Hacker News Reader

Comment by pippy | Hacker News Reader

After the FDA started pressuring 23andme in to not giving data relating to diseases to customers, I lost of respect for them in the genetic arena. Customers have a right to know what's in their bodies. At the moment 23andme is only cleared to give out data from a single disease, which is pathetic. Clearing the way for more genetic diseases will likely be a bureaucratic nightmare and would harm more people in the long run.

Given how conservative the FDA is in allowing companies to inform customers what's in their own bodies, it's unlikely companies will get any form of approval for modification.

View on HN

I'm not familiar with FDA that much (I live in EU) and I believe it isn't anywhere near perfect, but you want to be conservative when it comes to public health. Look at case of Thalidomide from 50s https://en.m.wikipedia.org/wiki/Thalidomide

  In the United States, pharmacologist Frances Oldham Kelsey M.D. 
  withstood pressure from the Richardson-Merrell company 
  and refused Food and Drug Administration (FDA) approval 
  to market thalidomide, saying further studies were needed.

Any example is pointless without considering opportunity costs. How many people were saved due to faster availability of drugs?

Then again, how many people would be lost if public confidence in drugs as being safe was lost? We can't get a sizable portion of people to even vaccinate their children anymore, and that was all sourced from a single fraudulent paper, which the author lost his license over[1].

1: https://en.wikipedia.org/wiki/Andrew_Wakefield

I imagine the reasoning is that give people incorrect information about a disease they may have could be extremely damaging. Having models that predict the accuracy of each diagnosis to within a few percent and presenting this percentage would be better (maybe requiring it be over a certain percentage), but the FDA is conservative enough I can imagine them requiring 99% accuracy or better on everything.

I don't know the details of the particular case of FDA and 23andMe (I remember only the one when FDA asked 23andMe to fix their marketing, or else), but there is a known problem with giving medically-relevant information to people - that armed with such information, fully predictably, they go batshit insane.

Most people are not properly equipped to handle such information. I see this as a failure of education, that could probably be easily fixed if e.g. high-school curriculum taught basic statistics and probability theory as applied to interpreting medical results. It's the simple things, like "just because you got a positive test on X doesn't mean you have X", "think of horses before zebras" and "if you read 'X causes cancer' or 'X increases risk of cancer by 1000%' in the news, take note if anyone is providing the base level - an increase of 0.00001% probability to 0.0001% means you can ignore the news and keep enjoying X", and "if commonly-used-in-food-X was really dangerous, you'd see people dying like flies around you; if you don't, it means it's mostly safe so don't panic".

While I agree with your assessment of how people act, I really hope that's not a reason that the FDA really uses to keep information from people[1]. It would be troubling to find out that the government had decided that I should not be able to freely access personal information about myself, given that it exists. Whether as a society we can deal with that information currently or not, there's no hope for us dealing with it in the future if we aren't given a chance.

1: I don't doubt it's some people's underlying motivation, I just hope it's not overt enough that it's valid reasoning in official reports.

Reply on news.ycombinator.com