Replies

"Wonder what the false positive rate is?"

That's a good question. A related, and even better, question is what is the positive predictive value (PPV) of the test? In other words, what is the probability that someone who tests positive really is HIV-infected? The answer depends on the false positive rate *and* the prevalence of HIV in the tested population. The CDC claims that about 1 million people in the U.S. are HIV- infected. Let's assume that all of them are among the 230 million people in the targeted test population of 16 to 64 year olds. So, the claimed prevalence of HIV is 1/230 = 0.00435, or slightly more than 4 per 1000. The CDC also says that the false-positive rate (one minus the "specificity") for the rapid HIV test that is being recommended is 0.2% or 2 per 1000. (The false positive rate is essentially zero.) That means that the PPV is 4/6 = 0.66, so one-third of all people who would be tested under this plan will test positive even though they are really HIV-negative.

Testing the entire population for a low-prevalence disease when there is a non-trivial false-positive rate is a huge waste of money. It's why we don't routinely test everyone for TB.

so one-third of all people who would be tested under this plan will test positive even though they are really HIV-negative

I'd be terrified of the damage that this could cause. I have nothing to worry about - but - Speaking only for myself, if I incorrectly tested positive, it would take a long time and a lot of negative tests to completely erase the worry. It's the old saw about a person with one watch knowing what time it is, but a person with two watches never being sure.

Not to mention damage to my marriage, etc. "I tested positive, then negative, so it's nothing to worry about, honey." Yeah, right.

"That means that the PPV is 4/6 = 0.66, so one-third of all people who would be tested under this plan will test positive even though they are really HIV-negative."

You mean one-third of people who will test positive are actually negative. Since the number of people testing positive is pretty miniscule (6 in 1000), I don't think that's unreasonable. More tests would be done to eliminate rule out false positives.