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

expert comment on different types of testing for COVID-19

Comments in answer to some questions from journalist about the different types of testing for COVID-19.


Comments sent out on Tuesday 31 March 2020

Dr Colin Butter, Associate Professor and Programme Leader in Bioveterinary Science, University of Lincoln, said:

“These notes below relate to testing for the virus.  The point of care tests for antibody responses are apparently on the way and will be very welcome.

“Why so slow?:

“The technique presently being used to test for the presence of virus is quantitative Reverse Transcription Polymerase Chain Reaction, or qRT-PCR.  Strictly speaking it does not detect antigen but viral RNA.  Reagents for the assay were very quick to produce once the viral RNA sequence was published by the Chinese.  Thousands of machines capable of doing the assay exist in the UK and a specific subset of these has now been requisitioned by PHE to be stationed at (I think) three centres including Milton Keynes.  To my knowledge there were 25 machines recruited from the Nottingham area alone so PHE must have, or have access to, hundreds.  Given the samples to run and the reagents to do this, each machine can do well over 1,000 tests per day, working 24h.  Each sample needs to be tested in 3 different ways so say 330 samples per day.  100 machines x 330 = 33,000 so why are so few being achieved?  Is this the processing of samples or the logistics of moving them about?  Where is the bottleneck?

“The PCR test was the correct first step as it was quick to produce, albeit relatively slow and messy, including a sample extraction stage, and depending on specialist machinery.  A much better and faster test would be one that detects a viral protein.  This depends on making a monoclonal antibody against the relevant protein, something that takes around two to three months.  Once these are available they can be made into a lateral flow device, like the antibody test.  This would operate as point of care, requiring no sample extraction or specialist machinery, with results available in 15 minutes.  What’s happing with developing these?  Which companies are working on them and do we have options to buy?”


Prof Ashley Woodcock, Associate Dean for Clinical Affairs and Professor of Respiratory Medicine, University of Manchester, and Clinical Director for Respiratory Medicine, University Hospital of South Manchester:

Tests for the virus:

“All these tests depend on upper respiratory samples.  They are crucially dependent on the quality of samples.  Sputum is better than oropharynx which is better than nasopharynx.

“Sampling itself is hazardous.  People sampling have described having sputum on their arms.  They need full PPE.

“The UK has been using a central laboratory setup of PCR machines for virus tests and is building greater capacity.  In spite of this testing type being the ‘gold standard’, there are a number of issues:

“• Capacity: turnaround has been too long – in many places 24-48 hours (but this is getting shorter).  Patients can still be sat on wards without a firm position on their infectivity for 24 hours.

“• Positive rates for patients with COVID syndrome are ~70%.  This means we have a lot of patients waiting for a repeat virus testing/diagnosis.

“• There are many false negatives (maybe 30%) meaning potentially infectious patients could be nursed in COVID negative areas, or sent home thinking they are non-infectious.  Reasons for false negatives include low levels of the virus in the early stages of the disease, and especially poor technique in sampling or conducting the test.

“Other options for testing include testing at Point of Care (POC) using a stick test with the same throat swab samples.  These tests have not yet been validated against the central laboratory PCR, but tests are beginning to arrive.

“A third option is conducting tests at POC with the same throat swab samples but using local PCR machines, which would ideally be located in emergency departments of hospitals.  This should give an answer in five to thirteen minutes.  However, local PCR machines are not yet widely available.

“In conclusion, central PCR testing will continue to be needed, particularly due to the volume of tests required.  However, within two to four weeks, patients being admitted to hospital will hopefully be diagnosed as virus positive at the first POC within 15 minutes at each large site. The best methods for this do still need validating.

Serology tests for immunity:

“All serology tests will likely need to be corroborated by virus diagnosis.  There will be central ELISA tests in due course, which should have a rapid turnaround and be in bulk and cheap.

“Currently the ‘stick’ test will be invaluable as POC diagnosis (e.g. in primary care).  They are becoming available in very large numbers (millions).  The UK Government has been trying to source millions of these tests.  They are mainly coming in from China, and a British test should be out in two weeks.

“Individual brands will need validation and probably need to be quality controlled for each batch.  There will be others; this is not complicated.

“These tests are cheap, about £5 per test.  They are a stick test like a pregnancy test.  It is easy to use with two spots of blood from a thumb prick, and takes 10 minutes for a positive answer.  There is no infection risk to sampling over and above that of a finger prick.

“There are two immunoglobulins that they detect, IgM and IgG.  Most sticks have both IgG and IgM.  IgM goes up quickly and is reputed to be 90-95% positive on admission.  IgG goes more slowly and reflects long term immunity.

