Several self-tests are claimed online that are said to indicate whether you have been infected with the coronavirus. According to the National Institute for Public Health and the Environment (RIVM) and the World Health Organization WHO, these tests are unsuitable for individual use and do not provide reliable results.

When you google 'self-test corona', you immediately come across several online providers who promise to tell you whether you have already been infected with the COVID-19 virus. However, authorities warn against tests of this kind: they could lead people to mistakenly think they are immune to the virus and are less likely to comply with measures against it.

Two types of corona tests

In the Netherlands, professionals use two types of corona tests.

The GGD uses a so-called PCR test. For this, a slime sample is taken from the throat and nose. This test only makes sense when someone may be infected and cannot tell you if you have had COVID-19 in the past.

In addition, the Sanquin blood bank, for example, uses a serological test to see how widespread the coronavirus is in the Netherlands. Many of the self-tests offered online are also serological tests.

Antibodies to the virus in the blood can be detected with this type of test. A test result can be positive or negative. With a positive result, antibodies are found in the blood, with a negative result they are not found. According to RIVM and WHO, serological tests can be useful for research, but are currently not suitable as a self-test or for determining whether someone is immune.

See also: For example, Sanquin investigates how many people have corona antibodies

Why personal results from serological tests say little

In a letter dated April 8, WHO explains why. First, according to the WHO, we are not yet sure whether the presence of antibodies in the blood actually means that someone is immune to the coronavirus. It is therefore unclear whether someone who demonstrably has antibodies can still spread the virus or become ill themselves.

The WHO also writes that a test result is sometimes incorrectly positive and therefore incorrectly reports that someone has had the virus. This is because the active substances in the tests sometimes react not only to antibodies to the coronavirus, but sometimes also to antibodies to other viruses that resemble the coronavirus.

In conversation with, RIVM therefore emphasizes that a positive result of a serological test does not mean that you no longer have to keep a distance of 1.5 meters or that you can visit your ninety-year-old grandmother with a cold.

According to the WHO, a test can also be incorrectly negative, for example if you have the test taken soon after an infection and at that time you have barely produced any antibodies.

What about population screening?

According to the authorities, these tests are therefore not useful at an individual level. But this type of testing is used, for example, to determine how widespread the coronavirus is in a country or region. How is this possible?

Sanquin previously explained to how it ensures that his research into antibodies among blood donors is reliable. The test that Sanquin uses to detect antibodies is also sometimes incorrectly positive. However, Sanquin knows how often this occurs, because it has also tested blood from its archive with the test. Of these, it was certain that it could not have been infected with the coronavirus, because it had contracted before the coronavirus outbreak.

Sanquin takes this into account when calculating the percentage of blood donors that have produced antibodies against the coronavirus, which means it can still call it a reliable percentage. Other corona research conducted among many people also sometimes uses such serological tests. Researchers then take into account in their calculations that the tests are sometimes wrong.

Calculation example: why serological tests at an individual level are now unreliable

  • With a calculation example we illustrate why serological tests at the individual level are currently unreliable. For this example we use an imaginary test. This test is always correct if someone has actually produced antibodies against the coronavirus - the result is always positive. And in only 3 in 100 people (3 percent) without antibodies to the coronavirus, this test gives a false result.
  • We assume that 3 percent of the Dutch have produced antibodies against the coronavirus and for the sake of convenience we pretend that the Netherlands has only 1,000 inhabitants. All these 1,000 Dutch people undergo our imaginary test.
  • 30 out of 1,000 Dutch people (3 percent) get a positive result, because they have produced antibodies against the coronavirus - this result is correct. Of the other 970 Dutch people, another 29 people (3 percent) get a positive result, because this test also reacts to something other than antibodies against the coronavirus - this result is not correct. The rest of the Netherlands is getting a negative result.
  • So, in this example, 59 people get a positive result. This result is not correct in almost half of these people. A positive result therefore tells you little in this case.
  • When the number of people who have actually produced antibodies against the coronavirus is higher, the chance that a positive result is correct is also higher. De Volkskrant and the Washington Post previously published a clear explanation of this with slightly different calculation examples.


Criticism of corona rapid test: 'There are also errors in RIVM tests'