How do I pick the best COVID-19 viral test?
By Kyle Tretina, Ph.D.
- There are two common types of viral tests for COVID-19: (1) molecular tests, which can detect virus RNA, and (2) rapid antigen tests, which detect viral proteins.
- Important aspects of the test to note are EUA status, turnaround time, limit of detection, sensitivity and specificity.
- A test is also an experience, and practical considerations from collecting the sample to receiving results need to also be considered.
Information on testing is all over the place.
If you are looking for a COVID-19 viral test and are confused, annoyed, lost or desperate, you’re not alone. According to a New York Times survey, over half of Americans:
- believe that COVID-19 tests are mostly inaccurate,
- don’t know the difference between antigen tests and molecular tests,
- and/or think that all tests are equally accurate.
With misinformation easily found on the web about COVID-19 testing, it is important to find a trusted source of truth. If you fall into one of these categories, or have other questions about choosing a COVID-19 test for yourself or a group of people, you’re in the right place. Here, we will describe the most important features of COVID-19 viral tests to consider and help you make your decision.
An EUA is like a hall pass.
While many tests require approval or clearance by the U.S. Food and Drug Administration (FDA) for diagnostic use, we recommend tests for COVID-19 that have an emergency use authorization (EUA) from the FDA. This special kind of authorization is set aside for situations where the Secretary of the Department of Health and Human Services decides that it is necessary in a public health emergency, such as a pandemic. In order to receive an EUA, laboratories need to go through a verification process. However, EUAs don’t last forever — they are only meant to get us through the emergency.
As is often the case during public health emergencies, the standards required for clinical trials haven’t yet been established, and so the EUA verification process focuses mostly on how well the test performs in a laboratory. Combined with the fact that there are laboratory tests that are available and do not have an EUA, it is even more important to understand how to compare and contrast the tests.
COVID-19 viral tests detect specific parts of the virus.
The first key to understanding the tests for COVID-19 is to learn the parts of the virus that causes it, which is called SARS-CoV-2. SARS-CoV-2 is one member of a very large family of coronaviruses, some of which infect humans and cause the common cold. As a coronavirus, SARS-CoV-2 has three essential components. At its core, there is the genetic material of the virus, the RNA, which is a single strand of molecules that contains the instructions for all of the viral structural proteins: the spike (S) protein, nucleocapsid (N) protein, membrane (M) protein, and the envelope (E) proteins (Figure 1).
Out of these five components (RNA + four proteins), two are the target of common viral tests. These tests are used to detect a current infection, and typically require the collection of a nasal swab or saliva sample (Table 1). These kinds of tests are generally recommended for people who have symptoms of COVID-19 or have specific reasons to believe that they might be at a higher risk of COVID-19, such as close contact with an infected person, have been travelling, attending a large gathering, or people who have been referred for testing by a healthcare professional.
Molecular tests vs Rapid Antigen Tests
The first group of viral tests are often called “molecular tests”, and can detect very small amounts of viral RNA in a sample (Table 1). We won’t go into details here regarding how these work, just know for now that they can detect the RNA part of the virus. If you go to a local testing site for a diagnostic test, there’s a good chance that you will be getting a molecular test, which needs to be submitted to a lab and could take a couple of days to get a result.
The second group of tests are often called “Rapid Antigen Tests”, can detect proteins on the surface of the virus, and are not as widely available. However, if you can get one of these you can get testing results in as few as about 5 to 30 minutes. Rapid antigen tests are also cheaper to produce. As a result, antigen tests are being used to screen large numbers of people, such as at airports.
If you test positive using either a molecular test or a rapid antigen test, that means that the virus was found in your sample and that you are likely infected with the SARS-CoV-2 virus. This doesn’t necessarily mean that you are contagious (especially if you don’t have symptoms), but you should self-isolate and contact your healthcare provider regarding what you should do next.
Pay attention to these test description numbers.
