A tough call

You seriously need to spend more then a second to know which one is more distrubing? It is easily the 23andMe thing.

I'm still forming my opinion. I'm the one who came down on Navigenics for the "medicine that's not medicine," but I also criticized DTC genomics for a lack of clinical utility.

BRCA testing and others has known clinical utility. That much, I'm pleased, because I wanted better clinical utility, and I got it. What greatly worries me is the clear breach in the chain of trust.

This isn't gray area anymore, this is "Crime and Punishment" Raskolnikov medicine. 23andMe believes that the rules do not apply to them because they are Great. Well, that theory didn't work well for Raskolnikov, but he was a big pussy nerd, so I'll have to think more about this...

What happens when the law lets these guys slide? Does that give a free license to some other less funded company to provide DTC Huntington's testing? What about DTC Cancer diagnoses? What if some company just started making up diagnoses akin to Bernie Madoff never buying stock but taking participants money? Hell, what if they don't even do the sequencing, but instead report back bogus results, with even more bogus answers?

This BRCA thing is pretty crystal clear. To let them slide would create a very slippery slope......even if they are "Big Pussy Nerds"

-Steve
www.thegenesherpa.blogspot.com

Frankly I don't want to be required to get some Dr Dilbert's approval before I can get genomic information about myself, and I don't care that much whether generally responsible DTC firms are Utterly Without Stain in their efforts to do what they do. Mild hypocrisy beats bureaucracy.

> Does that give a free license to some other less funded company to provide DTC Huntington's testing?

If some company provides mistaken huntingtin sequences, the government should zap em. If they provide correct huntingtin seqs, it should be nobody's business IMO. Product matters, not process. Yes, determining who has crossed the line and who hasn't is something of a subjective/gestalt-ish human judgment call. A "government of laws, not of men" has gone a little too far in this country (only a little, I emphasize). "Even the finest teaching is not the tao itself" - even the finest law is not justice itself. We all know the responsible DTC firms are responsible.

Steve:
"Does that give a free license to some other less funded company to provide DTC Huntington's testing?"

Effectively, yes, particularly across borders, though funding is irrelevant.

"What if some company just started making up diagnoses akin to Bernie Madoff never buying stock but taking participants money?"

That's unlikely, because few people actively do what they think is wrong because they know it's wrong. Like in science, you're very lucky if somebody or something is "wrong" because that's at least a single bit of data. Instead, what you usually get is nonsense in the guise of "some greater cause." The net effect is more noise, which is in many ways worse, because noise is harder to filter.

"even if they are "Big Pussy Nerds" "

No, I said that Raskolnikov was a pussy and that I wasn't convinced that 23andMe would share his fate which was that his guilt made him not "Great" and thus he still subject to the laws of society and its consequences.

BRCA tests are NOT medical testing.

First, they don't tell you if you have the disease (they are not diagnostic).

Second, you don't know if it's a "medical grade" test (anti-contamination procedures, sample handing procedures, etc.)

In my city, many bars have "drunk-o-meters" by the exit doors. They report your blood alcohol level for $2. They are not "police tests" or "medical intoxicants tests". They are simply useful tool.

And to Dr. Steve: can you get it in your head that some people just don't want or need your protection? Those that do are free to consult you. Those that don't can go their own way. Why should you have to right to "protect me" against my will?

For the record, I'm not arguing that BRCA tests should have to be ordered through a doctor, or that 23andMe is necessarily doing anything unethical (as opposed to unwise) in offering DTC BRCA testing. The main reason I find this disturbing is because I fear describing obviously clinically relevant tests as "non-medical" will rapidly bring regulators down on the industry in a knee-jerk, excessive response. In other words I'm disturbed for very different reasons than Steve is.

Oh, and N/A: dude, that's the second time you've taken a post literally when it was obviously not intended that way. Were you deprived of humour as a child, or what?

"obviously clinically relevant tests"

My big issue is this: in a few short years, DTC companies will be able to offer full genome scans for under $500. These will reveal that I am immune/resistant/susceptible/destined to hundreds of health conditions.

I see the Coriell and other supposedly "medically ethical" attitudes ("We only reveal to you what we want") as horrifying.

I WANT to know if I have a "bad gene" EVEN IF THERE's NOTHING I CAN DO ABOUT IT. I don't want to have to get "counseling" before having each major test, or have some doctor judge for me that I "don't need" or "aren't ready for" or "couldn't do anything with the results of" this or that test.

