How a 99.99% Accurate Test Can Be Wrong Half the Time
By Christine Maggiore
 
“A new article in Popular Science explains how allegedly accurate HIV
tests can be dead wrong when performed on people with no risk factors…”
 
===
 
An article in the July 2002 issue of Popular Science magazine gives
readers some surprising food for thought: the allegedly accurate HIV tests
can be dead wrong when performed on people with no risk factors.

Here’s
the text from page 78:
 
"HIV testing is 99.99 percent accurate, a doctor might tell his patient.

That suggests that if you receive a positive result, you almost certainly
have HIV. But this is not necessarily the case. The chance of a straight
man with no known risk factors contracting HIV is roughly one in 10,000.

That is also the rate at which an HIV test returns an incorrect result.
So
if 10,000 men in this low-risk group get tested for HIV, an average one
positive will come back from the man with HIV, and another man will test
positive even though he is not infected.

Thus, in our statistically perfect world, only one of the two men who test
positive actually has HIV. 50 percent of positive HIV tests in the
low-risk group turn out to be false."
 
The article incorrectly assumes that HIV tests can diagnose HIV infection
and that the claimed accuracy rate of 99.99% has been established through
careful scientific studies.

In fact, HIV tests have not been approved by
the US Food and Drug Administration for use in diagnosing actual HIV
infection, and all claims of test accuracy are based on estimates and
assumptions.