In the largest study of its kind, Imperial researchers have assessed the performance of a number of different home coronavirus antibody testing kits.
As part of the government-funded REACT (REal Time Assessment of Community Transmission) programme, which is tracking past and present COVID-19 cases across England, scientists have evaluated 11 antibody finger-prick tests for accuracy and how easy they are to use at home.
The testing kits, called ‘Lateral Flow Tests’, were analysed using samples from more than 800 individuals and compared with gold-standard laboratory testing, including two in-house tests developed at Imperial College London.
They found that the best performing tests are suitable for large-scale surveillance studies to monitor the progress of the pandemic. These could correctly identify individuals with coronavirus antibodies over 80% of the time, while also correctly ruling out those who don’t in more than 98% of tested individuals.
"We show here that some finger-prick tests are suitable for large scale testing in surveillance studies, which will help us understand the spread of the virus in the country and who has been at greatest risk." Prof Graham Cooke NIHR Research Professor of Infectious Disease, Imperial College London
Based on these findings, the programme has rolled out one antibody test to more than 100,000 individuals (see study results here) as part of ongoing efforts to understand the number of people in England who have had COVID-19, and how this varies by geography and demographics.
This knowledge will be vital to identify those at greater risk while also helping guide local public health responses and informing actions to control the disease.
The REACT programme, which has been commissioned by the Department of Health and Social Care, is being carried out in partnership between Imperial College London, Imperial College Healthcare NHS Trust, Chelsea and Westminster NHS Foundation Trust and Ipsos MORI. The latest findings have been published, after peer review, in Thorax.
Professor Graham Cooke, NIHR Research Professor of Infectious Diseases at Imperial, said: “There are still many unknowns with this novel virus, including the extent to which the presence of antibodies offers protection against future infections. We show here that some finger-prick tests are suitable for large scale testing in surveillance studies, which will help us understand the spread of the virus in the country and who has been at greatest risk. These data will have important implications as decisions to ease lockdown restrictions are made.”
Antibodies: signs of past infection
Testing for the presence of antibodies to the coronavirus can indicate that a person has previously had COVID-19 and recovered, and is believed to be the best way to assess past exposure to the virus. Laboratory testing with a technique called ELISA is the gold standard method to detect antibodies, but currently this is not possible to do at the scale needed to understand country-wide patterns of infection.
Some 200 antibody self-testing kits have been developed to enable community testing en masse, but there have been concerns over their accuracy and whether they are usable by members of the public.
The research involved almost 300 healthy volunteers, all NHS staff or close family members, who had a previous diagnosis of COVID-19 confirmed by lab tests. None of them had serious symptoms at the time of infection, nor had they been hospitalised due to the disease, and it had been at least 21 days since the onset of symptoms. Analysing test performance on individuals who have had mild cases is essential to understand their potential for widespread use in the community.
The volunteers first carried out the tests themselves using instructions provided by each manufacturer, followed by a trained technician. Blood samples were also taken for comparison with gold-standard ELISA testing and another similar laboratory technique called RBD-DABA. This enabled the researchers to work out how ‘sensitive’ the tests are, meaning how well they correctly identify individuals who do have coronavirus antibodies – so-called ‘true positives’.
In addition, the researchers evaluated LFT performance on a batch of 500 blood samples taken before the pandemic hit. These were used to calculate test ‘specificity’, how well they can correctly rule out those who don’t have antibodies. As all of the samples were negative, any that came out positive were therefore a ‘false positive’.
Measuring test sensitivity and specificity
Test specificity was found to be high; all LFTs could correctly identify the absence of antibodies 97% of the time and were therefore deemed suitable for use in national-scale antibody studies. And six had specificity over 98%, which the UK Medicines and Healthcare products Regulatory Agency recommends as the minimum level for use in the clinic. Sensitivity was much more variable, ranging from 21% to 96%, and in all cases was lower than what manufacturers reported.
Based on how easy it was to use, high specificity (98.6%), moderate sensitivity (84%), and availability for testing at scale, the Fortress test was selected for use in nation-wide surveillance studies and the first results are available here. Further evaluation of antibody testing kits is ongoing and new tests are being assessed as they become available. The researchers are also comparing performance to other potential home testing methods, including saliva samples and dried blood spots on paper.
Involving the public
In addition to evaluating how accurate the tests are, further REACT studies have carried out a major public engagement exercise and national usability study to assess how easily people can use them at home without supervision. Feedback from a pilot user testing study involving 315 people was used to guide the design of the testing pack, an instruction manual and how-to video for the test. The researchers then sent two different LFTs, including the Fortress test, to a total of almost 14,500 adult volunteers across England.
"Our study is the largest of its kind to explore the acceptability and usability of antibody self-testing with the general population." Dr Christina Atchison School of Public Health, Imperial
The public showed high levels of willingness to do the test, with almost all of those invited attempting it. Over 97% of participants successfully completed the test on their own by following the guides, and up to 94% had a valid result. Importantly, participants’ interpretation of the test results were very similar to that of a clinician, showing that in the majority of cases members of the public were able to correctly read the result themselves (more than 98% of the time for the Fortress test).
These findings have been published in the journal Clinical Infectious Diseases.
Dr Christina Atchison, lead study author, said: “Before we could embark on such an ambitious national surveillance study, it was vital that we evaluate whether LFTs are suitable for use at home. Our study is the largest of its kind to explore the acceptability and usability of antibody self-testing with the general population and we have built the critical evidence to underpin this approach, enabling us to scale up to testing more than 100,000 each month.”
Article text (excluding photos or graphics) © Imperial College London.
Photos and graphics subject to third party copyright used with permission or © Imperial College London.