Lateral flow tests: issues with mass testing

How accurate are they compared to PCR tests? And is this strategy likely to help prevent transmission?
13 April 2021

Interview with 

Jon Deeks, University of Birmingham


a picture of a medical swab


From the 9th April, people across the UK can ask to receive 2 Covid-19 lateral flow antigen tests every week. This is part of a £100 billion government initiative codenamed “Operation Moonshot”. But is this idea of conducting mass testing on the population to drive down Covid-19 cases and suppress the virus to get us back to normality based on a flawed test and flawed logic because it will miss half of cases? Jon Deeks is a biostatistician at the University of Birmingham, and he spoke to Chris Smith...

Jon - This test is very good at picking people up when they have high levels of virus in their systems. In the study that the manufacturers did, it got 96% of those people. But it doesn't work very well when people have low levels of virus, which is what everybody has when they first get infected, and some people never get above low levels of virus. So it'll miss people, and particularly in those first stages of infection. It'll also miss them if you do a less than successful swab.

Chris - So, in real-world circumstances, if we take a person off the streets and just test them, how good is it picking up a case of coronavirus infection?

Jon - The data we've got from this are all from Liverpool. They tested nearly 6,000 people. 70 of them actually came up positive on a PCR, which they did in the sample at the same time. And of those 70, only 28 were positive on lateral flow test. So that's 40%. So it missed six out of 10. Now we have to be a little careful because the PCR tests stays positive for days - if not weeks - after you've had coronavirus because it can't tell the difference between live virus and dead virus. And so they actually also produced analysis done in those who had the higher level of virus. They actually missed a third in the 39 cases in that sample, which had the high level of virus. I think one troubling thing is that actually these numbers of 70 and 39, this is the basis for our national testing policy; it’s a little bit shocking that we have only got this amount of evidence actually to tell us in the UK as to how well this test works.

Chris - Yes, they're not big, robust statistically large numbers are they! But in terms of what you've just said then, maybe as many as half of cases will be missed when people do lateral flow tests?

Jon - It seems to be about that amount. Whether those cases are infectious is part of the discussion because if they are dead virus, then it doesn't matter we miss them. But the concern is that some of those people will be infectious, they will get a negative result, and they may be falsely reassured by getting that negative result. They may be disinhibited from continuing with all of the social distancing, mask wearing, not cuddling your granny that we've all been told to abide by. Because they've got a negative test result, most people think, well, that means I haven't got it, and take a negative test result as a green light to go and start behaving in ways which they could spread the virus unfortunately.

Chris - So if we look at ways in which - let's take the UK example, the government here are seeking to deploy this test. They're being given to school children so that they will test themselves a couple of times a week. From the 9th of April, people will be able to access these tests so that they can, if they're an adult, test themselves. Is this a valid strategy or has it got flaws?

Jon - Well, there are two missing pieces of evidence straight away. One is we have no evidence as how well people will do this test when they do it on themselves. The other issue is actually we've got no evidence to say how well does it work if we keep on doing this test every Monday and Thursday, for example, the twice a week thing. So it doesn't seem right to me on any level that we're implementing a national policy when we don't actually have the evidence as to how well it's going to work.

Chris - We've considered whether or not the test can find positives in cases where the virus really is there, but what about in cases where the virus isn't there, but the test nevertheless still says positive, it gives a so-called false positive. How often does that happen?

Jon - The data that we've seen on this is showing that it could be 1 in 1000 or it could be even as low as 3 in 10,000 false positives will occur. The problem comes when we actually are using these tests in groups who are very unlikely to have disease because there'll be very few coming up with true positives. And so the ratio of those which are true to those which will be false becomes unfavourable. We saw this happen recently in the schools in the UK where the rates of test positivity was so low it was very clear that it must be that most of these positives are actually going to be false. Now there's a reasonably easy way of mitigating this risk and that is to test people with a PCR test if they've tested positive on the lateral flow test.

Chris - People are also talking about - and the Netherlands have explored the possibility of using testing in this way - for things like festivals or mass gatherings.

Jon - This is where the failure of these tests to detect cases is the big problem. Because if you started testing everybody going into a venue, there'll be some false positives. Not many, but those people will be wrongly turned away. But the bigger risk is actually that we'll miss half the cases as we started discussing. And so the venue will have, rather than having 20 people with COVID in it, it'll have about 10 people with COVID in it. So it's not making a venue safe. It's reducing the risk. It doesn't seem to be reducing it that much, in my opinion, to half the number of cases. However, we need to think that these are the first generation of these tests and lateral flow tests are available for other diseases, which work far better than the ones we've currently got for COVID. So there's hope, but I think we need to be very careful and do very good studies to actually understand whether this is safe or not.


Add a comment