3 ms·

I'd address your first point with a simple question: Suppose 23andme did no data analysis/interpretation and simply returned the raw data, would that increase

I'd address your first point with a simple question: Suppose 23andme did no data analysis/interpretation and simply returned the raw data, would that increase or decrease the scope of appropriate FDA involvement? To your second point, the site has a few "locked" areas that are described as containing potentially upsetting or confusing information, and the user is advised to consult his/her physician before looking at the results. There is also downloadable information available to physicians.

dragonwriter 13y ago
> I'd address your first point with a simple question: Suppose 23andme did no data analysis/interpretation and simply returned the raw data, would that increase or decrease the scope of appropriate FDA involvement? Reducing the scope of the product would inherently reduce the scope of the FDA involvement, not that that is in any way relevant to any point I made. > To your second point, the site has a few "locked" areas that are described as containing potentially upsetting or confusing information, and the user is advised to consult his/her physician before looking at the results. There is also downloadable information available to physicians. I don't see how that's relevant to anything I said. Particularly, its not relevant to either the second point in the first paragraph (which concerns 23andMe abandoning the regulatory clearance process) or the point of the second paragraph (which concerns 23andMe failing to provide studies supporting their claims about the accuracy of the reports).
- grandalf 13y ago
  It's possible 23andme has been sloppy with its compliance practices. If you take the raw data and download various utilities that match SNPs with the literature, the result is far more scary and misleading -- and it reveals just how thoughtful and reserved 23andme's approach is. 23andme provides ample links to evidence for every single result. The confidence metric is a function of the study size and reproducibility for the studies pertaining to each and every result. Most have very low confidence. IMHO the only legitimate beef is with the underlying scientific assumption behind 23andme, which is that statistical genetics is both useful and accurate. I consider it more of a form of entertainment than a likely way of gathering actionable information.
  - dragonwriter 13y ago
    > It's possible 23andme has been sloppy with its compliance practices. Then no one should be surprised that they are in trouble with a regulatory agency responsible for holding participants in the market they are in to compliance standards.
    - grandalf 13y ago
      Actually I think it's quite surprising. Why would a well-funded company pursuing a business model that clearly falls into a regulatory gray area gamble with its future by simply not complying?
      - dragonwriter 13y ago
        Seems to be not-too-rare practice with startups -- don't comply and hope the rules adapt to them rather than vice versa (either by legislation or litigation.) AirBnB, Uber/Lyft, 23andMe previously (with then-current California regulation requiring tests of the type they provide to be physician-ordered) have all done this with varying degrees of success.

dragonwriter 13y ago

> I'd address your first point with a simple question: Suppose 23andme did no data analysis/interpretation and simply returned the raw data, would that increase or decrease the scope of appropriate FDA involvement? Reducing the scope of the product would inherently reduce the scope of the FDA involvement, not that that is in any way relevant to any point I made. > To your second point, the site has a few "locked" areas that are described as containing potentially upsetting or confusing information, and the user is advised to consult his/her physician before looking at the results. There is also downloadable information available to physicians. I don't see how that's relevant to anything I said. Particularly, its not relevant to either the second point in the first paragraph (which concerns 23andMe abandoning the regulatory clearance process) or the point of the second paragraph (which concerns 23andMe failing to provide studies supporting their claims about the accuracy of the reports).

grandalf 13y ago
It's possible 23andme has been sloppy with its compliance practices. If you take the raw data and download various utilities that match SNPs with the literature, the result is far more scary and misleading -- and it reveals just how thoughtful and reserved 23andme's approach is. 23andme provides ample links to evidence for every single result. The confidence metric is a function of the study size and reproducibility for the studies pertaining to each and every result. Most have very low confidence. IMHO the only legitimate beef is with the underlying scientific assumption behind 23andme, which is that statistical genetics is both useful and accurate. I consider it more of a form of entertainment than a likely way of gathering actionable information.
- dragonwriter 13y ago
  > It's possible 23andme has been sloppy with its compliance practices. Then no one should be surprised that they are in trouble with a regulatory agency responsible for holding participants in the market they are in to compliance standards.
  - grandalf 13y ago
    Actually I think it's quite surprising. Why would a well-funded company pursuing a business model that clearly falls into a regulatory gray area gamble with its future by simply not complying?
    - dragonwriter 13y ago
      Seems to be not-too-rare practice with startups -- don't comply and hope the rules adapt to them rather than vice versa (either by legislation or litigation.) AirBnB, Uber/Lyft, 23andMe previously (with then-current California regulation requiring tests of the type they provide to be physician-ordered) have all done this with varying degrees of success.