Below find my ever-growing annotated collection of online responses to the FDA’s recent shot across the bow of 23andMe, the consumer genetics company. Until today, I was adding these links to the bottom of my own first reaction to the FDA’s stern letter. (I published my broader, more studied take, How 23andMe Broke the Rules: The F.D.A. Versus Personal Genetic Testing, on Nov 27 at The New Yorker.) I’m moving the ever-growing linkroll here, and arranging them in reverse chronological order (newest responses at top), to make them more accessible. Know a good one I missed? Add the link in the comments below, or email to email@example.com, with 23andMe in the subject.
If you’ve time for only five (after you’ve read my New Yorker overview, of course), try these:
- 23andStupid: Is 23andMe Self-Destructing?, by Matthew Herper. “This is not the way to deal with a powerful government regulator.” Herper’s been covering the FDA for 13 years. The depth shows.
- In FDA vs. 23andMe: How do we want genetic testing to be regulated?, Geneticist Michael Eisen with a bullet-point case for why we need direct-to-consumer genetic testing and how it might be regulated.
- In When 23andMe gives results that no one knows how to manage. Dr. Jen Gunter considers the dilemma a doctor faces when a patient gets a 23andMe result that suggests risk …no one knows what to do with.
- At Slate, Razib Khan argues that FDA letter to 23andme won’t mean anything in the long run. I think he’s right.
- At The DNA Exchange, Laura Hercher ponders how she and her fellow genetic counselors might use this disruption to find new ways to contextualize genetic information.
- The Sky is Falling for Personal Genomics! Oh, nevermind. It’s just a cease & desist letter from the FDA to 23andMe. [Dec 3, 2013]. Possibly the best account and commentary yet.
LINK COLLECTION STARTS HERE
Added Jan 16, 2014:
Regulation: The FDA is overcautious on consumer genomics. Well-argued worry by a Robert Green, a Harvard Medical School geneticist, and Nita Farahany, a Duke Law professor, that the FDA overreached with 23andme — and may also be considering regulating simpler health-risk services as well, such as phone apps that analyze your basic demographic, health, and family history information — the sort of stuff you could tell a doctor in an office.
In its recent guidance on mobile health applications, the FDA left open the possibility that it will regulate as medical devices information-based products such as questionnaires that evaluate the risk of a heart attack or the plethora of fitness trackers that help people to follow their weight, body temperature, heart rate, sleep patterns and more.
If that were the case, the FDA would essentially be making analysis of information a medical device. (“I wouldn’t eat that carbonara if I were you; could give you a heart attack.”) Green and Farahany may be overreaching themselves there, but it’s an important point to consider, and one that I’ve not yet seen FDA clarify.
My Risk-Benefit Ratio For Personal Genetics. Virginia Hughes, one of the sharpest followers of genetics, with a spirited second to Green and Farahany, with some interesting scenarios (such as a visit to a Romanian orphanage) mixed in. And do not miss Hughes’s 23 and You, a masterful look at how 23andme geneological information (which the company is still allowed to sell) upended one family’s view of itself.
New evidence shows the FDA was wrong to halt 23andMe testing. A good story on the Nature commentary, with extra data from another study as well.
Added Dec 5:
Added on Dec 4:
The Sky is Falling for Personal Genomics! Oh, nevermind. It’s just a cease & desist letter from the FDA to 23andMe. One of the most helpful things you can read, from Jennifer K. Wagner, aka @DNAlawyer.
23andMe’s Wojcicki: “We failed to communicate proactively” – Fortune Tech From the CEO’s first public appearance since the story broke. Thin but, well, all we’ve got.
A government ban on 23andMe’s genetic testing ignores reality | Rahul Rekhi | Comment is free | theguardian.com A doc makes a plea for patient access to genetic info. “While the FDA’s paternalistic policy on DTC genomics is admirable, it is also obsolete. The agency’s dictum, at its core, ignores the medical paradigm shift towards patient empowerment that consumer genomics and the internet is advancing each day.”
