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

“The latest results from the ONS Coronavirus Infection Survey (CIS) cover the week from 26 April to 2 May, so if there were any big effect of the loosening of lockdown restrictions in England on 12 April, it would show up in the CIS figures by now. But the excellent news is that there’s no sign of anything like that. The numbers of people who would test positive for the virus has continued to fall across the UK. ONS report that the numbers testing positive decreased in England, in Wales and in Scotland in the latest weeks. In Northern Ireland they consider that the numbers have kept level.

“One thing that the results do show is the difficulty of getting precise estimates of infection rates from a survey, now that the rates have got so low.  That’s particularly the case in the smaller countries, Wales and Northern Ireland, where the number of people swabbed for the survey is smaller than in Scotland and much smaller than in England, just because of the smaller populations. For instance, in Northern Ireland in the most recent two weeks, they swabbed 5,401 individuals, but there were only 6 positive tests. You just can’t get a precise estimate of the rate of testing positive in Northern Ireland from such small numbers. ONS report that 1 in 750 people would test positive in Northern Ireland in the latest week, but that the margin of statistical uncertainty around that figure goes all the way from 1 in 370 to 1 in 2,160. However, even if the true figure was right at the ‘worst’ end of that range, 1 in 370, that’s very, very much lower than it used to be. It does look, for the latest week, as if the infection rates in Northern Ireland and Scotland might be slightly higher than in England, and that the rate in Wales could be considerably lower than all of those – but the statistical uncertainty is, inevitably, so big now that it’s not even certain that there is any difference in rates at all.

“Ignoring that uncertainty for now though, for the latest week ONS’s central estimate is that 56,900 people across the UK would have tested positive for the virus that can cause Covid-19. The previous week, the estimate was 64,900. So the latest week’s estimate is 12% down on the week before. It could have fallen by quite a lot more than that, or quite a lot less, but it’s pretty clearly continuing to move in the right direction. The estimated positivity rate in England is still about one and a half times as high as it was last August, though in Wales it’s down to last August’s level. I can’t make that comparison for Scotland or Northern Ireland, because the CIS wasn’t running there last summer. It would be better, of course, if the rates fell even more to last summer’s lowest level or below – but because of the success of vaccination, even if the rates do stop falling and increase a bit, that’s unlikely to have the bad effect on serious illness or deaths that we’d have seen last year.

“The position on statistical uncertainty means that it’s even harder to make good comparisons of infection rates in regions of England, or in different age groups, because the numbers swabbed in a single region or age group are obviously far less than for the whole country. But ONS consider that the rate of testing positive decreased in the most recent week in six of the nine English regions, the exceptions being Yorkshire and the Humber, the East of England, and London, where they say the trends are uncertain. The estimated rate is highest in Yorkshire and the Humber, where ONS estimate that 1 in 470 would test positive. London is second with 1 in 760. But even those are low rates. By age group, ONS estimate that infection rates are falling in most age groups but are uncertain in those aged between 16-17 and 34, and those aged 70+. But for the 70+ group the rate is now so low that the statistical uncertainty is huge, so that the trend just can’t be statistically clear. The central estimate for the 70+ group is that only 1 in 2,500 would test positive, but the statistical uncertainty means that it could be somewhere between 1 in 1,250 and 1 in 10,000. So few people in that age group are returning positive swabs in the CIS that it can’t be clear exactly which way things are moving from the survey data – but the rate is clearly so low that there’s no reason for concern at all.

“As always, the latest data on the rate of new infections (incidence) is for a week earlier than the prevalence estimates for all infections, so only up to the week ending 25 April. But that’s still enough to catch any signal of increase from the 12 April changes in England. New infections are even harder to estimate that the rate of all infections, because there are even fewer of them. So there’s a great deal of statistical uncertainty. But, just looking at the central ONS estimates, across the whole UK they are estimating 3,700 new infections each day for the week ending 25 April, which is a fifth lower than the estimate for the previous week (4,610). Things do appear to be moving in the right direction on new infections too, and there’s certainly no clear sign of any increase after the loosenings of lockdown, though ONS emphasise that the trends are very uncertain.”



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



Declared interests

None received.

in this section

filter RoundUps by year

search by tag