select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to latest data from the ONS Infection Survey

The Office for National Statistics (ONS) have released the latest data from their COVID-19 Infection Survey.


Prof James Naismith, Director of the Rosalind Franklin Institute, and University of Oxford, said:

“This caps a good news week for the UK.  We have lowered prevalence to around 1 in every 1000 people, a further downward tick from last week. Once again the drop is seen across the UK’s nations.

“Over 70% of adults have antibodies against the virus, with over 25 % completing their second dose many of the most vulnerable are as protected as we can make them and in broad terms we are approaching two in three adults with at least one jab.

“Prevalence in the over 70s is now approaching zero.

“The new variants are here but they remain at very low levels.

“The current relaxation strategy is being vindicated.

“Those aged under 18 are not yet included in the vaccine campaign, the prevalence is higher in this group (3 in 1000 for school years 7 to 11).  This is a reservoir for the virus, thus completely abandoning all restrictions will see the virus come back.

“Provided the vaccine campaign continues, prevalence continues to fall and the new variants remain highly controlled the end is in sight.

“Similar conclusions can be drawn from data from the USA and EU.

“We have suffered too many tragedies to celebrate but we can be pleased by these data.

“What is happening in India is heart breaking.  I am sure we all want alleviate the suffering of other humans and self interest means such free circulation of the virus seen in India, some African and South American countries poses a threat to us.

“Every new infection is an opportunity for the virus to sample new mutations, this is an opportunity we do not want to give this virus.

“I strongly urge richer countries to do all they can to support vaccination across the globe.”


Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“There’s more good news on the progress of the pandemic, this time from the regular bulletin from the ONS Covid-19 Infection Survey (CIS).  This takes the data up to the week 18-24 April.  There’s been a certain amount of concern that the huge decline in infection rates, that we’ve seen in all data sources since the peaks at the start of the year, might level off or even reverse after the loosening of the lockdown measures that’s already happened.  But there’s no sign of that in the new data, with one possible, tiny, pretty unimportant exception that I’ll come to later.  In England, the Easter school holidays mostly ended on 12 April, and non-essential shops, gyms, pub gardens, and more opened again on that day.  Several similar changes have occurred in the other UK countries too.  If any of that had had a big and rapid effect on infection rates, we’d have begun to see that in the latest CIS data, but there’s pretty well no sign.  This doesn’t mean that an increase in infection rates will never happen.  People aren’t (yet) going back to the shops in the numbers they used to, and not everything that could reopen under the new rules has reopened.  (My local pub hasn’t reopened yet, and nor have several more round where I live, for example.)  And there will be further removal of restrictions before long.  So it’s possible that infection rates will start to go up again at some point.  But, if and when they do, the good news we’ve had about the effectiveness of vaccination in reducing the risk of serious disease and hospitalisation means that the impact on people’s lives will be less.  Vaccination is rolling out to more and more younger people, and they do need to take up the invitations to avoid any possible increase in serious Covid-19 in those age groups, if and when infection rates do start to rise.

“The latest ONS CIS data show falls in the estimated numbers who would test positive for the virus that can cause Covid-19 in all four UK countries, for the latest week compared to the week before that.  The rates do differ between the four countries – lowest in Wales with about 1 in every 1,570 people testing positive, highest in Scotland with 1 in 640 testing positive, with England (1 in 1,010) and Northern Ireland (1 in 940) somewhere between.  But those are estimates based on survey results, so inevitably there is some statistical uncertainty about them, and that uncertainty can be relatively large when we get down to such low numbers of people testing positive.  For example, ONS say that in Wales, the number testing positive could plausibly be somewhere between 1 in 740 and 1 in 4,610, rather than necessarily being close to 1 in 1,570.  Because of this statistical uncertainty, we can’t even be completely confident that the rates of testing positive differ at all between the four UK countries.  ONS estimate the total numbers of people (aged 2 and over) who would test positive in the community populations of each country – just adding up the central estimates of those numbers for the latest week, 18-24 April, indicates that around 64,900 people would test positive for the virus.  For the week before, the corresponding number was 104,000, so that’s a really big reduction (38%, or almost two in five) in a week.

“The estimates for the regions of England aren’t as precise as those for the whole country, simply because the number of people swabbed for the survey in a region is bound to be a lot less that the number for the whole country.  But ONS estimate that the rates of testing positive are falling in seven of the nine English regions, and in the remaining two (Yorkshire and the Humber, and the East of England), they say the trends are ‘uncertain’, so there’s no clear sign of an increase in any region.  The estimated rate of testing positive is highest in Yorkshire and the Humber, but that doesn’t mean it’s high in absolute terms there.  ONS estimate that between 1 in 340 and 1 in 930 would test positive there, so a large margin of statistical uncertainty, with a central estimate of 1 in 530.  That’s pretty clearly higher than the region with the lowest rates, South West, where between 1 in 1,250 and 1 in 11,320 (yes, over 11,000) would test positive, with a central estimate of 1 in 2,980.  But the rate in Yorkshire is still very much lower than it was – under half the rate at the start of April.  Looking at the numbers for age groups in England, ONS report that rates of testing positive are falling in people aged 35 and over, and in children of below school year 11 (aged 16 or 17).  Between those two ages the trends aren’t clear.  But again, none of the rates for different age groups is particularly high.  The highest estimate is for children in school years 7 to 11, where ONS estimate that 1 in 380 would test positive.  That’s clearly higher, even allowing for the statistical uncertainty, than the estimated rate for people aged 70+, where the estimate is that only 1 in 2,830 would test positive, but it’s still a very low rate compared to what we saw relatively recently.

