processâ(TM) used to determine which research ïndings to include in its genetic-testing package, a number of highly questionable tests continue to be offered to consumers.

http://www.ncbi.nlm.nih.gov/pubmed/20828856

Thord party analysis

[n6kuy]

> It would be great to see a secondary market
> in this kind of analysis emerge.

There are already companies (livewello.com, for instance) that will take your 23andme raw data and analyze it for certain traits and risk factors.

Re:23andMe is just plain bogus

[Anonymous Coward]

What's your definition of "useful conclusion"? Because I'm pretty sure it's different than what 23andMe was peddling and what all of us satisfied customers expected. Most 23andMe reports are equivocal on their face, in a good way. Some, like markers for Parkinson's or Alzheimer's, are also pretty clear. They tell you in plain English that the lack of a marker does not mean you're free from risk. If a marker is present, they give you a risk factor. But in no way did they diagnose you with Parkinson's.

Everybody thinks that they're smarter than most other people, and people like you just assume that most other people can't comprehend this stuff. Well, 1) we can and 2) it doesn't matter because 23andMe reports are of a different type entirely than what a professional genetic counselor would possess.

Show me one person who was duped by 23andMe? Even my own mother, who is quite naive about science-related topics, consumed her reports with a grain of salt, and without needing any warnings from me. Everybody understood that 23andMe is to professional genetic counseling as Walgreens is to The Mayo Clinic. In other words, professional and trustworthy in so far as a commercial vendor can be, but not in the same league as a real healthcare provider.

In an age where people obsess over stuff like HFCS and juice fasts, 23andMe cut through all the crap with hard data and reports which, despite all the inherent limitations, were damn interesting, if not in the rare case positively life changing. People understood the limitations, even the science-deficient anti-HFCS and organics crowd.

What the FDA did was pounce on some lofty marketing language. Big whip. The FDA just wants to shutdown the commoditization of genetic counseling, period. And it's ultimately going to cost all of us a ton of money.

Re:smart

[JWSmythe]

That quote isn't quite complete. They've only discontinued new user access to health related tests. They're still providing ancestry information.

The majority of the really useful information is genetic. The health information was secondary. Really, the whole health thing was vague at best, and it didn't take a rocket scientist to figure it out.

For example, gallstones. My risk is 6.2%, where avg risk is 7.0%. Not bad, I have a lower chance.

I had to have my gallbladder removed a year ago, because I had two golf ball size stones. Yup, anything greater than a 0.0% chance means there is a chance. Lucky me.

Anyways, here's the full email they just sent out to 23andMe customers today.

Dear 23andMe Customers,

I'm writing to update you on our conversation with the U.S. Food and Drug Administration and how it impacts you.

If you are a customer whose kit was purchased before November 22, 2013, your 23andMe experience will not change. You will be able to access both ancestry and health-related information as you always have.

23andMe has complied with the FDA's directive and stopped offering new consumers access to health-related genetic results while the company moves forward with the agency's regulatory review processes. Be sure to refer to our 23andMe blog for updates.

We stand behind the data we have generated for customers. Our lab partner adheres to strict quality standards that are part of the Clinical Laboratory Improvement Amendments of 1988 - known as CLIA. These are the same standards used in the majority of other health and disease-related tests.

You are among the first people in the world to ever get access to their genomes. You are genetic pioneers. Thank you for your ongoing support and we look forward to continuing to serve you.

Anne Wojcicki
Co-founder and CEO, 23andMe