select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to tweets from Prof Jon Deeks about lateral flow test accuracy at the University of Birmingham

Professor Jon Deeks has raised concern on twitter about the sensitivity and specificity of lateral flow tests carried out at the University of Birmingham and Scottish universities.


Please note the comments are based only on the information available in the tweets – there is no preprint or paper available:


Prof Richard Tedder, Senior Research Investigator in Medical Virology, Imperial College London, said:

“Lateral flow tests, essentially point-of-care assays for the detection of viral antigen present in the nasal passages of individuals infected with SARS CoV 2 (LFTAg), are obviously attractive for high throughput testing in at risk populations where social distancing and other non-pharmaceutical measures for preventing transmission may not be easily applicable or likely to fail.  Such a case describes a government intention to have schools open.

“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”.

“The second major flaw in this type of assay is that a negative result is by definition invalid.  To use a test which is going to inform an individual it is essential to know that the sample taken was adequate.  For this reason molecular tests should always have a method of detecting and confirming the presence of human material in the swab sample, this is termed the internal control.  A negative PCR with an appropriate signal in the internal control means that in spite of an adequate sample no viral signal was detected, therefore one can say to the person you are not shedding virus.  In such a situation the negative report is valid.  This is simply not the case with an LFTAg at present because such tests do not have an internal control and therefore cannot guarantee that a negative result has been given by an adequate sample, and for that reason a negative assay generated by a lateral flow test cannot be considered as valid.  This is one of the biggest criticisms of the use of such assays.

“If these assays are being used only to screen to tell somebody that they are “virus-positive” in principle for infection control, this may be useful.  Inevitably there will be false-positive reactions but these will not impact adversely upon the potential control of infection.  However it is absolutely essential that people are not told “you are negative” because the negative result is not valid and can simply be generated by an inadequate sample.  At best these assays will detect people shedding large amounts of virus but this depends entirely on the quality of the sampling and the accuracy with which the tests are used.”


Prof Gary McLean, Professor in Molecular Immunology, London Metropolitan University, said:

“Based on the information in the tweets and our own limited analysis of test results obtained at our University testing site, these Innova rapid tests appear to miss some positives – it could be due to self sampling technique rather than inability of the test itself although low levels of virus could be more difficult to detect by this method.

“We performed these at LMU earlier in December and found a very low positive rate, similar to other universities, but we did not perform confirmatory tests (that is up to the Govt).

“My feeling is these tests are of value but cannot be solely relied upon.  Remember they are designed for asymptomatic people testing only and therefore it is not surprising the positive rate is low.”


Dr Alexander Edwards, Associate Professor in Biomedical Technology, Reading School of Pharmacy, University of Reading, said:

“Commenting on the information that is available in the tweets:

We already knew from studies such as the Liverpool mass-testing evaluation report1 that lateral flow antigen tests to detect virus from swabs performed less well than PHE evaluation of Innova tests.  Unfortunately, most people rolling out rapid asymptomatic testing services simply don’t have the capacity to run in-service evaluations and so for many community testing services using this product – as recommended by the government2 – and so we just don’t know how well these perform in other sites.  That we don’t know how well these perform in other sites is worrying for two reasons: Firstly did every university student returning home after a negative test result, really understand that the low sensitivity of the test could not ensure they are ‘safe’ to return to their family?  Secondly, does wider use of less accurate testing undermine track, trace and isolate?

“There are two well-known weaknesses of this kind of lateral flow test.  Firstly, they are best at detecting really abundant targets, where the marker being measured is present at high levels.  When the level of virus is lower – which can often happen with swabs – they aren’t as effective.  Secondly, because they are portable and rapid, they are suited to community testing.  But community testing comes with an additional burden – the people operating the tests will be less experienced at diagnostic testing than technical staff in an approved laboratory, and the test results are recorded by eye.  Many past studies using the same lateral flow technology read by eye have shown that visually recording test results can show variability, especially with a weak test line.  For many mass-testing services rolling out, there is no way the service can be properly checked for accuracy.

“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.

“We are stuck with an urgent need to spot cases in the community to reduce spread, but still without a perfect tool for mass testing.  On paper lateral flow antigen tests have great potential.  The initial PHE trails indicated Innova tests may be the best compromise, but new data is emerging that suggests either the product accuracy has dropped further or that wider roll-out has revealed it’s harder to use than anticipated.”






(There is no preprint or paper.)



All our previous output on this subject can be seen at this weblink:



Declared interests

Dr Alexander Edwards: “I am involved in developing higher sensitivity rapid portable testing technology, and am shareholder in a diagnostic testing company, but this company does not currently offer any COVID-19 test products.”

None others received.

in this section

filter RoundUps by year

search by tag