“Rates of testing positive in England still aren’t as low as they were last summer, though, when only about 1 in 2,000 would have tested positive, compared to about 1 in 1,000 now.  In Wales, rates of testing positive are now getting down very close to the levels from last August, though they remain a bit higher (about 1 in 1,600 now, about 1 in 2,000 then).  This comparison can’t be made using CIS data for Scotland or Northern Ireland, because the CIS wasn’t running in those countries that far back.

“The CIS also estimates the numbers of new infections each day in the four UK countries.  That’s not the same as the total numbers testing positive, which includes people whose infection isn’t new but may have started a few days or even a few weeks before.  These so-called incidence rates are where the only tiny cloud in the good news appears, but it really is very tiny.  ONS can estimate these numbers of new infections, because people are tested repeatedly in the CIS so that there is some information about when someone’s infection actually began.  That’s an important strength of the survey.  But the estimation isn’t easy, particularly when infection rates are as low as they are now, so the margins of statistical error are pretty wide.  ONS estimate that the rate of new infections is falling in England.  In the other three countries, they say instead that it remains level, though in fact the central estimate fell in Wales and in Northern Ireland too.  The central estimate rose a little in Scotland, though by too little for ONS to be sure it’s a real rise (or that the declines in Wales and Northern Ireland are real decreases), because of the statistical uncertainty.  Overall, in the latest week, ONS’s central estimates for the number of daily new infections across the four UK countries add up to 4,610 a day.  That’s 44% down from the figure for the week before, which was 8,200.  I must emphasise again that there’s a lot of statistical uncertainty about all these figures – but the average trend across the whole UK is very encouraging.

Further information

“The incidence rates of new infections are the figures that you might expect to move in the same way as the counts of new confirmed cases on the dashboard at, because both are measuring new cases.  Those dashboard numbers do show confirmed cases continuing to fall in all four UK countries, including Scotland.  But they are not entirely comparable with the CIS estimates anyway.  To become a confirmed case in the dashboard counts, someone has to have tested positive, which obviously means that they have to have had a test.  Despite the increases in testing people without symptoms in some circumstances, most people who become confirmed cases on the dashboard would have been tested because they have symptoms.  So changes in how likely it is to have symptoms may bias the dashboard results, as may changes in how likely people are to come forward for a test if they do have some suspicious symptoms.  That bias won’t happen with the CIS estimates, because the CIS tests a representative sample of people from the community population only to provide estimates of the progress of the pandemic, regardless of whether they have symptoms.  There’s some evidence that being vaccinated reduces the chance that someone, who is infected, will have symptoms, so it’s possible that the size of any biases in the dashboard data may be changing because of vaccination.  The main message, though, is that infections are generally falling, however you measure them – though there may be limited exceptions to that in some places, perhaps because of limited outbreaks.”


Prof Paul Hunter, Professor in Medicine, The Norwich School of Medicine, University of East Anglia, said:

“Today’s report of the ONS Coronavirus Infection survey confirms what we have been seeing in the daily reports of new case numbers.  Taken together this is good evidence that case numbers are still in decline across all four countries within the UK.

“What makes this week’s results particularly important is that this would be the first week when there would be any evidence that the relaxation of the 12th April would have had a negative impact on the epidemic.  That there is in fact no evidence of an increased transmission risk is reassuring that for the time being at least it looks like the current Road map is still on target.”


Prof Rowland Kao, the Sir Timothy O’Shea Professor of Veterinary Epidemiology and Data Science, University of Edinburgh, said:

“The results of the ONS survey are consistent with other measures in saying that the numbers of randomly selected individuals testing positive for COVID-19 are declining.  This continued decline is good news and should be celebrated.  As it is a random survey, it is less subject to biases, but as it picks up a selection of all infections, rather than just newly reported cases (the way some other measures due) it will tend to more slowly reflect changes in trends.  As such it does not provide us with more information about what recent changes in restrictions are doing.  Thus we must continue to be aware of how those changes may be influencing the upcoming numbers of infections, including the possible spread of variants of concern and variants under investigation.  Key to those are the any impact that changes in longer distance travel may have on the rapidity of spread of any variants – should renewed measures be required (such as surge testing, or locally increased restrictions) longer distances would likely mean these would have to be broader, or more severe.”



All our previous output on this subject can be seen at this weblink:



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 received.

in this section

filter RoundUps by year

search by tag