The different types of Covid-19 tests are far from equal. Picking a test is generally a matter of speed versus accuracy and, most importantly, why you need a test and when.
Having received many questions about Covid-19 tests, Spotlight set out to find the answers to some of the most common ones. What do the different tests cost? What are the limitations of the various types of tests? Is there any quality control? What are the chances of false positives and false negatives?
Before getting into the different options, some terms are important to understand. The accuracy of a test is usually described in terms of sensitivity and specificity:
- Sensitivity is a measure of how well a test picks up the thing being tested for. If ten people with Covid-19 are tested using a test with 90% sensitivity, then nine of them will on average test positive. The higher the sensitivity, the less likely the chances of false negatives. In the example above, around one out of every ten people with Covid-19 will receive a false-negative test result.
- Specificity, on the other hand, is a measure of how well a test does at not picking up the thing being tested for when that thing is not present. If ten people who do not have Covid-19 are tested using a test with 90% specificity, then nine of them will on average test negative. The higher the specificity, the less likely the chances of false positives. In the example above, around one out of every ten people who do not have Covid-19 will receive a false-positive test result.
- While sensitivity and specificity describe the inherent accuracy of a test, additional contextual factors have to be taken into account when interpreting test results in the real world. These factors fall under the umbrella term “pre-test probability”. One key factor is the prevalence of a condition, such as Covid-19 in a community. If 1% of people in a community have Covid-19, and we tested 100 people with a test that has a sensitivity of 100% and a specificity of 90%, then chances are we’d detect the one true positive (since sensitivity is 100%), but we’d also detect 10 false positives (since specificity is only 90%). Thus, we would end up with 11 positive tests and for each person who tested positive, there would be only a one in 11 chance that they are truly positive. (This BMJ piece from earlier in the pandemic provides an accessible guide to how one might take these various factors into account.)
The sensitivity and the specificity of the different types of Covid-19 tests differ substantially, as does the sensitivity and specificity of different tests of the same type (although generally to a lesser degree).
To complicate things further, the sensitivity and specificity of a test will also differ depending on how many days it is since someone was potentially exposed to SARS-CoV-2 (the virus that causes Covid-19).
The different types of tests also test for different things - so thinking of them as tests that simply tell you whether you have Covid-19 is oversimplistic. Being highly sensitive to antibodies is different from being highly sensitive to viral particles.
There are three types of Covid-19 tests available in South Africa:
- PCR: PCR (polymerase chain reaction) tests are currently regarded as the gold standard for Covid-19 testing. A sample taken from the nose or throat is multiplied (the technical term is amplified) in a machine in the lab and the SARS-CoV-2 viral DNA is then detected if it is present in the sample. PCR tests have both high sensitivity and high specificity. Because PCR tests are highly accurate at detecting viral DNA, they are recommended for diagnosing active Covid-19 infection and it is the test most people with Covid-19 symptoms in South Africa would have received. Results can be available in a few hours or after a few days. (You will also sometimes see it written as rtPCR, which is technically more correct. The “rt” stands for reverse transcription, an extra step that is required to turn SARS-CoV-2 RNA into DNA.)
- Antigen: A rapid antigen test works by looking for proteins on the surface of the SARS-CoV-2 virus. As with PCR tests, samples are typically taken from the nose or throat, but unlike PCR tests, samples do not have to be sent off to the lab but can be analysed at testing facilities - with results usually available in less than half an hour. Antigen tests typically have high specificity (which means the chance of false positives is low), but they are significantly less sensitive than PCR tests (and accordingly, there is a greater risk of false negatives). Antigen tests also have a narrower window than PCR when it comes to timing. They work well when people have high viral loads, while PCR tests continue to work well even when people’s viral loads have declined.
- Antibody: Unlike PCR and antigen tests, antibody tests are not intended to diagnose current Covid-19 infection. Instead, they are used to detect previous infection. As the name indicates, antibody tests detect antibodies rather than viral DNA or viral proteins. It typically takes a few weeks from SARS-CoV-2 infection or vaccination before the body has detectable levels of antibodies. It is also worth noting that there are different types of antibodies that are picked up by different antibody tests. If you get an antibody test, it is worth asking what types of antibody the test will detect and what that means in terms of previous infection or vaccination. Results from antibody tests are usually available in 15 minutes or so - although there are more complete antibody tests that can take a day or so.
At the time of writing, there are 48 approved antigen tests in South Africa, 46 antibody tests, and 112 PCR tests. A list on the South African Health Products Regulatory Authority (SAHPRA) website is updated as additional ones are added. A few of the tests are local, but the majority are imported.
Where to get them and what they cost
To get any of these tests, you will either be tested when in hospital or have to go to a pharmacy or testing centre. Most antibody and antigen tests results are available within a 15 minute or so wait at the testing centre. Results of PCR tests take longer and are typically sent to you by SMS or communicated to you by your doctor.