The Failed Promise of 23andMe A psychiatrist argues for tighter FDA oversight of outfits like 23andMe.
23andMe suspends marketing after failing to meet FDA requirements | Science | theguardian.com This was news on Dec 3.
Added on Dec 2:
23andMe’s problems just got a lot worse. GigaOm reports that a customer has filed a class-action suit seeking some $5 million. “According to the complaint, which seeks at least $5 million under various California state laws, 23andMe makes false and misleading claims about the tests’ ability to provide relevant genetic information about breast cancer, diabetes, lactose intolerance and various other conditions.” Longer story at GigaOm
In other news, geneticist Joe Pickrell, expanding on Michael Eisen’s very sharp post of last week about regulating 23andMe specifically, asks Should the FDA regulate the interpretation of traditional epidemiology?, by which he means risk assessments based on things like family and personal health history. As he notes, many sites offer online “risk calculators” of things like stroke or heart attack, where you enter things like your age, weight, and small bits of medical history and then get a readout of your risk. It took me just 15 seconds to learn I supposedly have a 6.3% chance of having a stroke in the next 10 years. It took me another 15 seconds, pretending I was a woman with health history like some people I know well, to be told by the National Cancer Institute that I have a 3.5% 5-year risk of getting breast cancer and a 22.4% lifetime risk. Maybe I should get a mastectomy? This is precisely the nightmare scenario that the FDA raised in its stern letter to 23andMe. The main practical difference is that 23andMe’s assessment is based on a few snippets of genetic information, while the Nattional Cancer Institute’s assessment is based on a few snippets of demographic and health and family history information. Yet one is regulated, because it uses a machine that parses spit, and the other is not. Why?
That’s the question Pickrell asks here. Give his post a go.
Links added Nov 30:
Mike The Mad Biologist’s Thoughts on the FDA Action Against 23andMe.com was one of the best and most thought-provoking posts I read when this story broke on Nov 26; I’m not sure how I managed to leave this link out of this linkroll this long. (A chip-reading error.) We don’t see eye-to-eye on this, but — because of that — he’s a good corrective at thinking 23andMe is innocent victim. This is both a clash of cultures and, more broadly, a society trying to figure out to use an emerging body of information wisely. In any case, read Mike; he’s always lively and fun and smart.
Meet the phenome-wide association. Carl Zimmer focuses his weekly New York Times column on Linking Genes to Diseases by Sifting Through Electronic Medical Records – NYTimes.com. In these new “phenome-wide association studies,” he writes, “scientists start with a gene variant and then search among thousands of conditions for a match.” A nice look at a new approach that might complement the sorts of genome-wide-association studies that have struggled to find links between genes and disease.
Here are three papers on the (highly) limited effect that learning genetic information has on people’s health-related behavior:
Apparently knowing one’s genome rarely changes behavior. BUT: 3 things: 1. No other form of information seems to reliably change people’s behavior; it usually takes a good solid heart-attack scare or the equivalent; 2) I still think that examining one’s own genome is a matchless way to come to understand genetics; 3) the burden of proof lies on those who want to control genetic information, not those who want to control it.
But you can’t say 23andMe results are worthless, because Chuck Cody used his to make a song. Seriously.
In a post fun, lively, and smart,,John Wilbanks finds the FDA’s Culture Is Mendelian Dominant Over 23andme’s Business Model. Which is worrisome.
Links added Nov 29:
The editorial director of Entrepreneur.com is not impressed with the FDA’s move.
Cruwys news: 23andMe and the FDA This useful round-up includes 23andMe’s national TV ad, which many feel may have helped provoke the FDA’s move. The ad clearly does make health-related claims about genetic risk findings, and it’s easy to see why this ad might alarm the FDA — and anger them if they’re feeling ignored by the company.
