Covid-19: What you need to know today
India is now conducting in excess of 400,000 tests a day on average (the number tends to drop off over the weekend, but has been rising consistently). That’s still short of the million tests-a-day number that has become this column’s most popular refrain, but is way higher than what the country used to do even a month ago.
The recent surge in testing is being helped by the increase in antigen tests. These are also called Rapid Antigen Tests because the results are usually available within the hour — unlike the molecular tests which take longer, a couple of days typically, but with backlogs up to five days sometimes. If it is any consolation, US media is full of horror stories of how some people have been waiting up to 10-12 days for results as a wave of infections sweeps across that country. Both the Reverse Transcription Polymerase Chain Reaction (RT-PCR) test and the Nucleic Acid Amplification Test (NAAT) are examples of molecular tests.
Such molecular tests are highly accurate in identifying current infections of the coronavirus disease. They are so accurate that they do not need to be confirmed using any other test. They are poor tools to help measure prevalence of the disease in the population, though, because they cannot identify past infections — and often, both public health and local administrative authorities need to know just how much of a population has been infected because this will influence other decisions, such as what to open up and when.
There are some molecular tests that take much less time than the RT-PCR and NAAT ones but these are not in wide use.
Antigen tests are in the news in India. They helped Delhi scale up testing rapidly, and with the number of daily new cases in the Capital declining sharply, health administrators want other states to start using them as well. Unlike RT-PCR and NAAT tests, which search for the virus’ genetic material, antigen tests look for proteins that are usually found on the surface of the virus. Like molecular tests, they are good indicators of current infections (and like them, they cannot measure past infections).
There is a problem with such tests though — they are very accurate when it comes to what are called true positives, which means that if they show people to be infected, the subjects most probably are, but they are not so accurate when it comes to so-called true negatives, which means that at least some of the people they show to be uninfected are, in fact, infected. That’s the reason scientists recommend following up negative antigen tests with molecular tests, especially if the subjects being tested are symptomatic (but this does not account for the large proportion of people who remain asymptomatic despite being infected).
It’s important to understand that this does not mean antigen tests should not be used. There are times when it makes sense to use them — when infections are peaking, for instance, especially in one neighbourhood and the primary need is rapid testing — and how (preferably with a follow-up RT-PCR test for all those testing negative). In effect, they should be used to address the time constraint, not the testing-capacity one. They are also not good measures of whether it is safe for a person to return to work. I mention this specifically because at least some private testers are suggesting that they can be used for this purpose.
In theory (and in theory alone), the only tests that can be used to define an individual’s eligibility to return to work are those that test for antibodies. Unlike the molecular and antigen tests that use throat swabs, these use blood. In general, they are faster than antigen tests, although a review of current research literature shows that their accuracy varies widely. They do not identify current infections but are good measures of past ones — and therefore, ideal to measure prevalence (or immunity). The results of such studies are open to interpretation, though. Our understanding of the coronavirus disease, and the strength and duration of the immunity provided by antibodies, is still evolving.