Flutterby™! : new HIV test 2004-03-26 17:48:14.747945+00

new HIV test

2004-03-26 17:48:14.747945+00 by Dan Lyke 3 comments

"Studies show the rapid oral test is more than 99 percent accurate." Hmph. I believe the HIV infection rate (in the US) is something less than 0.5%. So just saying "no, you haven't got it" regardless of the evidence would be a 99.5% accurate test.

They need to give the false-positive and false-negative rates before you can tell whether this is worth anything at all.

Another factor is the timing of the antibody response. You could catch the plague on Friday and still test positive on Monday, since not enough antibodies could have been produced to trigger the new test. I believe the current tests won't show until you have been infected for a few weeks. No idea as to the sensitivity of the new test.

Also, as somebody pointed out on NPR, preferred tests are weighted toward false-positive. Nobody who tests positive with the new test will be declared so until they have had other, more established (and time-consuming)studies.

I once met a doctor who worked with CDC who knew of a man in SF who had been tested a few hundred times, basicly getting a test at a free clinic every two days or so. With the new test, he could do it several times a day. Better safe than sorry.

#Comment Timing, etc... made: 2004-03-29 18:27:07.344784+00 by: baylink

One of the stories I saw covering this said 6 weeks. So no, still not a silver bullet.

More importantly, though, it seems to me that the "false-positive; commits suicide" problem, which requiring-everone-to-get-their-results-in-person-even-if-negative was supposed to avoid, is not addressed by this. It's sort of like one-touch-recording on VCR's, which replaces the supposedly so annoying play-record from audio cassette recorders... which was introduced *for a reason*, when people kept accidentally recording over stuff. We did that for a *reason*, guys...