Some people will want hand holding. I encourage them to go through their doctors.

Others are fine on their own, or will only consult their doctors with specific issues. We have a right to avoid being forcibly "protected".

Of course the alternative is to regulate these services to death. That will be one more industry which will pack up and leave the USA, taking jobs and money away.

I am puzzled by guys like Murphy, who can't see that they are acting in their own self-interest while they pretend to be looking after my interests.

23andMe are quite obviously not making any diagnosis, nor prescribing any treatments, so what they are doing is decidedly non-clinical. Of course there are clinical applications for the same tests, but that's not justifications for doctors having better access to my genome than I do.

So Murphy's private parts are surely disturbing. 23andMe's level-headed rationalism? Not so much.

Vince

Vincent,

While you are at it, tell 23andMe to test your SNPs for AD, PD, *prostate cancer (or breast cancer), age-related maculardegeneration, and ALS. Have a good day.

*If you really are a male and not using some fake name

Which of the following is more "disturbing"?

1.) BRCA Cancer Mutations via 23andMe
https://www.23andme.com/health/brca/
2.) The Home Access® HIV-1 Test System
http://www.homeaccess.com/HIV_Test.asp

Both don't require a prescription, which is understandable since there is no medical reason to ever deny an individual access to either system under any circumstance. However, the HIV-1 test is "FDA Approved" whereas the BRCA test has "not been cleared or approved by the FDA." After much discussion with Dr. Murphy, I guess that this is the distinction that he is trying to highlight. I can understand this distinction, but I doubt that 23andMe can afford to seek FDA approval for every possible genetic-related discovery it documents. My question is, why can't there be a low-cost, non-FDA approved service that refers individuals to physicians for follow-up? Such an organization would be similar to a public health department or pharmacy that offers blood pressure readings and refers individuals with high blood pressure to physicians. Or, perhaps there should be a new regulatory regime setup for the 23andMe's of this world. Under such a regime, 23andMe-like organizations could be FDA approved as following a set of "best practices" (i.e. certified as a "private" health department) which would then exempt them from having to seek approval for each individual "test" that they conduct. I think that this may be an acceptable middle ground, especially since programs like prometheus exist and these programs are essentially resistant to all regulation.

While you are at it, tell 23andMe to test your SNPs for AD, PD, *prostate cancer (or breast cancer), age-related macular degeneration, and ALS. Have a good day.

Of course N/A* knows that 23andMe are not testing me for those conditions at all. Rather, they are simply genotyping me for SNPs associated with those conditions. N/A may NOT know the difference, but that should not be my problem.

Vince

*Assuming he really is N/A and is not using some fake name

I wonder what exactly people like Murphy object to.

Is it my knowing that certain genotypes at rs1447295 are associated with an increase risk of prostrate cancer?

Is it my knowing which genotype I have for rs1447295?

Or is it merely that I can find out both of those things without helping put a doctor's kid in braces?

Vince

I will leave it up to Steve to defend himself as I don't speak for him.

When you test(s) a SNP(s) at any location that has been associated with a health condition, YOU ARE DOING A MEDICAL TEST. I would not have a problem with these direct to consumer genetic testing companies if they were regulated, etc.

Sure you can get a test and only get a print out of the sequence. But that is garbage information, pure garbage. Trust me one that! I pay my bills by reading genomic sequences (for clinical trials). When I'm writing A/A, A/t on a piece of paper it has no meaning besides saying a heterozygous allele at this location. I have no way of translating the sequence into anything meaningful until I do a patient history, family history, medical examinations, physical exam, etc, etc, etc.

I have no way of translating the sequence into anything meaningful until I do a patient history, family history, medical examinations, physical exam, etc, etc, etc.

All of those things you do are useful diagnostic tools. No one wants to prevent you from doing them.

The point you are making, though, is that the genotype is not a diagnosis of condition unless and until other factors are considered. That's my position also.

If you consider family history in making a diagnosis, should that preclude me from knowing whether my parents or grandparents had any health problems?

If three of my grandparents died from heart failure, then my risk of heart disease is surely higher than average. How is knowing my genotype any different from knowing my family?

Vince

> When you test(s) a SNP(s) at any location that has been associated with a health condition, YOU ARE DOING A MEDICAL TEST.

