When I am asked (with two working days’ notice) to implement a large-scale medical testing operation despite having no relevant experience or qualifications, my first thought is not ‘where do we start?’ or ‘how can this be achieved?’, but ‘is this really a good idea?’.
Clearly I don’t have the can-do attitude, or whatever leadership traits I supposedly need, to jump into action immediately on receiving my instruction. Good leaders just get on with it, right? They are doers; grafters; deliverers.
You see, the issue to me is not whether we can do what is being asked, but whether we should. I don’t mean this in terms of if we should question whether this is really the role of schools, whether someone else is better placed to do it, or whether we should refuse until we are properly resourced (although these are all legitimate questions). I mean should we do this in terms of whether it will be, on balance, in the interests of the children and adults in our schools?
If I had an ounce of faith in the current administration, I would perhaps take it on trust that someone had already thought this through carefully – but I don’t. So that leaves me with more work to do. The school leaders who had already written to parents asking for volunteers within twenty four hours of receiving this latest dictat obviously have more faith in government than I do – either that or they were trained in the military and never question a direct command. Or perhaps they already know a great deal about the efficacy of lateral flow tests? I must assume they are either more knowledgeable, trusting, or reckless than me: let’s hope the former.
So here I go… diligence is due. Please be advised that what follows is entirely from the perspective of an educationalist, not a virologist.
I need to structure my thinking about this, so it may help if I am clear about the questions I particularly want answers to. I know the government’s objective is to keep children in school, which is laudable. However, more important than that in my opinion is whether what is being proposed will mean those who attend school, and those they might spread the virus to, are safer as a result of the proposals for widespread use of lateral flow tests (LFTs).
I also want to be confident that the time and effort spent setting up this process could not be better spent on doing something else to keep people safe and to educate students. Therefore, my key questions are as follows:
- Will introducing LFTs (in the way government suggest) reduce incidents of Covid 19 in school?
- To what extent will testing mean that more students and staff are able to attend school?
- Is the time/resource of setting up and running testing at the scale required justified by the benefits (health and educational)?
What is being proposed?
The proposal (option? instruction?) is that secondary schools provide lateral flow tests for all staff weekly, and for all students who have been identified as close contacts of an individual testing positive daily. The premise is that, by doing so, students can continue to attend school when identified as a close contact instead of isolating (and therefore studying remotely), that asymptomatic carriers are identified, and that those who have Covid-19 can be identified more quickly than they would by waiting for symptoms to emerge.
On the face of it, the plan appears to offer some benefits both in terms of containing the spread of the disease and keeping children in school. However, these benefits will only be realised if the tests are reliable in detecting the virus and can detect the presence of the virus in a host before they become symptomatic.
So, what do we know about the efficacy of LFTs?
Government position on use of LFTs in Care Homes
As yet, we have no evidence of the efficacy of LFTs in a school setting. However, the government have published an evidence summary regarding their use in care homes.
The point of using a LFT in a care home is to test visitors. This is advised as an additional protection to PPE, distancing, hygiene, and other precautionary measures. Testing is recommended as a precaution to reduce risk further, but is not ‘a panacea’. The implication is that the purpose of using LFTs is to filter out visitors who may have the virus, but that they are not sufficiently reliable to give great confidence that those testing negative do not carry the virus. In other words, it is not advisable to assume that someone does not have the virus just because they test negative.
This last point is important when it comes to the use of LFTs in schools. It may be advisable to use LFTs to identify potential positive cases (for isolation and the application of a more reliable test) but a single LFT negative test result does not give assurance that someone does not have the virus. We will return to this point later.
Evidence from university settings
The accuracy of LFTs has been tested in three university settings in the UK so far. At this point, I would advise you read this excellent explainer by Dr Rob Hagan, an ex-colleague of mine who, following an exchange on the subject on Twitter, kindly offered to write a short summary (aimed at staff/parents) about LFTs.
Here is an extract from Rob’s paper which points to the evidence of the three trials:
- Testing at the Porton Down by Public Health England together with the University of Oxford reported 76.8% accuracy (76.8% of persons who tested positive with the standard Covid tests in laboratories were measured as positive by the lateral flow test) in the tests when testing symptomatic individuals and higher accuracy for individuals with a high viral load. https://www.ox.ac.uk/sites/files/oxford/media_wysiwyg/UK%20evaluation_PHE%20Porton%20Down%20%20University%20of%20Oxford_final.pdf
- Testing at the University of Liverpool for University students under supervision reported 49% accuracy amongst asymptomatic university students. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/943187/S0925_Innova_Lateral_Flow_SARS-CoV-2_Antigen_test_accuracy.pdf
- Testing at University of Birmingham of asymptomatic individuals under supervision found 3% accuracy. https://www.medrxiv.org/content/10.1101/2020.12.01.20237784v2
Note that the Porton Down study is the one referenced by the government evidence summary for care homes, which happens to be the most optimistic of the three. It is also referenced in the NHS handbook issued to schools on 23 December, although in this document the claim is that this study showed the test had 99.8% “specificity”. This is a highly selective use of data and no explanation is given as to what this means. Out of context, this figure is very misleading.
