IS MY TEST ACCURATE?
A tough question to answer! Because each type of COVID test is different, they all face their own unique set of challenges in trying to be as accurate as possible. Before we look at the accuracy of each test, let's first look into why testing accuracy is important, and then learn two new vocabulary words that we'll be relying on in this section: sensitivity and specificity.
​
WHY TEST ACCURACY MATTERS
Though the US is currently focused on vaccines to prevent COVID, in other countries where vaccines are not yet available, testing is one of the best ways to help prevent the spread of the virus. By identifying those that are infected, they are able to isolate and prevent themselves from spreading the virus to others. Test accuracy is key to successfully identifying those who are infected and those who are not.
​
There are two kinds of inaccuracies within COVID testing: false positives and false negatives. A false positive is incorrectly identifying someone who does not have the virus as infected, while a false negative incorrectly identifies someone with the virus as not infected. Each presents its own potential issues should they occur.
​
A false positive requires individuals to self-isolate, as they believe that they have been infected with the virus. This has the potential to worsen already difficult financial conditions, as isolation may mean that they are unable to work during that period. 25% of U.S. adults say that they or someone in their household has lost their job because of the coronavirus outbreak, and not being able to work would only contribute to the income loss many are already suffering. (Parker et al.)
​
A false negative allows an infected individual to behave normally, continuing to interact with others and potentially spreading the virus. This is one of the reasons that mask-wearing has been so heavily promoted. In order to avoid spreading the virus unconsciously, masks act as a further preventative measure against transmission. This way we are able to curve some of the spread from individuals who may unknowingly be infected. Next, let's learn about sensitivity and specificity.
SENSITIVITY AND SPECIFICTY
You likely already know these words, but they mean something slightly different when talking about testing accuracy. Sensitivity refers to the ability of a test to successfully detect what it is looking for. In the case of an antibody test for example, it is the ability for the test to detect COVID antibodies. Specificity refers to the ability to only identify the thing it is looking for. For an antibody test, it is the ability to only detect the COVID antibodies. We'll use apples and oranges as an example here.
​
Let's suppose that we have a test to identify apples, and we run it on the 6 fruits below. The + means the test result detected an apple, and the - means the test did not detect an apple. The results below would have perfect specificity and sensitivity, since it identified all the apples as apple, and did not incorrectly identify any of the oranges as apples
To demonstrate sensitivity, let's say that the test wasn't perfect, and one of the apples was not identified as an apple:
This test did not show perfect sensitivity. It was unable to detect one of the apples. In COVID testing, this would be known as a false negative, where someone with the virus got a test that incorrectly came back negative. Note that this test still has perfect specificity, as the positive results for apples were all correct. No oranges were incorrectly identified as apples.
​
Let's now look at a different result:
This test did not show perfect specificity. It incorrectly identified an orange as an apple. In COVID testing, this would be known as a false positive, where someone without the virus got a test that incorrectly came back positive. Note that this test still has perfect sensitivity, as all of the apples were correctly identified.
​
In short, sensitivity is the ability to hit the target, while specificity is the ability to hit the correct target.
Test Accuracies
Note two things as you read this information: First, that these meta-analyses looked at a variety of tests and sought to find the averages among them. These averages do not guarantee the accuracy of any singular test within that category, it instead presents the bigger picture of testing within a specific method. Please do not use these numbers to confirm or deny your own personal results.
​
Second, know that human error was not included in any of the inaccuracies. While some tests may be more difficult than others, and human error could potentially account for some inaccuracy, it is not measurable. We instead rely on the methodology presented in studies of accuracy. If no flaws are found after reviewing the methodology of the study, we do take a small leap of faith in trusting the authors that they have followed everything to their specifications. If in doubt, others may attempt to repeat the study in order to prove its validity.
ANTIBODY TEST ACCURACY
Numbers
According to a meta-analysis done by Mayara Bastos et al., the pooled sensitivity
of antibody tests was 84.3%. This differed by methodology. We discussed an in-lab
testing method, and this category of test shows a 97.8% sensitivity. However, there
are more rapid test methods, similar to the method we examined in the antigen
test section. This method (known as lateral flow) is more instant and can be done
on site. This method of testing however, only has an average sensitivity of 66%.
The specificity for both methods ranged from 96.6% to 99.7%
​
Causes of Inaccuracies
One of the problems with antibody test inaccuracy is the rapid commercial development of tests. According to Elitza Theel, the director of the Mayo Clinic’s Infectious Disease Serology Laboratory, "We now have more commercially available tests for SARS-CoV-2 than any other infectious disease." (Kwon) In order to expedite the testing availability, the FDA announced in March of 2020 that companies could market their tests as long as they performed their own evaluation. Though they later heightened scrutiny of these tests, it initially allowed for financially motivated companies to decide whether their product was ready for market, which has the potential to lead to companies selling less accurate tests. In fact, the same meta-analysis found that commercial kits had a lower specificity than non-commercial tests.