What if I happen to have a geneological or other non-medical wissenschaftlich interest in that same SNP?

Look, I'll just be blunt almost to the point of crudity. S Murphy mistakenly thought some lab or firm was touting a sequencing protocol that yielded a 5% error rate in their final sequence - which, it is blazingly obvious, cannot be true. (It was of course 5% error per read in a protocol using scores of reads, with a final error of a sliver of a percent.) He may be on the whole a good person and a benefit to society; probabilistically he's likely to be a better person than me (I'm not saying that pro forma); but I am smarter. And I'd like to take full advantage of one of my few robust strengths. But you want to fine, threaten, harass, and arrest those who would help me sequence my genome unless I pay Murphy a bunch of money I definitely don't have. And you can't really quantify within reasonable bounds of uncertainty the risk to society that free genome sequencing will create, which is your justification for your views. The reality is that it creates very little risk.

I'm all for people knowing their genetic information. I just don't want consumers to get the sequences and not have a physical done, not have a family history tree made, and to see a doctor. The problem is many fold: doctors don't know how to review and interpret the genetic information (in general), many genetic conditions that don't result in hospitalization early in life don't usually appear until disease onset has already set in, etc.

What we need is disease prevention. This is why we DO need people getting genetic testing done. I just feel, like several others may feel, that sequencing a newborn genome is the most beneficial as a parent can help observe for any symptoms of a health condition that the child may be at an increased risk for. For instance, most cases of autism exhibit during the second year of life and an educated parent can see the symptoms early on and can seek medical care.

Knowing your family history information gives you more accurate insight into your potential future health than just a sequence alone. When you combine both, and have clinic tests done, you have the most predictive risk for a potential disease you can have.

Genotype: You just know a sequence.

Family history: Wow, I didn't know mom had two miscarriages...didn't know great grandma had breast cancer and ovarian cancer...didn't know dad and great great grandpa had colon cancer. When you combine this type of information with a genetic sequence you are getting good information. This is what DNA Direct does correctly.

All I want 23andMe (and the like) to do is do IRBs, contract with medical doctors that can read the sequence, and get approved as genetic testing company. The states where a person can order a genetic test without a doctor the consumer can purchase whatever test they want and then do the clinical thing.

Trust me, getting doctors back out into the community to visit homes, interpret genetic sequences with the combination of a family history tree can help lower disease burden on our society.

This talk about double testing is dumb. Double testing increases cost, not decreasing costs. There are only so many dollars around to order the materials to do the sequencing.

As I mentioned before, I don't speak for Dr. Murphy. He is a doctor, I'm not (I'm just a little pre-med who works a job in a genetics area).

Insulin, don't go around saying you are smarter than others. Do me a favor and go on some clinical rounds and see just how smart you really are.

I am in favor of free genomic sequences, but not the way that one group in the New Jersey area is doing it because of the limitations they have in location.

I'm in favor of having newborns having their genome sequenced for free. The problem comes with the costs of sequencing these new babies. Just about all States are having budget problems every year. So we have to figure out a way to do this.

You all should read up on the Wisconsin Genomics Initiative. This is how you do it correctly: http://www.marshfieldclinic.org/patients/?page=genomics

You all should read up on the Wisconsin Genomics Initiative.

A fine initiative. IF I wanted something that integrated genomic, clinical, and environmental data and recommended a course of treatment or preventative care, that is exactly the kind of program I'd use. Assuming I lived the two years it will take to accomplish anything.

It just so happens, though, that is NOT what I want.

To further clarify my thoughts on this matter, I do think that 23andMe might be going a step too far when they say things like this on their website:

"Breast/Ovarian Cancer is one of the diseases that 23andMe analyzes. Our service includes the following information: Whether or not you are a carrier for Breast/Ovarian Cancer..."

That is a very strong statement. A lot stronger than saying something like:

"We looked at [SNP1], [SNP2], and [SNP3] in conjunction with papers [PMIDx], [PMIDy], and [PMIDz]. Overall, individuals with your genotypic composition at these markers appear to have a greater risk (quantify) of developing [Disease] according to [Algorithm]. Consult with your physician for more details."

