[jacobolus]

I am sorry for your horrible situation.

I don’t really understand what your implication is. Are you wishing you had not done the genetic tests? Suggesting you should have aborted the 20-week fetus based on unusual skull appearance? Suggesting that the meningitis was a result of some malpractice? Mad at medical professionals who are now “dropping hints” without saying something more substantive?

The problem with these tests highlighted (confusingly) by the NY Times article is that they test for extremely rare conditions but have a high enough false positive rate (by my napkin math, on order 0.05% = 1/2000 false positive rate) that they end up dramatically over-diagnosing these conditions.

That doesn’t necessarily mean the people designing the tests were lacking in “sound science” or were behaving maliciously. Their tests just need to be almost perfect to not overdiagnose rare abnormalities, and perfection in medical tests is a very high bar. Arguably we shouldn’t do tests for such rare conditions unless the false positive rate can be dropped by another order of magnitude or more, and if we do such tests communication to those who test positive should be much clearer. On the other hand, since genetic abnormalities are so life altering, even a 10% or 20% chance of an accurate positive might be grounds to abort a very early fetus.

It sounds like your specific problem was that the genetic tests did not return a positive result, even though it seems plausible that your child had some genetic abnormality. Not picking up every possible genetic condition is a different problem from the false-positive problem highlighted by the article.

[savanaly]

I've always wondered-- is the error between tests correlated? Why does it not suffice to gather all the positive outcomes and test them again, then again and again until you have sufficient power depending on the rarity of the disease in the population? I know that for covid tests for example the test outcome errors aren't independent, which is why even taking several tests you can't be confident you're negative even if you tested negative. But I figured that was a special case and it was because their error comes from variance in viral load or something. But would that apply to these genetic marker tests?

[tagoregrtst]

I think of it like this:

Unlike a false negative where there isnt enough to trigger the test, false positive implies that something triggered a positive.

You have something, just not what the test thinks you have. For example, men with testicular cancer, apparently, show up as pregnant in pregnancy tests. Clearly thats a false positive for pregnancy.