Im confused, technical questions for you

[jammienix]

Hello, i posted in the public forum which im sure you saw from my posts. I am curious about when the tests and people talk about the sensitivity and specificity of a test. From what i read if sensitivity is 99.7 % then means there is a fraction of people that are truly pos that are testing neg because of this or does this only have to do with the window period?

From my understanding the tests may still give a false negative even well beyond the window period regardless of producing antibodies or not, to be simple the test just misses. Please correct my understanding of this as for i have been reading different stuff on this and would like to get a direct answer to someone who knows about this, not that i 100% distrust Teak in the public forum yet i cant find your opinion on this particular question.

Thanks for reading this question.

[Edward W Hook, MD]

Yes, it menas the same thing.    There will be no further answers.  EWH

[smith125]

Does 12 week (84 days) or 3 months test mean the same? Or does one recommend a 13 week test (90 days).

[Edward W Hook, MD]

I've explained this as best I can.  For further information I am afraid you will need to talk to a statistician.  This will conclude this thread.  EWH

[jammienix]

sorry I'm not asking what ifs I just don't understand, at least I don't think I do.  The 99.9% is just because you can't say 100 when testing I'm guessing rhats what your getting at?
and I know 2 years is ridiculous I wasn't saying myself or others should I was saying out of example of someone testing wayyyyyyyy beyond window period that was all...

[Edward W Hook, MD]

It is not 99.5 % of persons, it is 99.9% (note the difference in number) of persons with HIV, a very small proportion of the total tested.   Further, the time frame for such recommendations is 8 weeks or, at the outside if your are (in my opinion, overly) conservative), 3 months.  The idea that it takes 2 years to have a reliable test is nothing short of unvarnished paranoia.  The tests are to be believed. If you cannot believe them retest to your hearts content.

Watch out for meteorites.  EWH

[jammienix]

I'm not worried about false neg due to lab tech screwing anything up, I'm curious as to the sensitivity of a test. If say someone has hiv and doesn't take his till 2 years after last exposure and gets neg because the test is only 99.5% so a percent of people will slip through regardless of how long they waited or if a test tech screwed up or not..is this correct in my understanding? If that's the case shouldn't people take 2 tests

[Edward W Hook, MD]

Welcome to our Forum. I have reviewed your interactions on the HIV Prevention Community site and agree with what has been said.  As a generalization for both you and other readers, you must realize that we VERY frequently get questions asking if different types of test are 100% accurate or if exposures or prevention measures are 100% effective (while the questions are not posed in terms of sensitivity or specificity, but that is what is being discussed).  The answer to that is that it is scientifically impossible to be 100% sure.  For a variety of complex mathematical reasons far too complex to go into here, all one can do with well conducted scientific studies is estimate probabilities (this is in part because, unless you test every possible person, you cannot be "sure" that you would not have gotten a different answer).  By definition, any estimate, cannot be 100% certain.  On the other hand, when Dr. Handsfield or I say that something is very close to no risk or of minimal risk, or use any other term to indicate a very small risk, that means "close to zero" in a world where zero cannot be attained.  This just as there is chance that you will be struck by a meteorite while reading this, I am very confident that that will not be the case.  Can I be 100% confident about this, no but I am sufficiently confident that I can tell you that this is not going to happen.

There are few tests which have even been as thoroughly evaluated as the performance of HIV tests.  When we evaluate the performance of any laboratory test there are many things that can go wrong.  The technician could do the test wrong, the participant from whom the specimen was obtained could not have told the truth in their historical information, the specimen could have been mislabeled, the materials in the test could nor performed as they are supposed to, or aliens could have interfered with the test.  None of these happen often enough to not have complete faith in the results of the tests for HIV when done as recommended.  When we provide figure on the Forum we are providing the best data we have available.  The development of antibodies to HIV does not occur at the same time in all people, this is the reason for the so-called window period.  A very few things can delay this process but not many and the chance of someone have a rare reason for delaying antibody development following an exposure to a partner who statistically was unlikely to have HIV and even if they did, would be unlikely to transmit infection is so low that we have great confidence in our recommendations and figure.  I hope you can as well.  If you cannot, please be careful to look out for meteorites as well.

I hope this helps.  EWH