In order to properly judge a COVID-19 test, you need to understand the timeline of infection (Figure 2a). After exposure to the virus, there is a period of time where the virus grows unnoticed, and about 4–5 days later, symptoms appear. In some cases, the incubation period can be as soon as 2 days and other times it can take as long as 14 days. People can become infectious to others for a few days before they get any symptoms, and this is thought to be a time period that leads to a lot of viral transmission. Interestingly, there can be a couple weeks of time after a person is no longer contagious where they will still show up positive on a viral test.
Turnaround time: a test is useless if it takes too long.
This is often the most important test metric you will run across, as studies have shown that turnaround time is even more important that test sensitivity for surveillance purposes. Any test result that you receive over 3 days from sample collection is probably not useful. Fortunately, there are many testing options, both rapid testing, where results are often available minutes from sample collection, and molecular tests, where results can often be received within 24–48 hours.
Limit of Detection: How little virus can this test detect?
The limit of detection (LoD) of a test, is the smallest number of viral molecules needed to be detected by the test (with a 95% probability), so a lower number for the LoD is better. Figure 2a shows how the LoD is one of the most important numbers to know for any test that you use. A nasal swab can have as many as 10^8 copies of virus at its peak, but early and late during the infection, the virus can be present at as low as only a few particles per sample. The LoD can therefore determine whether you can detect an infection before symptoms appear, but where the person is still infectious. Or, whether you can detect the presence of the virus after the person is no longer infectious (Figure 2a). A look at actual LoDs from tests on the market reveals a large range of possibilities, with >10,000 fold differences between the least and most sensitive tests (Figure 2b). The best tests have an LoD ~100 copies / mL or lower. It has even been estimated that the tests with the highest LoDs will not be able to detect 70% of infections.
Sensitivity is how often the bullseye is hit.
For a COVID-19 test, sensitivity is a measure of how often the test identifies an actual infection (“true positives”). So if a person actually has COVID-19, sensitivity indicates the likelihood they will actually test positive with that test. For a good test, this number is very close to 1. For example, a sensitivity of 0.95 means that 95% of the time a “true positive” will test positive with that test. This corresponds to a 5% false negative rate. The best tests have a sensitivity of 0.95 or better.
Specificity is how often the noise is ignored.
Conversely, specificity is a measure of how often a test identifies someone who is not infected (“true negatives”). If a person does not have COVID-19, specificity indicates the likelihood that they will actually test negative with that test. For a good test, this number is also close to 1. For example, a specificity of 0.95 means that 95% of the time a “true negative” will test positive with that test. This corresponds to a 5% false positive rate. The best tests have a specificity of 0.95 or better.
Price and third party reimbursement amount
For a single COVID-19, the price can vary from around $20 to $850 (as of March, 2021), with the most common test price range $100-$149. In early 2021, the Biden Administration provided new guidance that insurers have to cover testing without cost sharing for asymptomatic individuals and without requiring medical screenings. However, this is not required if the testing is a part of an employee return-to-work program or for public health surveillance purposes. This means that some patients with health coverage could receive bills for COVID-19 diagnostic tests, and those bills can vary widely. The availability of community-based testing sites varies by state and is provided by the U.S. Department of Health and Human Services, but there are some online services you can use to get your test more conveniently.
A diagnostic test is more than a sample collection.
As with any product, a COVID-19 test is more than just a price or a performance measure, and other factors have to be considered that are more specific to your circumstances, such as the convenience of ordering the test, whether the test can be taken at home or the sample collected under supervision of a medical professional, what sample type is required (e.g. swab, saliva, etc.). Is shipping included in the cost of the test? How clear are the instructions for collecting the sample? How will the results be delivered to you (email, in-app, etc.)? In the end, this decision comes down to trust as much as test performance. Whether you are testing in preparation for travel, to get out of quarantine, or to screen for an event, will your test provider have active customer service to help if you need them?
Limit of Detection: the minimum number of viral particles needed to detect SARS-CoV-2 with a 95% probability.
True positive: Sick people correctly identified as sick
False positive: Healthy people incorrectly identified as sick
True negative: Healthy people correctly identified as healthy
False negative: Sick people incorrectly identified as healthy