Tests that test for pregnancy…also used to test for a “specific” virus?
From the review:
“A strongly reactive ELISA result is more likely to indicate that the person is infected (true-positive result) than a weakly reactive result.16-17
- There are many ‘levels’ of reaction, but no standard as to what constitutes a true versus a false positive.
The ELISA screening test is inexpensive and is designed with a low threshold so that few infected individuals will be missed. However, like other screening tests, it will also capture some uninfected individuals.
- Good to know! So who are the unfortunate false-positives? And how do you know how many are really false, if there are no standards for what is true?
“This is especially a problem when testing is expanded from a high-risk population to a large low-risk population.18”
- When moving from poor and brown-skinned, to white and wealthy, the “accuracy” of the tests decreases. The results may be identical, but while you can call any poor, Black woman “Hiv Positive” with total cultural approval, you can’t do the same to the white, wealthy, suburban housewives of America (or Europe).
“The PPV of the ELISA test has been reported to be as low as 2% for a weakly reactive ELISA result in a low-risk population and as high as 99% for a strongly positive ELISA result in a high-risk population.14”
- Two to Ninety-Nine percent? Again, good to know. That is, the tests are only as accurate as the level of poverty – not sex – that you’re dealing with.
“Expanding the test population to include more low-risk people can be expected to result in a proportionate increase in false-positive test results and to decrease the PPV.”
- So don’t test the “low-risk” (white and wealthy and straight), just the poor, gay and Black.
“False-positive ELISA test results can be caused by alloantibodies resulting from transfusions, transplantation, or pregnancy,13, 19-20 autoimmune disorders,13, 15, 20 malignancies,15, 19 alcoholic liver disease,15, 19, 21 or for reasons that are unclear.”
- Good to know, huh?