When is a positive lab test really positive?

On a pregnant woman’s first visit to the antenatal clinic it is routine practice to send a blood specimen to the microbiology lab to test for blood borne viruses, such as hepatitis B and HIV as these can be passed on to the baby around the time of birth. We recently had a mum-to-be whose blood specimen tested positive for HIV, a result which would be devastating at a time of joy and anticipation. So how did we approach this highly sensitive issue?

Before we answer this question it’s worth looking at how we interpret test results in the lab, in particular two features of the test used: sensitivity and the specificity.

The sensitivity of a test is a measure of its ability to identify correctly the infection in the patients who are tested. No test is 100% sensitive and this means that there will always be some patients who have the disease which is not picked up by the test.

Specificity is the capacity of a test to correctly exclude those patients who genuinely don’t have a particular infection or disease. Again, no test is 100% specific. So, there will be some patients who test positive for an infection but don’t actually have the infection – these tests results are called “false positives”.

Although we can never achieve 100% sensitivity or specificity, many current tests have sensitivities and specificities of 98 or 99%. This is pretty good, but sensitivity and specificity isn’t the whole story. We also need to take into account how common an infection is. If it’s very common, then the positive predictive value of a test is high. The positive predictive value is a good way of estimating the probability that a positive test really is positive and that the patient does, indeed, have the infection.

The positive predictive value is obtained by dividing the number of true positive results by the number of all positive results (that is, true positives + false positives).

HIV infection in the area served by my hospital is rare, so the chance of this positive result reflecting that our expectant mum truly has HIV are actually quite low, even though we use a test with good sensitivity and specificity. If, however, I were working in a hospital in sub-Saharan Africa where HIV is common then the chance of her being HIV-positive would, sadly, be high.

Where did this leave our patient? Recognising that this could be a false-positive, we sent the blood to another lab, which tested for HIV using a number of other tests. These were all negative. Our test was, indeed, a false-positive meaning, fortunately, the patient didn’t have HIV after all. So we have to be cautious in interpreting results of the tests we do in the lab as “positive” may not really mean “positive”.