The cost of a false positive test

False Positive (Can This Marriage Be Saved?) – tells a poignant story. I will make the case that overtesting can lead to problems because of the false positive problem.

Over the years I have taught many students and physicians about diagnostic tests. Several principles concerning diagnostic tests are very important. First, each test has a sensitivity and specificity. Sensitivity defines the probability that a patient with the disease has a positive test. Specificity defines the probability that a patient without the disease has a negative test. The false positive rate equals 1 minus the specificity.

Now that I have demonstrated my nerdiness once again, let us understand the problems that these definitions cause. The problem we are discussing today is the false positive problem. Patients and physicians often worry about false negative tests – missing a diagnosis. However, as this case demonstrates a false positive diagnosis carries important costs. I usually use the example of a colleague’s patient (who happened to be a 45 year old lawyer). This lawyer went to give blood. The initial screening test for HIV read positive. The patient had no known risk factors for HIV. It took my colleague 3 months to prove to everyone’s satisfaction that the patient was HIV negative. What was the cost to the patient? This occured early in the HIV epidemic, prior to any antiretrovirals. Our testing was rudimentary. Prognosis for AIDS was less than 1 year.

Multiply this example by the extensive testing that many patients get. Even with a specificity of 99% (pretty damn good), 1 of 100 healthy patients will receive a false positive diagnosis. And they will have anxiety and doubt.

Now read the story – especially these quotes:

“Buenos dias,” I say. “Soy Doctor Eshleman. Cómo se siente hoy?” and reflexively brace myself for the torrent of Spanish that usually follows my feeble attempts at bilingualism. Instead, the nurse speaks to me in English, telling me that Mr. Guerrero is here for treatment for syphilis and that he really doesn’t understand how this could be. He is very sad because last month, when he first learned of the diagnosis from a blood test, he was advised to have his wife tested. His wife was shocked when she was told that she needed to be checked for a venereal disease. She accused her husband of being unfaithful and moved out of their home and flew back to El Salvador. That’s why he had been crying.

I knew there was a problem before I entered the exam room.

In reviewing Mr. Guerrero’s medical record, I noted that he had had two recent blood tests for syphilis. One was the VDRL test — VDRL stands for “Venereal Disease Research Laboratory,” the U.S. Public Health Service (PHS) facility where the test was developed — while the other was an FTA test. (Those initials stand for “fluorescent treponema antigen,” another PHS test. Treponema pallidum is the bacterial species that causes syphilis.) I also noted that he had received two injections of penicillin.

The VDRL is a screening test. A positive VDRL does not necessarily mean one has been exposed to syphilis. If the VDRL is positive, clinical laboratories will routinely perform the confirmatory FTA test. If that’s positive, it clinches the diagnosis. In Mr. Guerrero’s case, his VDRL was weakly positive and the FTA was negative.

I found all this confusing, because my interpretation of his test results was that he did not have — and has never had — syphilis. Instead, his test result was almost certainly a biological false positive. Lots of other diseases, including lupus, rheumatoid arthritis, mononucleosis, malaria, HIV, mycoplasma pneumonia, leprosy and Lyme disease can give a biological false-positive VDRL. In parts of the tropics and of Latin America, the diseases yaws and pinta, which are caused by a bacterium that is a close cousin of T. pallidum, will also give a false-positive VDRL. And sometimes there’s just no explanation for a biological false positive. But false-positive VDRL tests for syphilis are common, which is why the confirmatory FTA test — the one that actually shows the presence of the bacterium — is so important.

This case (and the one I recalled) put faces and feelings into those numbers. We should never forget the faces and the feelings.