They are not perfect, but a test doesn’t have to be perfect in order to be useful.
This article was originally posted in the Montreal Gazette
Christopher Labos MD, MSc | 30 Dec 2021
Starting this week, Quebecers are able to get rapid test kits from pharmacies so that they can test themselves at home. Compared to going to a test centre where lineups are starting to stretch into hours-long waits, the convenience of an at-home kit is obvious.
The tests can be done anywhere and are easy to perform. Everything you need is in the testing kit. They are less expensive than PCR tests and provide results in 15 minutes.
On the surface, the appeal of these tests should be obvious, and yet their rollout here has been slow compared to other countries, especially in Europe, where they have been widely used for months.
The fear has been that these rapid tests would be less accurate than standard PCR tests. But to truly evaluate their usefulness we have to look at both their sensitivity and their specificity.
A sensitive test would correctly identify everyone with COVID-19 as positive. A specific test would correctly identify everyone without COVID-19 as negative.
Ideally, you would want a test to be both sensitive and specific, but practically speaking you often must sacrifice one for the other.
The issue with rapid tests is an issue of sensitivity. In other words, some people with COVID-19 do not test positive.
There are multiple reasons why this might happen. The rapid antigen tests detect a protein that makes up part of the virus and can tell you whether you have an active infection.
But if you test too early in the course of the infection there may not be enough replicating virus in your nose for the test to detect it.
Similarly, if you test too late after symptoms start, then the viral load may be decreasing and also be undetectable.
Reviews of various rapid antigen tests show they perform better if done in the first week of infection.
Technique also matters. If the swab gets an inadequate sample, then the test may be inconclusive or falsely negative.
So while the procedures may be daunting for people unfamiliar with them, explanatory videos of rapid tests should go a long way toward demystifying the procedure, which actually is fairly simple.
The other issue that can affect the sensitivity of the test is whether someone is symptomatic. Because you need to need a certain amount of virus to be detectable by the test, asymptomatic individuals are more likely to have either mild infections or be too early in the disease course for the test to pick them up.
In the Cochrane review of rapid antigen tests, sensitivity was higher in symptomatic vs. asymptomatic individuals (78 per cent vs. 58 per cent) whereas specificity was high for both groups (99 per cent vs. 98 per cent).
In short, the tests are good at ruling in the disease, but perform less well at ruling it out, especially in asymptomatic individuals.
One final wrinkle is that the accuracy of these tests changes depending on the epidemiological situation. If one uses a different branch of statistics, called Bayesian statistics, one can see the probability of false positive and false negatives can change depending on how much COVID is circulating in the community.
When the prevalence of COVID-positive people goes up, the positive predictive value of the test goes up while negative predictive value goes down. With a test positivity rate of 7.4 per cent as of Monday, the likelihood of a rapid test being wrong is quite low , but not zero.
A negative rapid test is not a licence for carelessness and should not be used as a justification to throw caution to the wind and start holding large parties. But if used properly in tandem with all other public health measures, they can help reduce viral spread. We must remember that they are not perfect, but a test doesn’t have to be perfect in order to be useful.
https://www.mcgill.ca/oss/article/covid-19/some-things-know-about-covid-rapid-tests