No self/over-the-counter rapid antigen tests have so far been approved in South Africa. Different countries have taken different approaches to such self-tests. The United States and United Kingdom have given them the green light, while for now they are still prohibited in Australia (though processes are underway to make them available there as well).
The costs of the tests in the National Health Laboratory Service (NHLS), which handles Covid-19 testing for the public sector, are R507 for PCR tests, R150 for antigen tests, and R200 for antibody tests.
As expected, the private sector prices are slightly higher than those in the public sector. Sandile Sihlaba, spokesperson for Lancet Laboratories says the National Pathology Group (NPG), a group consisting of South Africa’s three largest private pathology companies, has agreed on pricing for all affiliated laboratories of R850 for PCR tests, R200 for antigen tests, and R248 for antibody tests.
How are these tests approved?
Yuven Gounden, spokesperson of SAHPRA, says the role of the regulator is to look at the safety and performance of products, including for these test kits. Before tests are used in South Africa, technical documents have to be submitted, which are reviewed by a committee to ensure that they meet the requirements for sensitivity and specificity.
“If the requirements are met, the kits are further validated by the NHLS to ensure the tests perform accordingly,” says Gounden. “The test kits are authorised for use and distribution via Section 21 authorisation.
The company is issued an establishment licence listing for the approved test kit. The applicant will share the application with the relevant documents as per the guidelines published on the SAHPRA website. The application will be reviewed for compliance as per the set requirements in the guideline.
“When all requirements are met,” Gounden says, “SAHPRA will issue a letter for performance evaluation with the National Reference Laboratory [which is the NHLS]. The laboratory’s test results, together with the product technical dossier, will be reviewed. If all requirements are met, the product will be included on the applicant license and on the SAHPRA website.”
Different tests, different uses
Andy Gray, senior lecturer in the Division of Pharmacology at the University of KwaZulu-Natal, says As long as the test being used is SAHPRA approved, then the performance should be as expected. “There is a trade-off between cost and speed and performance. The rapid antigen tests are faster and cheaper but not as sensitive or specific as PCRs, but that is well understood by the NHLS, which is why they also use both,” he says.
He adds that the PCR test is extremely sensitive and specific, but the rapid antigen tests less so, though still useful, provided the pre-test probability is taken into account and negative tests are confirmed when pre-test probability is high.
“An antigen test is used for screening in a high prevalence setting where the patient is more likely to test positive,” says Mzi Gcukumana, senior communications manager at the NHLS.
“This test is generally considered accurate. Antigen tests can be 100% sensitive in the correct population. Sensitivity drops if the patient is asymptomatic or if the test is performed too early,” he says.
“The antibody test is used for retrospective diagnosis, seroprevalence, and specific patient populations like children with multi-system inflammatory syndrome (MIS-C),” says Gcukumana. “The accuracy of this testing method depends on the test, the patient population, and the timing of testing.”
According to Gray, antibody tests are not useful for diagnosing acute cases and should not be used as such. But these tests can be useful research tools to conduct seroprevalence studies,” he says.
To complicate matters further, it turns out the timing of a test and how the sample is taken also matters a lot.
According to a recent article in the Journal of the American Medical Association (JAMA), PCR tests’ high sensitivity isn’t always a good thing. They can return positive results weeks after someone has recovered from Covid-19 and is no longer infectious.
In infected people who become symptomatic, the highest risk of transmission appears to occur two days before symptom onset through to five days after, while asymptomatic people seem to clear the virus even more quickly.
In theory, a benefit of rapid antigen tests is that they allow for more, and more regular, testing. But as indicated in the JAMA article, research is beginning to suggest that they aren’t as sensitive in the real world as they appeared to be in clinical trials. “That’s not always the fault of the tests themselves but is sometimes due to user error, such as testing too soon or too long after a possible SARS-CoV-2 exposure or not correctly swabbing the nose,” according to the study.
Gray concurs that the way a sample is taken may influence the test result. “There is some suggestion that a mid-turbinate specimen (middle of the nose) is less sensitive than a nasopharyngeal specimen. About samples contaminating each other, again, this is a theoretical risk but has not been born out in studies,” he says.
But Gray says one has to remember no test is 100% accurate, but different reasons may apply in different cases and with different tests. Both false positive and false negative tests are possible. And repeated tests may give different results, sometimes appropriately, such as when the viral load is higher and more easily detectable.
Does the tester have any legal obligation to inform the client about the limitation of the tests?
“Not specifically for these tests,” says Gray, “but there is an obligation on all health professionals to inform patients about any tests conducted on them.
All tests have limitations.” He advises people that they familiarise themselves with the tests and obtain the correct information about which test to conduct and why, and then get assistance on how to interpret the test results and what to do thereafter.
*This article was produced by Spotlight - health journalism in the public interest.