As Rob points out in his paper, the vast differences between the accuracy rates in the three studies probably comes down to two things: 1) whether those being tested symptomatic or asymptomatic 2) whether the tests carried out by trained personnel or self-administered.
In the Porton Down tests, participants were symptomatic. Remember that the point of testing in schools is to pick up asymptomatic cases and/or those who are yet to become symptomatic. Therefore, the other two accuracy rates would appear to be a better indicator of what we might expect in schools. These were self-administered tests by adults (albeit mostly young adults). We can reasonably expect that tests self-administered by 11-18 year olds would be less reliable.
Given the above, it would be prudent to assume that LFTs will accurately detect less than 50% of asymptomatic students/staff in schools. But were the individuals missed infectious? We do not know. If the viral load was low (which is likely to be a contributory factor to the tests not picking up the virus) then perhaps those carrying the virus were not yet infectious to others, although as this article points out, we need to be careful not to conflate viral load with infectiousness. Or perhaps they had carried the virus for some time (i.e. caught Covid some time ago) and were no longer infectious. In short, we cannot know the extent to which low accuracy is a problem in real terms.
We should also remember that the intention is that students who are identified as close contacts of those testing positive will be tested using a LFT every day for 7 days. As viral load increases, it becomes more likely that a LFT will detect the virus and generate a positive result. Also, if the tests previously ‘misfired’ (i.e. generated a false-negative result even though viral load was high, perhaps because the test was not administered properly), subsequent tests will hopefully generate a positive result. Repeated testing will certainly mitigate the low accuracy of the tests significantly.
None of this would matter much if we were using LFTs to complement all the other precautionary measures, as is done in care homes. Their use would mean that we pick up more carriers of the virus, earlier (if only a few). Then it just becomes a question of whether it is worth all the effort. However, that is not what is proposed. These tests are being recommended (strongly) by government to schools as a way to avoid having to ask close contacts to self isolate. This means that all those who would have isolated for 10 days would now remain in school on the basis that the daily LFT will pick up the virus if they have contracted it.
Can we be confident, given the above, that those the virus has been passed to will test positive before they themselves become infectious? In other words, is the test sensitive enough to detect the presence of the virus before the host becomes contagious to others?
To help answer the question of whether LFTs detect the virus before the host becomes infectious to others, consider the view of Professor Richard Tedder, a Senior Research Investigator in Medical Virology at Imperial College London:
“There are two serious flaws with the use of these assays to screen young children for identifying those infected with the coronavirus. The first relates to the sensitivity of the assay which is going to be many thousandfold less sensitive than the current PCR-based molecular assays. To signal with a LFTAg a person must already be shedding significant amounts of virus. Since these tests are being used in the asymptomatic individual, this essentially means that the person, if infected, is in the incubation phase and that the viral load will continue to increase with time even if the test gives a negative result. This means that in principle testing needs to be repeated in a short period of time implying the need for daily testing. This is important conceptually because the whole object of having these tests can only tell you at the time the sample was taken whether the individual is shedding virus or not and does not inform about whether they will be shedding one or two days in the future. To use a test of minimal sensitivity such as LFTag to identify a person in the early phase of the infection is inviting a false sense of security in people who are told “your test is negative.”
Again, I am no scientist, but my reading of this is that an individual must already be shedding the virus (i.e. infectious to others) before a LFT will give a positive test. If, as suggested, students come into school and are tested when they arrive, those that test positive (if this is not a false-positive result) will have been infectious on their way to school at the very least, and possibility for a longer period since the last LFT was done. Identifying an infectious individual soon after they become infectious is certainly better than waiting for them to become symptomatic to do so, but the use of LFTs will not, as far as we know, stop infectious students from being in circulation. This approach must therefore be more risky than an enforced isolation for close contacts.
What are we to make of this?
We do not know at this time whether LFTs are accurate enough to fulfil the purpose of controlling the spread of Covid-19 in schools by identifying those with the virus sufficiently early to stop them spreading it to others. It seems likely that the approach being suggested will be less effective than an isolation programme, but we do not know enough to judge whether this loss of efficacy is ‘worth it’ in terms of the benefits of being able to keep students in school.
