translated from Spanish: Are Chinese tests done at train stations “defective”?

Yesterday a note came out in an important national newspaper stating that the tests that had begun to be used at train stations were “defective” and that, therefore, they might not be useful in fulfilling the purpose they would come to fulfill, which is to monitor the epidemiological situation. The note cites a report by CONICET and the Ministry of Science and the opinion of some biochemists who, with a rare argument, argue that obtaining eight positive cases in a sample of 1200 tests performed is too little, and that therefore it is necessary to suspect their validity. This argument is rare, I say, because when the samples are taken completely randomly (these are not symptomatic, or suspicious cases, but of volunteers who undergo the test) it is expected that the positivity will be low, especially given that, because of quarantine, the circulation of the virus remains low in our country, even in CABA and the province of Buenos Aires (the most affected districts). Let’s look at this with one example: the AMBA has a population of about 15 million; of those 15 million, so far there are 4000 confirmed cases. This means that we have approximately one case confirmed by every 3750 inhabitants. Even if there were ten times as many cases as confirmed, which is perfectly possible, it would be expected to find a case every 375 random tests performed. I mean, it’s not uncommon for 1200 tests to look at eight positives. With this I want to be clear: the positivity that is expected in a PCR test, which is done according to the definition of suspicious case, is not the same as that expected in a randomly performed serological test. In fact, if positivity is low, it is not a reason to worry but to suspect that the disease is poorly dispersed. What are these tests?
Remember that the only type of test used for diagnostics is PCR, which I explained in this note. The tests that were done in the seasons use another technique and measure something else: while the PCR test detects the genome of the virus, and therefore can tell whether or not a person is infected at the time the sample is taken, serological tests measure antibodies, one of the substances that our body produces to fight the virus when infected. Two antibodies are detected against SARS-Cov2: IgG and IgM. The problem is that these antibodies are not detectable from day 0 of infection but take several days to reach enough concentrations to be detected. IgG, for example, begins in 90 percent of cases after the second week of infection. And besides, they can be detected when the virus is no longer in the body, because they were circulating in the blood. That’s why they’re not used to diagnose the disease: having the antibodies doesn’t mean you’re infected, nor does it mean you’re not infected. It means you had contact with the virus.  This has been known long before testing how effective these tests are in particular. The tests used at train stations are serological tests that were not purchased by the state but were donated by a company (Petroquímica Cuyo), which bought them in turn from Zhuvai Livson Diagnostics, a Chinese company. What the state did, once it received the donation, was to have a rehearsal to see how they worked and determine what they could be used for. This test, conducted by the Ministry of Science and CONICET, concluded that in 80 per cent of cases the detection of IgG worked well, and that for IgM the detection was very weak. So they don’t work… do they?
This doesn’t mean they don’t work. It means, in any case, that they do not serve to measure IgM but to measure IgG, not to diagnose (which, I insist, we knew before) but to get an idea of the state of circulation of the virus. When designing the epidemiological study, the degree of sensitivity of the test to be used to do so is taken into account. If igG is known to be 80 percent, it will be considered that if, for example, eight cases tested positive, those who have had contact with the virus probably were not eight but ten. Since it’s not about diagnosing, but having a circulation map, the data is still useful. So to say that the tests are “defective” is simply incorrect. They are serological tests, are not used for diagnosis and remain useful for research purposes. In this note:

Original source in Spanish