The second statement is more consistent with what they started out doing. This kind of statement should probably not be regulated since it is relatively transparent, backward-looking, and talks about population risk, not individual risk. Obviously, it might not be very "user-friendly", but it makes 23andMe similar to Yahoo Finance - a site just supplying information, not making predictions. The first statement, though, seems very deterministic and patient, not population, directed. This would be like Yahoo Finance stating that "[stock x] is bad" or "[stock x] will go down" or even "sell [stock x]". In the financial world, those statements fall under regulatory authority and I think that similar "hard" statements should fall under the purview of the FDA in the world of genetics.

Contrary to what most people believe, Myriad's BRACAnalysis test isn't approved by the FDA either. It's a laboratory-developed test, or homebrew, which means that as long as their lab is CLIA-certified they can offer the test. 23andMe doesn't have a lab in-house, but the data is generated at LabCorp, which is CLIA-certified, so what's the big difference?

Quote: "so what's the big difference?:

- TOS violation just to start.

@Anonymous

Well, to the best of my knowledge, Myriad is not selling BRACAnalysis on a direct-to-consumer (DTC) basis (even if they are advertising on a DTC basis). It seems to me that a test sold DTC should require FDA approval if it makes personal/patient specific claims like OTC HIV-1 tests. This is the "big gray" area, moving from population specific claims to personal/patient specific claims. For instance, suppose a person tells a website like the New York Times that she is an African-American and it reports back that:

"African-American women face the highest risk for death from heart disease, and their rate of heart attacks is increasing."
http://health.nytimes.com/health/guides/disease/heart-attack/risk-facto…

This is not the practice of medicine. The person is simply identifying herself as a member of a class and then getting results about what has happened, in aggregate, to members that were sampled from her class in comparison to members sampled from other classes. It does not mean that she is at high risk for a heart attack. That determination would depend, at the very least, on a "patient history, family history, medical examinations, [and] physical exam." I believe that 23andMe needs to make sure that they only identify a person as belonging to a class based on genotype and then that they only report on scientific studies that compare members sampled from that genotypic class to members sampled from other relevant genotypic classes. If 23andMe wants to "get personal" and say things like "you are a carrier for Breast/Ovarian Cancer" or that "you are at risk for breast cancer" then I agree with Dr. Murphy; they either need to go the BRACAnalysis route and team up with physicians or they need to go the Home Access® HIV-1 Test System route and get FDA regulatory approval.

All the proponents of heavy regulation ignore one thing: no one is forcing you to be a customer of 23andme.

If you want physicals, medical advice and all the rest you can go to your doctor.

But many of us are perfectly happy getting interesting though non-medical information at a low cost. By what right do you seek to impose your views on us? And if DTC companies are forced to relocate outside the USA but continue to offer their services by mail, what will you have accomplished?

"All the proponents of heavy regulation ignore one thing: no one is forcing you to be a customer of 23andme."

Nobody is forced to invest in a Ponzi Scheme, but that does not mean that Ponzi Schemes are or should be legal. I think that the charm of 23andMe is that it promises not to offer any hand-holding or paternalism when it comes to you accessing your genome and the latest scientific research relevant to your genotypes. Furthermore, almost everyone agrees that if you want hand-holding, then that hand-holding should be done exclusively by physicians or genetic counselors working under physicians. Now, with their latest move into BRCA testing, 23andMe is making paternalistic statements and supplying a degree of inaccurate hand-holding which is misleading or, worse, fraudulent. There is a fine line between saying that "you have genotype A/G at some position and that, in the past, a sample of individuals carrying A/G at that position had an increased risk for breast cancer" versus saying "you have an increased risk for breast cancer." The truth of the first statement is determinable, free of hand-holding, and should be protected under the 1st Amendment even in a commercial setting. The truth of the second statement cannot be established with just genetic information and is probably fraudulent speech in a commercial setting. 23andMe needs to clean-up the language on their site ASAP or they do run the risk of provoking regulatory action.

A Ponzi scheme is outright fraud, where the consumer is kept in the dark. That's a ridiculous comparison.

In some cases if you have SNP X, then no matter what your family history is, you have increased risk of disease Y. This is a fact that can be checked, and is covered by free speech.

I don't need family history and a medical license to tell a 400 lbs woman "You have increased risk of heart attacks."

PS: When you talk of BRCA testing and 23andme, I don't think you understand the definition of "paternalism"
http://www.merriam-webster.com/dictionary/paternalism

"In some cases if you have SNP X, then no matter what your family history is, you have increased risk of disease Y."