At this point, I will move away from the science and consider the risk management and organisational issues that arise given the little that we do know about the efficacy of mass testing.
Firstly, it is worth explaining what happens at the moment if we are informed of a positive test result for a student or member of staff. In this event, we trace close contacts for this individual and ask these contacts to isolate for 10 days (previously 14). In reality, no-one isolates for 10 days. For example, if a student becomes symptomatic on a Wednesday evening, they may get tested on Thursday and, at best, get the result on Friday. We contact-trace and ask close contacts to isolate for 10 days since the last contact they had with the individual, which could be from the Wednesday, or Tuesday if they had no contact with them since then. It is therefore likely that at least three or four days have passed before close contacts are asked to isolate. Our experience is that we are often on day five before close contacts are identified and informed. This isn’t a significant problem as it takes at least a few days for the virus to incubate, so the close contacts are unlikely to be infectious in the first 3 days even if they have caught the virus.
The result of the above is that many close contacts will isolate for no more than one week, and miss no more than five days of school.
When there are multiple positive cases, the contact tracing becomes very time consuming and therefore there is a delay in notifying close contacts. In this event, it is safest to ask possible contacts (e.g. a whole year group) to stay at home whilst the contact tracing is completed. For those who would urge schools to get on with it quicker, I would remind you that schools have received no additional funding or resource to perform this task.
It should be noted that contact tracing is not a precise science, therefore it is likely that individuals remain in school who may have caught the virus. It is a numbers game – you are trying to identify most of the people who may have caught Covid. Given the virus will only infect a small number of these, if any, you bank on those getting it being in the isolating group.
How does mass testing change the above?
The answer is that almost all of the above still needs to happen. The close contacts must still be identified, but instead of asking these to isolate, the school will then need to test all of these every day for 7 days (with this being self-administered at home over the weekend). One assumes that 7 days has been chosen as, given the time lag of receiving a positive test and identifying close contacts, this period would take you at least to the 10 days of isolation, after which the host is less likely to be infectious to others.
So, if the virus has passed to one or more of the close contacts of the original case, these individuals will still come to school. If these individuals then pass the virus on to more than one other person then the rate of infection will increase exponentially. If the virus is passed on to only one other person then the chain of transmission continues, but does not accelerate. If these individuals pass the virus to no-one, then the spread stops and the testing has been as effective as isolation.
Which of the above scenarios is likely to play out? We don’t know because we do not know enough about how accurate the tests are generally, we know nothing about their use in schools, and we do not know enough about transmission rates in schools (the ‘R’ value).
We also do not know how people will behave when they receive a ‘negative’ test result or as a consequence of the false security of rolling out mass testing. If those in schools relax other precautionary measures and have a false sense of security, that could increase transmission in the school environment. How testing is portrayed, and how negative test results are explained, will be very important. The government propaganda which over-sells the accuracy of these tests may end up being counter-productive.
When faced with this much uncertainty and potential risk, my instinct is to proceed very cautiously.
But students will be in school!
Whilst we know little about whether mass testing in schools will contain the spread of Covid-19, we do understand the difficulties of students and staff having to isolate. For students, even five days out of school can be extremely disruptive to their learning. And for those in school, having cover teachers because their regular teachers are isolating is also problematic. There are significant gains to be had if we can avoid having to ask students and staff to isolate.
It is possible (I am not saying likely) that the use of LFTs in the way intended by government will lead to more time out of school for students and staff if their inaccuracy means that there are more outbreaks in schools. However, it is also possible that testing will be successful at containing the virus and significantly reduce disruption to learning. We do not know which way this will swing.
Are we any closer to answering the key questions?
Let’s return to the questions I wanted answers to:
- Will introducing LFTs (in the way government suggest) reduce incidents of Covid 19 in school?
We do not know. Based on the evidence we have, I am not even prepared to make a guess.
2. To what extent will testing mean that more students and staff are able to attend school?
We don’t know. What we can say is that if the tests are successful at keeping the R value at 1 or less then there are significant learning gains to be had. However, if they do not, then we could see more disruptive outbreaks. We will not know the answer to this until we run some trials, or just launch this nationwide and see what happens!
Effort to reward ratio
Is all this worth the effort? Don’t get me wrong: if we can do something to reduce the spread of Covid-19 or to keep students in school more then we will, given the expertise and resource to do so. But is this the best way of achieving these goals? And is what is being asked even possible?
I do not want to enter into detail about the logistics of delivering what is being asked (and I am sure you don’t want me to either), but having digested the information given to schools so far I make the following observations:
- I am not sure that what is being asked is logistically possible for my school, certainly not by next week. We do not have the space, manpower, or expertise.
