Dr. Michael Mina:

I have spent the last year or so studying or developing the science on rapid testing, particularly rapid antigen testing. At the beginning of the pandemic, people had not asked how to use testing to really limit spread. That has led to a lot of confusion. I have seen it. There is a lot of confusion in Ireland but it is not just in Ireland; it is also in the United States and every other country. I have been advising governments across the world on this issue.

The biggest issue that I can see people are continuing to have regarding antigen tests concerns the questions of whether they are accurate and how to understand the likes of the Cochrane review, which showed a sensitivity rate in the order of 50%. That needs to be placed in the context of the goal of testing. This is where the biggest questions have arisen. Testing is almost always used in the context of medical practice. That is how we have devised all the metrics we use to evaluate a test but, in the middle of a pandemic, the question of how to use a test is much more about how well the test is limiting the spread and detecting people who are actually infectious. If that is our question, the sensitivity metrics of the rapid antigen tests change dramatically. We have published on this in Science, The New England Journal of Medicineand many more of the world's top medical journals. All of the conclusions are essentially the same in that where there is an attempt to limit spread, the speed at which a result comes back is much more critical than getting every ounce of sensitivity to detect the RNA of the virus.

That is because the PCR can detect down to 100 or 1,000 copies, but people do not spread when they are at 100 or 1,000 copies of this virus. People spread when they are at a million, billion or even a trillion copies of this virus. If the question is whether we can detect a billion copies of the virus, these tests do it very well, down to about 100,000. That is sufficient to detect anyone likely to be infectious.

The other downfall, if we compare these tests to PCR, is speed. What we have to do when evaluating the sensitivity of a comparator test, like a PCR test, and which is not currently done because we do not even have the metrics to do it, is to question the effective sensitivity of the PCR. A test taken two days before flying, for example, is a much poorer performing test than a rapid antigen test taken an hour before flying. People are often worried about missing the first day after somebody is exposed, for example, and whether the antigen test would pick it up, but testing two days before flying means giving an inferior test for that flight versus a rapid antigen test one hour before the flight. That is because if the real concern is transmission on the flight, or thereafter, we want to be able to test as close to the event as possible. That will always be the better solution.

We have shown this innumerable times now, mathematically, and it is really not a question anymore of the mathematics or science. It is now a question of how to create new language. That is really what the world is confronting with testing in this pandemic. We have never had the language before even to describe a test whose primary mode or purpose of use is the obstruction of transmission. It is a very different set of metrics and uses compared with what we normally see.

False positives have also been a major concern for many people. However, with many of the newer rapid tests and some of the older ones, such as Nova and Abbott, both of which we evaluated, for example, we are seeing about one in 10,000 false positives. It is very small number of false positives. If a test is positive, the nice thing is a second rapid test, or even a second rapid molecular RNA test, can confirm right then. The rapid tests can be used for their greatest ability, which is that they are fast, so it does not mean a ten-day purgatory from a false positive. It means a 20-minute purgatory and then the test is confirmed. Again, those instances are rare if the right tests are chosen and selected.

I could go into as much detail as anyone wants, but I really want the take-home message to be that these tests are very accurate if you are concerned about limiting transmission. If you are trying to diagnose somebody who has symptoms because of Covid, use a PCR test. However, if the goal is trying to use a test as a way to limit spread, then use a rapid test, whether it is an antigen or rapid molecular test, both of which will be available in future. Speed is much more important if your goal is to limit spread, especially on something like an aeroplane or at an event etc. I will stop there but am very happy to take questions.