There are very, very few cases like this and they all deal with inborn errors of metabolism, not complex diseases like breast cancer.

"I don't need family history and a medical license to tell a 400 lbs woman 'You have increased risk of heart attacks.'"

Actually, you do, especially if you are getting paid for giving out that advice. Perhaps both of her parents died of non-cardiac causes, she has low LDL and high HDL, passes a stress test, and has normal heart sounds. If you then tell her that she has an "increased risk of heart attacks" you are making an unsupportable statement. Furthermore, I would say that you are being paternalistic (i.e. acting clumsily like a father that somehow "knows" that her true risk for a heart attack is elevated above the general population even though you do not have a clue). All you know is that overweight individuals have, in the aggregate, an increased risk for heart attack and that this women is overweight - nothing more.

Hi Matthew,

I understand the distinction you're trying to draw between making verifiable statements of past research and genotype and suggesting that one knows a customer's risk based solely on genotype, without other knowledge.

But does this distinction apply to BRCA1? Even Steve Murphy seems pretty clear that having the well-known mutations makes you a deleterious carrier at vastly increased risk.

"Vincent,
It's not that you have a A at position x....it is that the A now makes you a deleterious BRCA carrier.....That is a very important distinction. I have no problem with the telling you about your As, Gs, Cs and Ts......just not what it means when it gives you an established medical disease/risk."

So in this case, what is the responsible message? And what happens if someone got genotyped/sequenced for ancestry purposes, and then the ancestry service knew that the customer was a BRCA carrier? How should that discovery be couched, or should it be reported at all? Or should ancestry-only services be regulated because of the potential for discovering medically relevant information?

If you're going to use your own definition for words, then there's nothing I can do for you. I go with Webster.

As for weight/vs/heart attack: Of course the more data you add the more refined the statement becomes. But even in the example you gave, she could have some epigenetic/environmental/unknown condition which puts her back into the high risk category.

By your argument therefore, we could never give anyone any information on risk. There could also be more data lurking in the shadows about to shift risk back and forth.

The fact is that people who have SNP X are often at greater/lesser risk of condition Y. The fact is also that the more we know about a person, the more we can refine the risk profile. So what? That analysis is a job for doctors, but shouldn't stop DTC companies from giving us access to our own genome.

An M.D. used to have to kill a rabbit before early pregnancy was confirmed.
Now a pregnancy test is a few bucks at the 7-11 convenience store

Same thing will happen to genetic tests.

The high and mighty need to get over it.

"I understand the distinction you're trying to draw between making verifiable statements of past research and genotype and suggesting that one knows a customer's risk based solely on genotype, without other knowledge."

Thank you for elegantly summarizing my ramblings!

"But does this distinction apply to BRCA1?"

Yes, I still think that there is a distinction.

"So in this case, what is the responsible message? And what happens if someone got genotyped/sequenced for ancestry purposes, and then the ancestry service knew that the customer was a BRCA carrier? How should that discovery be couched, or should it be reported at all?"

Let me be clear, I have always envisioned 23andMe as being more than just an ancestry service but less than a pure genetic counseling service. In its idealized form, I would view it as an automatic newspaper reporting on scientific publications relevant to my genotypes. Now, deleterious BRCA carrier testing is a complex subject and one that I am barely competent to talk about, but it should be pointed out that Myriad's BRACAnalysis [1] is more comprehensive than 23andMe's reporting [2]. More specifically, 23andMe's reporting is primarily based off of a single scientific paper [3] which has an external validity limited to Ashkenazi Jewish females. Thus, in the case of an ancestry-service that explicitly offers updates on scientific literature, I would propose the following rule:

IF (( I am an Ashkenazi Jewish female ) AND (( I have a 185delAG mutation in BRCA1 ) OR ( I have a 5382insC mutation in BRCA1 ) OR ( I have a 6174delT mutation in BRCA2 ))) THEN {I should be alerted to the existence of the aforementioned scientific paper [3] and provided with a summary of its chief findings.}

This of course does not apply to an ancestry-only service where the customer acknowledges that there will be no medically relevant information explicitly provided to him/her even though such information may be implicitly derivable.

[1] http://www.myriadresourceguide.com/pdfs/Sample-Result-BRCA-No-Mutation…
[2] https://www.23andme.com/you/journal/brca/overview/
[3] http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=Search&db=pubmed&term=9145…