- The diversion of school resources will significantly undermine other aspects of the school’s work e.g. quality of provision, supervision of students.
- The funding provided is unlikely to be sufficient to cover costs.
There also appear to be some better options on the table. Keeping exam years in school and moving to full remote learning for other year groups for January, whilst the vaccine rollout gains momentum and infection rates in the community come back under control, would allow schools to focus their efforts entirely on providing education (which is what they know most about, after all). This option would also enable the use of school facilities for vaccinations for local communities and allow more planning time if testing is to go ahead.
What we know is that what schools are being asked to do is not within their experience or expertise, will damage the quality of education, and will possibly push some headteachers over the edge. Given we know so little about how effective mass testing in schools will be in controlling the virus and keeping students in school, this seems like a big leap of faith.
So to my third question:
3. Is the time/resource of setting up and running testing at the scale required justified by the benefits (health and educational)?
Again, I don’t know, but my gut says ‘no’. There may be a role for LFTs in schools, but perhaps not as currently intended.
During my research into LFTs I have steered towards the most reputable sources and qualified opinions. I am afraid that I have only found words of caution and dissenting voices about this plan. Here is what Dr Alexander Edwards, Associate Professor in Biomedical Technology at the University of Reading, has to say:
“These same tests are supposed to be used in schools in January to help our children return to school safely – yet the emerging information from Birmingham suggests they may offer false reassurance. It’s very hard to understand how all these new testing sites can be checked to ensure that they are accurate enough. It’s hard enough for schools to teach in current environment – we shouldn’t be expecting them to setup diagnostic testing services on top, without being sure that the tests used are accurate enough.”
And from a recent editorial in the British Medical Journal:
“Whatever decision making process the UK government used, it ignored key evidence and dismissed expert international advice. The result is a considerable burden on care home staff, universities, NHS staff, public health teams, and schools, with minimal additional safety compared with existing risk mitigation measures. Asymptomatic lateral flow testing is an unhelpful diversion from the important task of vaccination rollout.”
And a quote from John Deeks in New Scientist who was involved in the Birmingham screening programme:
“Mass screening using lateral flow tests is set to be introduced among all secondary school students in England, starting in January. ‘They certainly should not be used like this,’ says Deeks.“
Whilst I am conscious that dissent plays out better in the media than endorsement of government policy, these are reputable sources, not the gutter press. Who are we to trust?
What to do?
Every secondary school head teacher across England will be grappling with this issue right now, and each must make their own decision. I am jealous of those who have just run with this – life would be easier if I could shed my caution and doubt. I’m holding out for a u-turn, or a decision which will take the dilemma out of my hands, but I fear that the government will blindly and belligerently stick to the plan.
So it is time to be pragmatic. If schools don’t deliver mass testing we will once again be slagged off by the media. If we do deliver it, government will take the credit. These are things I cannot change, but I can try to make the best out of yet another fine mess.
There will be a middle ground, a compromise, which will help steer a course through this uncertainty. I just haven’t found it yet. But then, what are school holidays for?
NHS Schools and Colleges Handbook
Government Evidence Summary for Lateral Flow Devices
Science Media Centre: responses to Professor Deeks’ Twitter comments
2 thoughts on “Lateral Flow Tests in Schools”
Will introducing LFTs (in the way government suggest) reduce incidents of Covid 19 in school?
Assuming that you test the people that you would previously isolate, I cannot see a mechanism by which incidence of Covid 19 would be reduced. At best it would be the same, in the unlikely event that all secondary cases are caught before the virus is transmitted further. It must be more risky, as you said, “This approach must therefore be more risky than an enforced isolation for close contacts.” The risk is entirely on the downside.
It also seems particularly poor timing to attempt the experiment when we have a dramatically more infectious strain circulating. The prospect of it all going very badly wrong is all too plausible.
I like your analysis of considering the R value in school. It seems to me that a much better use of LFTs would be to do additional random testing to seek out and isolate “silent carriers”. This might initially result in more isolating of these and their contacts, but if it reduces R, there could be a subsequent reduction in the prevalence of the virus, increasing safety for students, staff and the community and improving attendance.
We cannot defeat this virus by pretending that it isn’t there – that leads to exponential growth – we have to try to get ahead of it. Mass testing with LFTs in addition to the traditional isolation of contacts is another tool to reducing R. Using LFTs as a substitution for isolation cannot reduce R, and may very well increase it, at a time when it may increase anyway due to the new variant.
In short, LFTs are not fit for the purpose of indicating that a contact is negative.
I’m not medically trained either. I teach a STEM subject.
I think you are right about a better use of tests being random sampling to identify asymptomatic cases. Thank you for your thoughtful response.