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:
“It’s concerning that, in the latest results from the ONS Covid-19 Infection Survey (CIS), the estimated numbers of people who would test positive have increased in the latest week, in pretty well all of the subdivisions of the UK population that the researchers looked at. They report increases in England, Wales and Scotland, though they regard the trend in Northern Ireland as uncertain. They also report an increasing trend in infection in all nine regions of England, and in all the age groups in England. Obviously that won’t feed through in terms of serious illness, hospitalisations and deaths in the way that we used to see with previous variants and before vaccination, but we are seeing increases in hospitalisations on the Government dashboard.
“I want to comment on a different aspect of the CIS results, though. As well as their estimates of numbers of people who would test positive, whether their infections have just begun or have been going on for some days or even weeks, the CIS researchers present weekly updates of their estimates of the numbers of new infections. These so-called incidence estimates, in principle at least, should match more closely onto estimates of new confirmed cases of Covid-19, on the Government dashboard and elsewhere. These CIS new infection estimates aren’t usually reported prominently in the media, possibly because (for technical reasons) the latest ones relate to an earlier time period than the weekly estimates of people testing positive. For instance, in this week’s CIS bulletin, the latest positivity estimates are for the week ending 12 March (so just last week), while the latest ‘official’ incidence estimates are for the week ending 26 February. Also, again for technical reasons, there’s more statistical uncertainty about the incidence (new infections) estimates than the positivity (current infections) estimates.
“But the CIS incidence estimates can tell us things that the dashboard new case figures can’t. That’s because the dashboard figures are affected by the numbers of people being tested for the virus, and indeed by who is asking to be tested. Other things being equal, if fewer people are being tested, or fewer are reporting their test results to the Government, fewer new cases will be found for the dashboard. The CIS results aren’t affected by that sort of bias, because the people in the survey are tested simply to track the pandemic, and regardless of whether they have symptoms or are working in certain jobs. And in the most recent available weeks, a rather major discrepancy has emerged between the CIS estimates and the dashboard numbers, that throws some doubt on what the dashboard figures appear to be telling us about trends.
“You wouldn’t expect the numbers of new dashboard cases to match the CIS estimated new infection numbers exactly, for example because if someone is infected but have no symptoms, they are considerably less likely to show up as a new case on the dashboard, but will still count towards the CIS estimated of new infections, and generally they will still be able to infect others. So generally, since the CIS began to produce incidence estimates in England in June 2020, the CIS estimated number of new infections each week has usually been between two or three times the number of new cases on the dashboard for the same week.
“The latest available week of the CIS’s ‘modelled’ estimates of new infections goes slightly beyond the latest ‘official’ estimate, so covers 24 February to 2 March. The estimated number of new infections across the UK in that week was about 2.1 million, though there’s quite a wide margin of statistical error around that figure. However, the number of UK-wide new confirmed cases on the dashboard for that week is very much smaller, at about 250,000. In other words, for that week, CIS estimate that the number of new infections was over eight times the number of new cases on the dashboard.
“That’s a big change in the pattern of how these two sets of figures are related – but it’s not just a blip for that one week. As I said, generally in England the number of new confirmed cases on the dashboard has generally been between a third and a half of the estimated number of new infections according to the CIS. But in late January and early February, the dashboard case numbers fell to about a quarter of the CIS estimates, and they have gone down relatively further since, to about an eighth. The most obvious explanation for this is that people just aren’t being tested as much as they were, so that the dashboard isn’t picking up new infections in the same way that it did. (Either that, or the CIS estimates have somehow gone haywire, but I simply don’t believe that.) In fact the number of virus tests being carried out has fallen considerably, according to the dashboard, though it’s recently started to rise again.
“This discrepancy is a very important reason why it’s important to continue with the CIS. The indications are that the trends in cases on the dashboard may well not be representing what’s actually going on. That can work both ways, of course. Perhaps the marked decline in new cases on the dashboard since January was, to some extent, not a real decline and was partly due to people not being tested or not reporting their lateral flow test results. But also, the increase in new cases that we’ve seen on the dashboard since late February is perhaps partly due to more people being tested, and not entirely to real new cases. (We can’t tell yet from the CIS whether that’s true, because the CIS incidence estimates only go up to about the end of February, so far.)”
Dr Simon Clarke, Associate Professor in Cellular Microbiology, University of Reading, said:
“The latest statistics from the ONS show once again how transmission of Covid-19 quickly follows changing patterns of human behaviour. Concern about the virus among the public seems to be at an all-time low since the start of the pandemic, with a big reduction in people working from home, and less than a third of people choosing to regularly maintain social distancing from others.
“At the same time, we are seeing an increase in infections, with around 1 in 20 people infected in England, and even higher rates in Scotland and Northern Ireland. Hospitalisation rates of people with covid have increased, and while deaths remain low, there are still hundreds of people who are dying every week with this infection. The BA.2 version of Omicron seems to be behind this uptick in infections, which again shows how quickly the situation can change as the virus evolves into new forms.
“I understand that the UK government is keen to stop spending on expensive testing infrastructure, and people have had enough of isolation requirements. But the government must be careful not to dismantle all the systems which have allowed UK planners to stay ahead of the Omicron wave through a successful vaccination drive. Other countries have fared much worse, we cannot simply wish away this coronavirus. The ONS survey is now one of the few tools we have left to provide data on infection rates so it’s vital that it doesn’t become prey to internal party-political pressures.”
Prof James Naismith, Director of the Rosalind Franklin Institute, and Professor of Structural Biology, University of Oxford, said:
“The ONS survey to 18th March shows evidence for a sharp uptick in the cases of covid19 (omicron, with BA.2 now dominant).
“Prevalence in Scotland is quite remarkable, 71 per 1000 people, and testament to the infectivity of the new strain. It is dramatically improved compared to the original strain.
“England is not far behind with 50 per 1000; I would expect England to match Scotland. The reality is that differences in control measures between the countries are meaningless.
“Prevalence is increasing in every area and across age groups. Without vaccines, this would be bleak for the this country. Countries with poor vaccination rates face a very challenging future.
“At this level of prevalence and the decision not to halt the spread, the most likely outcome is that almost everyone who is susceptible is going to be infected.
“For the UK, my main concern is for the vulnerable for whom this disease is serious and for those whose lives will be blighted by long covid19.
“Every effort must be made to triple vaccinate as many people as possible, quadruple vaccinate the most vulnerable and make available antivirals.”
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.”
None others received.