Weekly newsletter:
January 2021
Level difficulty:
The sensitivity of tests is secondary to the speed of results
This can be represented schematically as follows:
Since highly contagious people have no symptoms for the most part, or have not yet shown symptoms, the re-opening of society and control of the epidemic would be made easier by massive testing of the population. Diagnostic tests and the isolation of positive cases could be enough to break the chain of transmission and make control of the virus’ spread possible.
Confidence in testing depends principally on their virus detection sensitivity. The RT-qPCR test is the benchmark for diagnosis, with a detection limit of 103 copies of viral RNA per millilitre (cp/mL). However, RT-qPCR is expensive and results take 24 – 48 hours to come through. Certain tests are quicker and can reduce costs as well as having a waiting time of just a few minutes. This is the case for example with saliva or antigen tests. But these tests have a 100 times lower detection limit (105 cp/mL).
Researchers have investigated the impact of regular testing on the transmissibility of the virus between individuals, by modelling the viral charge of 10 000 people simulated on a database. The percentage of contagious individuals that could be isolated through screening tests was calculated by comparing diagnostic sensitivity tests of 103 and 105 cp/mL and different screening frequencies.
The analysis shows that there is a slight difference in levels of reduction of virus transmission between the 2 types of tests (RT-qPCR and the quicker tests). In addition, the more individuals get tested regularly, the greater is the reduction in rates of virus transmission between individuals. Likewise, it was shown that someone who gets tested infrequently is at greater risk of testing positive. Finally, for viruses with SARS-CoV-2-like infection kinetics, modelling shows that a rapid test is better than a more sensitive but slower test.
This can be represented schematically as follows:
Overall this study shows that large-scale public testing can help to limit the spread of the virus. In addition, these tests allow infected individuals to be isolated at the start of infection, but also enable people in the process of getting better to come out of isolation when their viral charge is too low for the virus to be passed on.
But these analyses have their limits. Firstly, the sensitivity of a test is not limited to the detection of viral load. Poorly carried out sampling may, for example, modify the test’s sensitivity, although certain methods of sampling such as saliva tests are less affected by this problem. In addition, the modelling of participation in testing is statistically independent of the individuals’ backgrounds, whereas we know that behavior differs from one segment of the population to another, in terms of self-protection or propensity to be screened. Likewise, some groups of people refuse to self-isolate after a positive result, thereby adding to the difficulty of controlling the epidemic.