In a comment below my New Yorker story, geneticist Gholson Lyon explains why, in the anecdote described at the end of that article, he was not able to share with the patient’s family his identification of a gene that caused the death of an infant. (A couple of other commenters had said they found that situation implausible.) It has to do with whether a genetic test is done in a lab certified by a federal program known as the Clinical Laboratory Improvement Amendments, or CLIA. As Lyon explains, “[C]urrent law stipulates that genetic test results in America returned to participants must be obtained in CLIA-certified labs, and that much of the research enterprise is NOT performed under such settings. Furthermore, still, the vast majority of exome and whole genome sequencing is NOT performed under CLIA standards, including lack of very rigorous sample tracking and not being performed in any CLIA-certified labs. “
Two particularly sharp posts came today from American Science, a group blog about the history of sciene. First, Science, Regulation, and the Epistemology of Big Data offers that
Although I cannot say so with certainty, I have a strong suspicion that what the FDA is really objecting to here is the use of big data techniques in biomedicine. Traditionally, if a company wants to bring a certain pharmaceutical or medical device to market, it is expected to conduct extensive clinical trials. In so doing, there might not be an expectation that every aspect of the drug or device’s mechanism of action is fully known. However, there would at least be clinical or experimental data to support its safety and efficacy.
But these are precisely the kinds of data that big data will not produce.
In the other American Science post, What’s better than a holiday card? “Possibly the worst FDA letter of all time, tries to clarify what the FDA might be after here:
As FDA has made clear in the past through their correspondence with DTC firms, the worry with a medical device is that it be accurate and reliable. That is FDA’s rationale for post-premarket review. But back in 2010 Guitierrez, the head of FDA’s Office of In Vitro Diagnostics, did not want the agency to be viewed as paternalistic and said, “We really don’t have any issues with denying people information. We just want to make sure that the information they are given is correct.”1 So, now we are back to worrying about analytic validity (the reliability of the actual genetic test, meaning laboratory performance) and clinical validity (whether the genetic variant actually corresponds to the condition or trait).
In this post at Forbes, the author is disappointed that his 23andMe results don’t seem to line up with reality — that is, with his health and family history. I find this a great example of the dissonance between the probabilistic results of genetic effects and the common expectation that genes produce traits in straightforward fashion. The FDA worries that such expectations will lead people to get mastectomies they don’t need. In the comments section, some people are trying to enlighten the author.
Links added Nov 28:
23andMe revealed a condition it took my doctors six years to diagnose. This piece by Shaheen Pasha shows well the value of a test like 23andMe not as a solid diagnosis — which it isn’t — but as information about risk, to help direct diagnostic efforts.
When 23andMe gives results that no one knows how to manage. Dr. Jen Gunter takes a smart and nuanced look at the dilemma a doctor faces when a patient gets a 23andMe result that suggests risk … that you don’t quite know what to do with. Gunter reacts not by suggesting we ban such testing; rather, she accepts it as inevitable, and says “somehow we are going to have to figure out what to do with this kind of raw data.”
In a red letter day for consumer genomics, Dr. Stuart Hogarth, who was in on some of the early FDA-industry meetings on regulating personal genomics kits, ponders what the FDA letter seems to mean, what about the letter mystifies him, what may have gone amiss between 23andMe and the FDA.
Why The FDA Can’t Be Flexible With 23andMe, By Law. Very helpful explainer/contextualizer by @DavidKroll
In this interview at BusinessWeek, Alberto Gutierrez, the FDA official who wrote the warning letter to 23andMe, elaborates on why the FDA feels it needs to call the company to heel. His arguments seem mainly aimed at how scared patients can push their doctors to do unnecessary and sometimes harmful procedures. This reveals an interesting tension: He’s right about fear driving procedures, but that fear comes from many places and is based on many kinds of information, some far, far less solid than even the shakiest of the findings that 23andMe reports. In a world where companies can more readily help people gather information about themselves, where does the FDA draw a line on what constitutes a ‘medical device” that delivers medically relevant information? That’s one of several tough questions that needs a good, thoughtful answer to really solve this dilemma.
