select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
The Office for National Statistics (ONS) have released the latest data from their COVID-19 Infection Survey.
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“The latest release of results from the ONS Coronavirus (Covid-19) Infection Survey, the CIS, takes the estimates of the numbers of people who would test positive for the virus up to the week 5-11 December.
“The CIS results are beginning to pick up clear signs of the rise of the Omicron variant, particularly in London, but up to the latest date in the report, 11 December, the Delta variant was still very dominant across the UK. Of course, that might well have changed in the subsequent week.
“I mainly want to comment today on some reasons why the CIS will be particularly important in understanding the pandemic over the next few weeks, as Omicron takes hold, but I’ve made a few brief points on other aspects of this week’s CIS findings at the end of my comments.
“Results from the CIS are going to become increasingly important with the rapid rise of the Omicron variant. You might question why we need them at all, given the monitoring every day of new confirmed cases of Covid-19, published on the dashboard at coronavirus.data.gov.uk, as well as other information about Omicron new cases specifically. But the snag with most of those numbers is that they are based on the results of PCR and lateral flow device (LFD) tests carried out within the standard testing programmes that we’re all encouraged to use. So, to a considerable extent, they can be affected by who is choosing to be tested. If the numbers of people being tested, or the types of people being tested, change over time, then that can introduce potential biases in the trends in numbers of confirmed cases. In other words, they might not accurately represent the real trends in infections in the country.
“There certainly have been considerable changes in the numbers being tested in recent weeks. It’s difficult to get consistent detailed figures for the whole UK, because of differences in how the data are measured between the different UK countries. But, just for England, the average daily number of LFD tests conducted in the week 9-15 December was 823,657 according to coronavirus.data.gov.uk, and that’s almost a quarter higher (23%) than the figure for just one week before (670,517). There’s been a pretty clear upward trend in the number of lateral flow tests since mid-November at least. I’m not complaining about that at all – it’s generally a good thing, in my view, that testing is increasing – and there are many possible reasons for the increase. The Government has been advising us to test regularly. People may be concerned about infecting others at Christmas or when Christmas shopping, and may be testing to be careful about their own health and to be considerate to others. The publicity about the Omicron variant will have led to more concern and more tests. There’s been an increasing trend in the numbers of PCR tests in England too, since early November.
“The problem with the increased testing is that, in very crude terms, the more you test, the more cases you find. That can be particularly true of a disease like Covid-19, where a substantial proportion of infections don’t result in any symptoms, but the infected person can still pass on the virus to others. More testing can lead to more of these asymptomatic cases being detected, which is a good thing for reducing the disease spread, but which makes the figures for new confirmed cases harder to interpret. If the new case numbers go up, as they have been doing since early November, we can’t tell exactly how much of the rise is because the number of infected people out there has really risen, and how much might just be because we’ve detected more cases that were always there. Don’t get me wrong – I’m not claiming the huge rises in cases in recent days aren’t real. There’s certainly good evidence of big real increases. It’s just that we can’t be sure exactly how big, and that can be important for governments and for individuals in deciding what to do.
“This is where the CIS comes in. It doesn’t test people because they have symptoms, or because they have decided to get tested for whatever reason. Instead it tests a representative sample of the UK community population simply to track the progress of the pandemic. So it isn’t affected by biases arising from the number and type of people who choose to be tested in the standard testing programmes.
“If the case numbers arising from the Omicron wave become as high as has been predicted, that will be another compelling reason for the importance of the CIS. So far, the numbers of confirmed cases, while they are getting alarmingly large, have still been within the capacity of the existing standard testing programmes. That is, we can test enough people to detect and record a big proportion of the new cases. But if cases increase a great deal more, there just won’t be the available testing capacity to detect enough of them, so that it’s possible that the daily confirmed case numbers can no longer measure the pandemic accurately. That won’t happen to the CIS.
“That doesn’t mean that the CIS is a perfect tool for tracking the pandemic. No perfect tool exists. An issue is that its results can’t be produced instantly – today’s latest figures are for the week 5-11 December, and it’s nearly a week since that week ended. That length of delay wasn’t a major problem for tracking the variants we saw before Omicron – but the very rapid Omicron rises may unfortunately mean that even the very rapid, accurate and efficient CIS can’t quite show the current picture clearly enough. We’ll see.
“There is of course another survey, in England rather than the whole UK, that also aims to track the pandemic by testing a representative sample of people. This is the REACT-1 survey, run from Imperial College. However, because that runs in a series of rounds, with gaps of some weeks in between the rounds when nobody is being tested, it can often be subject to even longer lags in its ability to report (depending on where we are in relation to the survey rounds), and that could well be an issue in using REACT-1 findings in relation to Omicron.
“A few brief points about today’s CIS results. ONS report that the trends in the percentage of people who would test positive in the most recent week (5-11 December) in England, in Wales and in Scotland are all uncertain, but the trend is downwards in Northern Ireland. Infection levels do remain high in all four countries. The estimated percentages testing positive increased in some English regions (London, East of England) and decreased in others (North East, Yorkshire and the Humber). In England, the estimated rate of testing positive in the most recent week was again highest in the youngest age group (2 years of age up to school year 6), where an estimated 1 in 20 would test positive. That is 17 times as high as the estimated rate in people aged 70 and over.”
All our previous output on this subject can be seen at this weblink:
www.sciencemediacentre.org/tag/covid-19
Declared interests
Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee. I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic. My quote above is in my capacity as an independent professional statistician.”