Re:Ask HN: My child food allergy from prick test. Can't find...
by brudgers on Wednesday September 21, 2022 @12:13PM
(#32927309)
The science is statistics.
Fifty percent of positive results are false positives.
5% of negatives are false negatives.
Most results are negative.
Lets assume 99% for this example but for most uncommon allergies it is probably lower.
So out of 100 tests, the results are expected to be accurate 99 * 0.95 + 1.0 * 0.50.
That's essentially 95% of the time.
In addition, the harm is in not treating an undiagnosed allergy. So there's no direct harm from treating a false positive. The harm is in not treating an a false negative.
The other thing about the prick test is that it is fast and easy.
Fifty percent of positive results are false positives.
5% of negatives are false negatives.
Most results are negative.
Lets assume 99% for this example but for most uncommon allergies it is probably lower.
So out of 100 tests, the results are expected to be accurate 99 * 0.95 + 1.0 * 0.50.
That's essentially 95% of the time.
In addition, the harm is in not treating an undiagnosed allergy. So there's no direct harm from treating a false positive. The harm is in not treating an a false negative.
The other thing about the prick test is that it is fast and easy.
The alternative is controlled diet.
It's way more work.
Good luck.