Hacker News Reader

Comment by SCAQTony | Hacker News Reader

Never once did they attack the technology but rather the potential for error or the consequences if the public "can't handle the truth."

The closest they come to calling her a "quack" is when they state "...We still do not have any assurance that the firm has analytically or clinically validated the PGS for its intended uses..."

The bigger question is why is the FDA having a seizure over this? Could it be the potential for added treatment and preventative healthcare measures that insurers vis a vis the Affordable Care Act and not looking forward to paying for? (Not a rhetorical question, just asking)

View on HN

I think the line of reasoning makes sense: people will make medical decisions based on this information, but 23andme has never given us any reason to (or not to) believe their tests are accurate. Without someone reviewing their process, their error rate is unknown. That can be dangerous. This is similar to the problem with many alternative medicine claims. 'Alternative medicine' may be very helpful to an individual, but it's important not to overstate its effectiveness or reliability.

how would you define 'error rate'? Do you mean "lab handling error rate"? Or "information presenting" error rate? The way 23andMe works, is they run your DNA and they show "percentages" based on already done studies. Then they throw in caveats (like, if you're not of the correct race relative to what the study was done in). So what would you suggest? That the service track all of their users over a long-term and then see whether the percentage of their users that get condition X matches the percentage they predicted? And then do some sort of statistical analysis on a possibly moving number (because 23andMe updates their information if academic science improves) with a population that might turn out to have differing biases...? And then prove this information over what, 100000 SNPs on the chip?

What you are basically arguing is that 23andMe should not be allowed to offer its service.

Ultimately both error rates matter, and are measurable. It's not necessary that these error rates are zero, it's necessary that they are measured, controlled, and acknowledged, which is what 23andMe is not doing.

Even the FDA isn't arguing that 23andMe should not offer it's service, just that it shouldn't market it. And really, given that they don't seem to be following the basic quality assurance practices of the industry, they should not be marketing at all.

> Even the FDA isn't arguing that 23andMe should not offer it's service, just that it shouldn't market it.

They aren't even really arguing that either; they are more arguing that if they are going to market it, then they need to provide studies supporting the accuracy and take other steps to get it approved to be marketed for the particular uses it is being marketed.

They are saying in the Warning Letter that 23andMe must stop marketing the product because 23andMe has abandoned the approval process (and apparently was not preventing 23andMe from marketing the product as long as they were working through the approval process), not because the product is inherently unmarketable.

False. Read the letter carefully. They're specifically concerned with false positives and false negatives - these are both very real possibilities - one is documented up thread.

> The bigger question is why is the FDA having a seizure over this?

Um, they aren't having a seizure. 23andMe was informed of the regulatory requirements in a letter in June 2010 [1], and (as discussed in the Warning Letter just issued) submitted required documentation to begin the approval process in July and September of 2012. The FDA sent letters identifying issues with the submitted documentation and requesting additional information in order for the approval process to move forward on September 13, 2012 and received no response from 23andMe.

The FDA set follow up letters in March and May of 2013 indicating that due to 23andMe's failure to address the issues raised in the FDA's earlier letter, the applications submitted in 2012 were deemed withdrawn -- which meant that 23andMe was not only not in compliance, but no longer considered to be actively working toward compliance -- but still took no substantive action, giving the company time to begin working toward compliance again.

Now, more than a year since raising issues with 23andMe's original submission, more than 3 years since notifying them that they needed to come into compliance, and 6 months after notifying 23andMe that their failure to respond to the issues raised with their initial submissions had caused those submissions to lapse, they've given 23andMe notice that they will face regulatory action if they don't either stop marketing the products or provide an acceptable plan within 2 weeks.

That's not exactly "having a seizure".

[1] http://www.fda.gov/downloads/MedicalDevices/ResourcesforYou/...

Reply on news.ycombinator.com