The government was wrong on its own data on Covid-19 tests for care homes
23 July 2020
What was claimed
125,000 care home staff had been tested for Covid-19 as of mid-May.
This is not correct. Government data shows 125,000 tests on staff in care settings had been processed. It is not possible to tell the number of staff tested but it is likely much lower. The figure also includes tests on other people including family members, social care workers and care home CEOs.
What was claimed
118,000 care home residents had been tested for Covid-19 as of mid-May.
Incorrect. When the Prime Minister made this comment, 57,000 residents had been tested in public labs, and 61,074 tests were processed in commercial labs. Adding these together gets to 118,000 but that’s not appropriate as this would lead to residents being double counted.
“Mr Speaker, I’m afraid he is simply in ignorance of the facts. Because the reality is that already 125,000 care home staff have been tested, 118,000—perhaps he didn’t know that Mr Speaker—118,000 care home workers have been tested and we are absolutely confident that we will be able to increase our testing not just in care homes but across the whole of the community.”
"Since the start of the pandemic, nearly 125,000 workers in care settings and over 118,000 care home residents have been tested through DHSC and PHE testing routes."
Department of Health and Social Care press office, 20 May 2020
During Prime Minister’s Questions in May 2020, Boris Johnson claimed that 125,000 care home staff had been tested for Covid-19. We wrote a story about it at the time and the fact he mistakenly said 118,000 care home workers had also been tested. The Department of Health and Social Care (DHSC) confirmed to Full Fact that nearly 125,000 staff in care settings and 118,000 care home residents had been tested since the start of the pandemic. However, new government data shows that both Mr Johnson and the DHSC were incorrect—and, therefore, so was our story.
The new data, released on 16 July 2020, includes a section that specifically addresses the claims made by Mr Johnson about the testing in care homes. This clearly shows that the figure of 125,000 (or, to be precise, 124,906) care home workers relates to the number of tests done, rather than the number of people tested.
This is an important distinction, because we know that people often receive more than one test, and so it is very likely that the actual number of people tested at this point was considerably lower than 125,000. The government’s own testing methodology notes: “For clinical reasons, some people are tested more than once. Therefore, the number of tests will be higher than the number of people tested.”
Under a section on data limitations, the government release says: “This is a count of tests processed not individuals tested, so may count multiple tests for an individual both within and across testing routes.” What it really means is that we do not know from this data how many people were tested, and both Mr Johnson and DHSC were wrong to claim that we do.
We can use other figures to get a sense of how accurate they might be though. These figures show care home staff were tested under pillar 2 testing. The government’s test and trace data tells us 1.25 million people in England have been tested under pillar 2 testing between 28 May and 8 July, but does not publish the number of tests. DHSC data tells us just over 2 million tests were processed across the UK under pillar 2 in this time period, but does not have a figure just for England. As England makes up 84% of the UK population, we can very roughly estimate that 1.7 million of these tests were processed in England.
This means the 1.25 million people tested is about 75% of the 1.7 million tests done in England. Unless something very different is happening with care home staff, we could expect to see a similar disparity between the number of tests and the number of people tested.
There is also another issue with their claims. Although both Mr Johnson and DHSC said the figure was for the number of staff in care settings that received tests, the actual data shows that these workers are not the only people included in the 125,000.
For care home staff, the figures also include “other social care staff such as social care workers, CEOs of care homes, cleaners and administrative staff.” Social care workers included in this bracket include domiciliary carers and children’s social care providers, so may not have had contact with care homes. Some “symptomatic household members” of these workers are also included. It is not possible to tell from this data how many people who received tests were workers in care homes.
As for care home residents, although in May DHSC told Full Fact that “over 118,000” had been tested, the data also shows this is incorrect. It shows that, as of 12 May 2020, 57,000 residents in care homes had been tested in Public Health England labs, and, as of 18 May 2020, 61,074 tests had been carried out on residents of care homes in commercial labs. As above, we know that 61,074 tests are likely to translate into fewer actual residents being tested.
A single care home resident who was tested in a Public Health England lab and then who later submitted two samples to a commercial lab could be counted three times if you were to simply add the figures together, as Mr Johnson and DHSC did.
Full Fact asked DHSC to explain the discrepancy between the claims made by Mr Johnson and its own press office and the data shown in the new government release. We were pointed back to the government release, which says: “These figures were based on the best available estimates at the time and were therefore derived from a range of sources with different definitions and scope.”
The data release also says: “The figures quoted by the Prime Minister on 20 May included people tested and a count of tests, which gave a reasonable estimate of the scale of testing in care homes. Over time, as more people are tested more than once, the figures for number of tests and number of people tested have diverged more. That means that this method of calculating an overall estimate can no longer be used.”