“There are three immediate uses:

“• Point of care diagnostic: done by a doctor or nurse in the emergency department.  If positive, then the patient can immediately be triaged.  False positives are rare.  There may be 5% false negatives (IgM may not have yet gone up), but this can be checked in patients with the syndrome by re-testing the next day and corroborated by throat swab/virus test.  This is more accurate than current virus PCR tests.

“• Tests for staff going off work with symptoms to confirm COVID (along with taking throat swabs for PCR testing).

“• Tests for staff returning to work to confirm COVID immunity (these staff will be invaluable, but we need to be sure staff had COVID and not a seasonal flu).”


Comments sent out on Monday 30 March 2020

Dr James Gill, Locum GP & Honorary Clinical Lecturer, Warwick Medical School, said:

“Currently there are two tests in use, a PCR lab test looking directly for the virus, and the newer antibody test which is looking for evidence that the body has been exposed and reacted to the virus.

“PCR testing – as used by the CDC and WHO initially – is very labour intensive, and has several points along the path of doing a single test where errors may occur – which may lead to headline issues of a false positive, the test showing evidence of the virus when it’s not actually there, or a false negative, suggesting someone doesn’t have the virus when in fact they do.

“The false negative is the more concerning of the two outcomes, as they can occur up to 30% of the time. As such, these tests have more often been used to confirm an infection, rather than to give someone the all clear.

“This is because the PCR test is a much more hands on laboratory intensive process both in terms of collection and sample processing. As a result there have been times when shortages of both swabs to take samples from patients, and also the reagent chemicals allowing the testing to be performed, have limited speed.

“During the course of the outbreak, the PCR testing has been refined from the initial testing procedures and with the addition of greater automation to reduce errors. As such, we now have an 80-85% specificity – i.e. the chance the test is detecting the virus. Remember as we are looking at swabs taken from people, who have lots of other organisms floating around, we are essentially dealing with the question of how “right” the result we are looking at is.

“The newer test coming on line now is the antibody test. This has simply taken time to become available, as it has had to be designed and manufactured from scratch.

“With PCR you are essentially looking for RNA evidence of the virus. With the antibody test, this is much more sensitive, but in the early stages of the outbreak it has been a challenge to create the antibody test in the first place.

“For the antibody test, you’ve had to physically get patient blood, then find the antibodies in that blood – molecules smaller than the virus – create a reproducible antibody in the lab, and THEN create a reagent test which will react in the presence of that antibody which can be read by health teams.

“As a result the antibody test is far superior, but has meant creating something from scratch whereas the PCR test is a small addition to an existing test

“There are also further improvements in the pipeline, such as an approval pending for the FDA for a PCR based test which may also provide a bedside test, similar to the antibody testing.

“In terms of overall speed, it’s related to the scale of the outbreak – a lot of hospitals actually have a significant proportion of their complex testing work – such as virology – pooled at a main hospital with more advanced labs. As a result, in the initial outbreak, Public Health England took direct control in order to allow the scaling of lab work at local centres, which required validation of their work. This has obviously taken time.

“Initially all testing was performed at Colindale, so a simple logistical complication is added to each test from around the country. This workload became excessive, and testing capacity was at maximum until additional labs came online, sharing the workload as a collective to ensure that UK diagnostic services were able to withstand the strain.

“The inverse is true for antibody testing, whereby the test needed to be devised, then manufactured, before being able to be shipped out en-masse.

“As testing capability has increased, so has the accuracy of the UK case load data, allowing more accurate forecasting and planning. This in turn allows for improved reaction and deployment of resources in an attempt to support the NHS and reduce the likelihood of the service being overwhelmed.”


Prof Eleanor Riley, Professor of Immunology and Infectious Disease, University of Edinburgh, said:

“Traditionally there are two types of diagnostic test for infectious organisms – tests for the presence of the virus itself (current infection) and tests for antibodies to the virus (prior infection).

“Tests for the organism (current infection) are often called “antigen” tests – where antigen refers to some component of the virus, typically the external (coat) protein of the virus. However, the test being used for COVID-19 is actually looking for viral RNA (which is technically not a viral antigen). So when the CMO talks of “antigen” tests and the CSO talks of tests for viral RNA or “PCR tests” they are actually talking about the same thing. I would prefer if the CMO would stop talking about antigen tests and if everyone would simply talk about tests for the presence of the virus – which covers all eventualities.

“The confusion continues in that viral antigens (proteins) are the basis for developing the tests for anti-viral antibodies. These assays use viral proteins as “glue” to trap the antibodies present in serum (a bit like Velcro where the viral protein/antigen is stuck onto a solid surface and the antibodies in the test serum sample then stick to it).”


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


Declared interests

None received.

in this section

filter RoundUps by year

search by tag