Links added Nov 27:
My own Nov 27 piece at The New Yorker, How 23andMe Broke the Rules: The F.D.A. Versus Personal Genetic Testing
In FDA vs. 23andMe: How do we want genetic testing to be regulated?, UC Berkeley geneticist and open-science advocate Michael Eisen lays out a concise, bullet-point case for ) why we need direct-to-consumer genetic testing and b) how it might be regulated. A key point that lies at the heart of the dust-up: “Genetic tests are simply not — at least not yet — medical devices in any meaningful sense of the word…. The FDA and companies like 23andme need to come up with standards for accurately and honestly describing the current state of knowledge for genotype-phenotype linkages and their application to individual genotypes.” Those two sentences hit the root of this thing spot on.
Why 23andMe is Not for Me — Yet Writer Ricki Lewis expresses well both some important caveats about genetic testing in general and her own reasons for not getting into it … yet.
23andMe Is Terrifying, But Not for the Reasons the FDA Thinks Charles Seife writes that while 23andMe is indeed using its growing database of genes and health conditions for useful medical research, the company also “reserves the right to use your personal information—including your genome—to inform you about events and to try to sell you products and services.” Seife worries that, while the company so far promises it won’t sell your genetic data to anyone else, there’s good reason to worry that could change.
Nathaniel Comfort adds some long-cycle perspective, including the history of hyping anything genetic: 23andMe, FDA, and the history of hype | Genotopia
23andme CEO Anne Wojcicki elaborates a wee bit with An Update On FDA’s Letter to 23andMe | The 23andMe Blog. Her message is a bit hard to square with the FDA letter saying the company had not communicated with the agency for 6 months. But perhaps the FDA has the company’s attention. “This is new territory for both for 23andMe and the FDA,” Wojcicki writes. “This makes the regulatory process with the FDA important because the work we are doing with the agency will help lay the groundwork for what other companies in this new industry do in the future. It will also provide important reassurance to the public that the process and science behind the service meet the rigorous standards required by those entrusted with the public’s safety.”
Going meta: Here’s a nice wrap-up of by Lindsey Alexander, of MedCity News, of 6 other posts on the FDA-23andMe dustup. They include John WIlbanks, with 23andme Gets A Nastygram For The Holidays and Why 23andMe Terrifies Health Insurance Companies, which may be a bit overwrought … or not, since, historically it has been easy to underestimate the greed or power of the health insurance industry.
Links posted Nov 26:
23andStupid: Is 23andMe Self-Destructing?, by Matthew Herper, who says, “This is not the way to deal with a powerful government regulator.” Herper allows that, by the FDA account, the agency gave 23andme plenty of chances but the company didn’t step up. I hope we’ll get full accounts that make clear whether that was the case, or whether the FDA itself is being less than forthright.
Body blow: How 23andMe brought down the FDA’s wrath | The Verge This provides some needed context on how conversations between the FDA and 23andme may have broken down.
At Slate, Razib Khan argues that FDA letter to 23andme won’t mean anything in the long run:
This isn’t the story of one firm. This is the story of government response to very important structural shifts occurring in the medical delivery system of the United States. The government could potentially bankrupt 23andMe, but taking a step back that would still be like the RIAA managing to take down Napster. The information is coming, and if there’s one thing that can overpower state planning it is consumer demand. Unless the US government wants to ban their citizens from receiving their own genetic data they’re just putting off the inevitable outsourcing of various interpretation services. Engagement would probably be the better long term bet, but I don’t see that happening. I think he’s essentially right.
Three years ago at WIRED, Daniel MacArthur expressed concerns that the FDA might make a restrictive move at some point in Did The FDA’s Jeffrey Shuren Mislead A Congressional Hearing?
A while back here at Neuron Culture, genetic counselor Laura Hercher makes The Case for Selective Paternalism in Genetic Testing.
Image: Fencing match, courtesy University of Wisconsin.