[anewvillager]

> In the 23andme case there was a worry that people might commit suicide or make massive life changes, if they discovered they had a potential terminal illness. Here, the 2x2 matrix of positive vs negative and true vs false, seems most dangerous on the false-positive side for people who believe it grants them freedom to do whatever they like. False negatives are dangerous for those who then feel certain they don’t have the disease currently and thus expose others unwittingly (but this is already likely!).

So you suggest it's better to not test people?

[jmchuster]

Isn't that the suggestion? That if there are groups of individuals who you know are going to make wildly irresponsible choices because they don't have the ability to interpret such information, you would prefer not to give it to them?

[greglindahl]

It's entirely reasonable to assume that tests with high error rates cause problems for everyone, not just "groups of individuals who you know are going to make ..."

For example, a PhD biologist friend of mine agonized for months before she was willing to look at her personal 23-and-me results, after the FDA certified them as accurate enough to show to consumers.

It's a real problem for everyone.

[boulos]

No, I was just calling out the "you know, the outcomes just aren't that bad here". As I ended on, I think they'll come to the conclusion relatively quickly that there is minimal risk to the public from having such diagnoses, even if they're misinterpreted.