16/09/2020 By Desmond Swayne
On Tuesday the following exchange took place in the Commons:
Sir Desmond Swayne (New Forest West) (Con) To what extent is there a possibility that it is the exponential increase in testing itself, in identifying genuine new cases, and the very significant possibility of false positives, that is giving a distorted impression of the trajectory of the disease?
The Secretary of State for Health and Social Care (Matt Hancock) I like my right hon. Friend very much and wish that that were true. The reason why the Office for National Statistics does the surveillance testing is to ensure that we are constantly looking, on a national representative sample, at what the case rate is, as well as, of course, using the tests, and as we increase the testing numbers, we will inevitably find more of the cases that are there. The ONS survey published on Friday shows a rise in the numbers commensurate with the rise in the numbers of tests that have come back positive, and that does take into account the point about false positives, which is an important one.
The reason I put the question is that I have received a number of representations from clinicians across the country (presumably prompted by the stand that I’ve taken) who believe that Covid-19 testing is giving a misleading impression about a second wave. They argue that of course we will detect more genuine cases of the disease as we do more testing. It doesn’t necessarily mean that the disease is increasing: we may simply finding more of it because we are looking for more of it.
More worrying is the belief that the a false impression is being created by the test which is unable to distinguish between active cases of the disease, and those who may have recovered from it some time ago but have remaining non-infective traces of it.
I’m not sure that the Secretary Of State’s answer provides sufficient reassurance. He is stating that the Office of National Statistics conducts a survey by testing a representative population sample and then applying a statistical calculation to estimate the number of cases in the whole population.
If the ONS test suffers from the same inability to distinguish false positives it doesn’t escape the possibility of exaggerating the incidence of the disease.
The ONS does publish the following caveats:
“we do not know the true sensitivity and specificity of the test, as COVID-19 is a new virus….
The estimates presented in this bulletin contain uncertainty. There are many sources of uncertainty, including uncertainty in the test”
The Government website -when I last looked- stated the daily number of tests processed in the UK was 227,075. https://coronavirus.data.gov.uk/testing and the daily number of people tested positive is 3,105. https://coronavirus.data.gov.uk/cases
Which means that 1.37% of those tested had a positive result. Clearly it doesn’t take a mathematician to tell you that a relatively small percentage of false positives will have a significant impact on our estimate of how many people are infected. In fact a report by the ONS in June put false positives at 2.4%: which even higher than the actual proportion of positives!
We are threatened with more restrictions to our liberty which will lay waste to even more of our employment opportunities, and will be more damaging to the health of anyone suffering from a medical condition other than Covid-19, all driven by policy reliant on what may be very misleading data
At some stage we will have to have a reality check and remind ourselves that whilst Covid-19 is a horrid and dangerous disease, it is only 24th in the league of UK killers at only 1.4% of deaths (the top scorer being dementia at 11%).
If we want to live normally we will have conquer our fear