You asked
Reliability of SARS-CoV-2 coronavirus test used by UK Government
The Office of National Statistics has indicated in its publication of 10 May 2020, about the national Coronavirus Infection Survey, that the currently used swab test is of uncertain reliability in that the rate of false positives and false negatives is not known:
"Uncertainty in the test (false-positives, false-negatives and timing of the infection) We do not know the false-positive and false-negative rate of the current swab test for the virus. False-positives and false-negatives could also come from the fact that participants in this study are self-swabbing. We also do not know if all individuals testing positive are still infectious. Some may have had the coronavirus (COVID-19) in the past but still test positive."
Please can ONS please clarify the basis upon which it stated: "We do not know the false-positive and false-negative rate of the current swab test for the virus."
Please include in your response the source of this information eg. the government department or public body which made this assertion.
Please disclose the communication which conveyed this information to ONS.
Please also advise if ONS has queried and or raised any concern with the government, any government department, public health England or any other public body about the fact the false negative rate and false positive rate for the current test is unknown.
Please disclose any correspondence querying or raising concerns about this situation.
Please advise if ONS is collating data on the results of any other microbiology test for which sensitivity and specificity (false negative and false positive rates) is apparently unknown.
We said
Thank you for your request.
When considering the false-positive and false-negative rate, it is important to note that no test is perfect. We are striving to be open and honest about this by communicating to users in our statistical release that the exact false-positive and false-negative rate is unknown.
To provide helpful and transparent information to users we are using the latest evidence and academic expertise to put a plausible bound on the sensitivity/specificity, which we continually review. We are not making strict judgements on the sensitivity/specificity as to a certain extent we do not know the answer, and want and need to be honest in conveying this information.
Please see our bulletin, where we have included a section on the test sensitivity. By using Bayesian analysis and available academic literature we have calculated what prevalence would be in a couple of scenarios with different sensitivity; we have found that even if there was a relatively high rate of false-negative results, the positivity rates would still be fairly close to the true figure throughout the study.
It should be noted that, as COVID-19 is a novel virus, it is more difficult to assess what sort of benchmark we should be using for the tests. For context, tests were developed for patients in hospital with the virus, who have high viral load – even if a test performs well in that situation, this does not equate to equivalent sensitivity in patients with a milder or asymptomatic infection. So, as we cannot assume equivalent sensitivity, calculating the false-positive rate is more difficult. Comparing results of different tests to try to identify the 'better' test is also difficult as any differences are a combination of false-positives and false-negatives, and there is no one way to easily distinguish between them. In short, there is no obvious 'gold standard' to which our tests should be performing.
In processing the test results ONS are working with the UK Biocentre, as part of this we have worked to understand test sensitivity (considering the various difficulties outlined above) and continue to do so. We will consider undertaking 'triangulation' work to compare our results with other similar studies, which will consider the various methodology differences between them. This could include assessing their statements on false negative and false positive rates to check for new evidence or information that we should evaluate. And finally, as this is a pilot study, we expect to provide further updates in publications as we reach the next stages of the study.
With regards to the specific answers to the questions that you raised:
Please can ONS please clarify the basis upon which it stated: "We do not know the false-positive and false-negative rate of the current swab test for the virus."
This has been described above; the basis is that these rates are an unknown variable.
Please include in your response the source of this information e.g. the government department or public body which made this assertion.
This information is from discussions with our partners in the study, the University of Oxford and the UK Biocentre.
Please disclose the communication which conveyed this information to ONS.
The study was set up very quickly and the pace of work necessitated regular, lengthy meetings between ONS and our partners for the study. As such, the communication that disclosed this information was verbal.
Please also advise if ONS has queried and or raised any concern with the government, any government department, public health England or any other public body about the fact the false negative rate and false positive rate for the current test is unknown.
As explained above, when setting up the project, discussions took place with our research partners. This included discussion with the UK Biocentre around false-positive and false-negative rate rates. No explicit concerns were raised, as we understand that no test is perfect, and we have used the expertise of our research partners to understand the implications of the false-positive and false-negative rate. We continue to work with our partners to understand test sensitivity; as explained previously, this included conducting analysis on the effect of different sensitivities. Further evidence concerning sensitivity and specificity will be made available when appropriate.
Please disclose any correspondence querying or raising concerns about this situation.
No explicit concerns were raised and discussion on the subject was undertaken verbally at meetings.
Please advise if ONS is collating data on the results of any other microbiology test for which sensitivity and specificity (false negative and false positive rates) is apparently unknown.
As part of the Covid-19 Infection Survey, we are also testing for antibodies to Covid-19 in blood samples from approximately 10% of the survey participants. Further information about this assay will be included in a future bulletin