In article <37hu6i$clh at www.interramp.com>, <pp000165 at interramp.com> wrote:
>>Brian .... This is not meant to be a flame but do I understand you
>that if a donor to your blood bank comes back HIV positive on
>Elisa you don't believe it? Do you use the blood anyway?
>If you dont use the blood, then you dont notify the donor so
>they can go on donating blood which is going to be re-
>jected because of the off-chance the Elisa might be correct?
Eh? This is a very strange interpretation of what
I said. What I said was that there was a 98% probability
that a positive HIV ELISA is a false positive,
at least in my blood bank. Predictive value of
a positive test in a setting with low prior probability
is low. We certainly would not transfuse a unit
with a 1 in 50 chance of having HIV!
>Why dont you tell your donors they came back Elisa positive but it
>may not mean anything and recommend them see their doctors?
>Mainly because it would be malpractice to do
so. We do what all blood banks do - perform confirmatory
western blot assays to determine whether the
donor actually *is* infected with HIV. What is
done with those who *don't* confirm positive
(ie, the 98%) is a very complicated and contentious
area of medicine, and I don't want to go into that.
The point of my post is that ELISAs and other
screening tests are not suitable as the sole
means for determining such a crucial question
as whether a patient has HIV. For good reason.
Brian
