FDA tells 23andMe genomic test to stop marketing (US)

Quantified Selfers and the D3H (Digital Health Hypester Horde) are in a swivet. This past Friday, FDA slammed the door shut on the 23andMe Personal Genome Service (PGS) saliva test. This past summer, the company broadly marketed to US consumers, including a TV campaign [Charles Lowe, TTA 7 Aug]. The FDA cease-and-desist letter cites that 23andMe never provided requested data on their July and September 510(k) filings, which are now ‘considered withdrawn’, and cites that “after these many interactions with 23andMe, we still do not have any assurance that the firm has analytically or clinically validated the PGS for its intended uses, which have expanded from the uses that the firm identified in its submissions.” The danger is that people will make medical decisions based on the testing information and that the results produced may be faulty. It appears from FierceHealthcare that the kit has actually been marketed for five years. According to MedCityNews, it is backed by Google Ventures (the CEO/co-founder is the estranged wife of Google head Sergey Brin), New Enterprise Associates, MPM Capital and the Moscow billionaire Yuri Milner. A private citizen is petitioning the White House to overrule the FDA (as if that extra-legal move would be possible, but who knows with the influence of the Googlesphere?) and states that the agency ‘grossly overstates the risks’ (also MedCityNews). As of 2 Dec there are 3,306 signatures of the 100,000 needed; one suspects this administration has bigger slices of uncooked turkey on its plate such as Obamacare and a kind-of-achieved 30 Nov deadline on Healthcare.gov, which is now clearly seen as just one problem.

The 23andMe website is still fully up and still selling kits.

Editor Donna sorts through the noise for possible reasons why: While worthy, genetic testing such as 23andMe can have personal consequences beyond consumer information and to be useful, should be reliable and also taken by the end user seriously. It’s apparent that 23andMe initially played with FDA the right way: they filed as required, met extensively with FDA, then broke it off earlier this year. In the FDA’s account, 23andMe was given a lot of chances to prove their stuff. Moreover, 23andMe is apparently still selling in defiance of the FDA letter.

Is it that the clinical validation wasn’t there? Was some of the genetic testing flawed? Is it that they didn’t have enough people of diverse backgrounds roped into trials to get the clinical proof, and decided not to go the ‘investigational device’ route for financial reasons?

Is it that the investors said ‘damn the torpedoes, full speed ahead’?

This Editor’s opinion after consideration (and written before reading Forbes, below): The broad consumer marketing effort this past summer seems to point, in this scenario, as a last-ditch effort to bypass FDA, take the risk and leverage awareness for publicity. Ms Anne Wojcicki has been pounding the conference circuit, and there is that breathless sense of conspiracy expressed in this article in Fast Company (which has featured her prominently–it’s the FDA in cahoots with the insurers!), in various LinkedIn groups and of course a breathless and overclaiming ‘petition’. I note that 23andMe’s Terms of Service are carefully written to put all responsibility onto the back of the user. Notably way down the list is the disclaimer “23andMe Services are for research, informational, and educational use only. We do not provide medical advice. The Genetic Information provided by 23andMe is for research, informational, and educational use only. ” Yet everything else on the website presents their testing’s medical consequences as potentially life-saving or critical lifestyle modifying information, not ‘fun’ and educational.

You can disclaim till the cows come home, but once you get past the cognoscenti and enter the mass market of millions, if you 1) truly care about your customers and 2) want to protect your company, you need to haul back on the promises, assume that your users just trust you more and work through worst case responses. Or else put aside a budget line for legal defense and chew up a lot of your staff’s time in court.

As a marketer, I understand several things that appear to have eluded 23andMe. Underpromise and overfulfill. Get what your product does (your implied warrant of service) rock solid (23andMe is not at this point) and backed up by studies. Structure your claims as if a trial lawyer will come after you, because they come with the territory. Until that point, don’t full-bore market DTC. Write your materials at a (US) 6th grade level. And live with the fact that you are in FDA (and CE, Health Canada) land.

Finally, if you are going to pick a fight, go for the KO, not points!

Also worth reading is this balanced but damning take in Forbes by Matthew Herper, 23andStupid: Is 23andMe Self-Destructing? A closer-in look from a journalist who’s been following the company, he susses that their real business model was not just the liberation of health information, but building a monetizable database of millions sharing genetics, health history and more for research purposes. Downright Google-esque, in fact, but that was a different world and territory. He discounts their strategy of libertarians plus D3H–this Editor never discounts the power of Google and its money in the political world.

Online petitioning site Change.org has nearly 3,000 4,225 signatures (2 Dec) expressing the libertarian POV that personal genomics kits should not be banned–in other words, not regulated–because people have the right to make decisions based their genomic information. Nothing about whether the FDA actually has a legitimate role in ensuring the accuracy of the data.

Updated 2 December

Categories: Latest News, Opinion, and Soapbox.


    • Donna Cusano

      Thanks David. This was a miss that spelled a problem to be fixed before a wider rollout. However, their response seemed OK: 23andMe knew what happened from the outsourcing and handled it fairly well. Also that their ‘beta testing’ used leads from Ancestry.com and WOM. Red flags, especially with outsourced testing, have to abound. Which is another way to consider their goal of 1 million records from the reported present level of 500,000–how much of this would be known as flawed in a vastly larger database? What quality procedures would be put into place to ensure this?

      Having debated whether this would be of interest to our readers and within our scope, Charles’ initial article ruled it in, with the news in the first paragraph (NY time). In thinking about it overnight, the reasons why a company would actually ‘step on Superman’s cape’ began to nag at me, thus my thoughts above. The cutoff, after a long dance with the FDA, says ‘calculated risk with an ace in the hole’.

      Perhaps I have drunk a bit at the Well of Conspiracy, but 23andMe may have a secret weapon in their political connections. Or they may be going for broke, as well financed as they are, to break the FDA structure on genomics testing. FDA is badly stressed just dealing with mobile health, and now this.

      My concern is that once in a kit, the darn thing better work as advertised–not that the average citizen has no right to it.The libertarian view that FDA should concern itself only with urging interpretation by medical staff of these tests does have appeal (see Reason’s blog article), but it is way too caveat emptor with something emotionally loaded, as the examples you provided proved to be. 23andMe should also build another database–genetic counselors.

      In any case, we’ll be sorting through a fair dose of overheated D3H rhetoric from The Usual Suspects before this is all through.