​
False positives may also occur with other coronavirus antibodies. SARS-CoV-2 is not the only type of coronavirus, and if other antibodies are present, they may be similar enough to register as SARS-CoV-2 antibodies.
​
Another issue is the timing of the test. An antibody test is designed to measure previous exposure. If an individual was recently exposed and has not developed antibodies, they will test negative. Remember that the antibody test looks for a previous infection, and should not be used to test for current infection.
“COVID-19 Strip Test.” Guyana Palace, www.guyanapalace.com/20-new-covid-19-cases-recorded-from-162-tests/.
I'm a paragraph. Click here to add your own text and edit me. It's easy.
ANTIGEN TEST ACCURACY
Numbers
In a meta-analysis done by Jacqueline Dinnes, the average sensitivity of
antigen tests: 55.6%. It should be noted that this study found a wide
range of sensitivities in the tests it investigated, from 0% - 94%. The
specificity of these tests averaged to 99.5%.
​
Causes of Inaccuracies
The timing of an antigen test is important. It needs a certain amount of virus present in a sample in order to be detected. If the virus is present, but in low quantities, it will give a false negative result. This has been one of the challenges with the virus, as there is a delay between infection and an effective testing period, making it difficult to know immediately if someone is sick. If an individual was infected 24 hours before taking the test, there will not be enough viruses present in order to register a positive result.
​
False positives are also possible after a person has been infected and recovered. These individuals can continue to shed viruses after their infection and recovery, though a majority of the time these viruses are unable to replicate. While typically in low quantity, it could present a false-positive result if there is enough virus present in the sample.
​
Collection of the sample is also important. As you may have experienced, samples are often taken with a nose swab that reaches quite far back. This is to ensure that a sample is being taken from your upper respiratory system, one area in which the virus resides. If the swab does not reach far back enough, it may not collect enough virus to be measurable. So if you ever have to endure the eye-watering swab of doom, just know it's for your own good.
Krystal. COVID-19 Test Kit. www.mindmusic.online/rapid-covid-19-testing-events-launched/.
RT-QPCR TEST ACCURACY
Numbers
RT-qPCR is the gold standard of testing, with specificities up to 100%
(Vogels et al.). Dr. Emily Volk, an assistant professor of pathology at the
University of Texas-Health in San Antonio and president-elect of the
College of American Pathologists (CAP), said that if performed correctly,
RT-PCR swab tests “would be pretty close to 100 percent accurate.”
(Curley) PCR is often used as the control when investigating other methods
of testing (as done with some of the previous studies), because it is so
accurate. From personal experience, I know that this is a test used by
UW medicine in their COVID testing facilities. 'Sko Dawgs!
​
“RT-PCR Lab Test.” Edward-Elmhurst Health, 2020, www.eehealth.org/about-us/newsroom/news/2020/06/covid19-tests-can-detect-active-or-past-infections/.
However, it is not without fail. When the virus RNA drops in quantity, the specificity also drops to 0%-50%. (Vogels et al.)
Causes of Inaccuracies
The timing of a PCR test is important. As with antigen testing, enough initial virus needs to be present in order for the RNA to be replicated and detected. If exposed too quickly after infection, the virus will not have replicated enough to be measurable, giving a false-negative.
​
Collection of the sample is also important. As you may have experienced, samples are often taken with a nose swab that reaches quite far back. This is to ensure that a sample is being taken from your upper respiratory system, one area in which the virus resides. If the swab does not reach far back enough, it may not collect enough virus to be measurable. So if you ever have to endure the eye-watering swab of doom, just know it's for your own good.
Again as with antigen testing, false positives are possible after a person has been infected and recovered. These individuals can continue to shed viruses after their infection and recovery, though a majority of the time these viruses are unable to replicate. While typically in low quantity, it could present a false-positive result if there is enough virus present in the sample.
TESTING TAKEAWAYS
By looking at the specificities and sensitivities of the various tests, you may have noticed that specificity remains quite high, while sensitivity can be less reliable. From our understanding of these measures, we know that this means false positives are unlikely, while false negatives have a higher chance of occurring. Interpreting these results, we can know to take positive results very seriously, as they are highly likely to be correct. Negative results however require more caution. We cannot guarantee that this result is correct, and in the interest of caution, should still institute a partial quarantine. Washington state currently permits a 7 day quarantine after receiving a negative test and experiencing no symptoms. (“Isolation and Quarantine for COVID-19 :: Washington State Department of Health”)
​
Testing timing is also very important. Testing too early can increase the likelihood of false-negative results, which as we have just discussed are already higher than we would like. After exposure, you should quarantine and wait at least 5 days before getting any test. This ensures that you are not accidentally exposing others to the virus, and ensures an accurate result. (CDC)