Replies to post #16986 on Biotech Values

[DewDiligence]

>>This is why if you test positive for HIV (or some other diseases), they use a different type of test to double check the results.<<

I used the term standalone test to allow for the possibility of a follow-up by either another test or continued use of the same test.

In the puzzle problem, the test is weak when run only once, but it may become considerably more useful if repeated. Based on the problem specification (and assuming for simplicity a zero rate of false negatives) then the probability of actually having HIV after two successive and independent positive results is about 40%, and the chance after three such positive results is close to 90%.

One caveat to the above is that the false-positive error rate in the test may represent a systemic bug in the test design rather than a random hiccup, which invalidates the assumption that successive tests on a given patient are independent of one another.

A second practical caveat is that our medical reimbursement system may not be smart enough to recognize the value of repeated testing with the same test.