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:

“This week’s bulletin from the ONS Covid-19 Infection Survey (CIS) takes the data up to the week of 23-29 May, that is, last week. In broad terms it looks as if it doesn’t tell us anything much that we didn’t already know, from the data on new confirmed cases on the dashboard at and from various reports about new variants. In fact, though, it plays an important role in confirming that the messages we’ve been getting from those other sources aren’t misleading. The case data on the dashboard, and most of the reports on new variants that come from the excellent UK sequencing labs, are based on swabs and other tests done on people who were tested because they have symptoms, or because of their jobs, or perhaps because they live or work in a place where surge testing is going on. So it’s possible that those figures may be biased, particularly if the pattern of reasons for testing changes over time. The CIS results are based on a survey of a representative sample of people across the UK community population, with appropriate adjustments being made to allow for any remaining lack of representativeness, so that don’t suffer from those possible biases. That’s why the CIS is as important as it always was – indeed arguably it’s increasing in importance as we move out of lockdowns.

“We’ve seen many reports, generally based on the new case counts on the dashboard, saying that cases have been increasing, in total across the UK, in the past two or three weeks, after a period of being roughly level in late April and early May. The CIS results confirm that position, particularly in England. ONS report that the percentage of people testing positive for the virus in England increased in the latest week, and quite considerably so. For the previous week, they estimated that 48,500 people in the English community population would test positive – that’s about 1 in 1,120. For the most recent week, that estimate has gone up to 85,600, or about 1 in 640. I should point out that there is quite a wide margin of statistical error around that estimate – it could plausibly be between 1 in 540 and 1 in 760. But there’s very clear evidence of quite a big increase on the week before. It’s important to realise that the infection rate is still extremely low compared to what it was back in January and February. It is at roughly the same level in England as in mid-April, and back then we were delighted at how low it had become. But it was decreasing then, it’s increasing now, and it’s currently at about three times the level it was in July and August last year. There’s no way to tell from these data to what extent the increase is due to increases in new variants, to the openings after stage 3 of the roadmap in England on 17 May, or to something else entirely.

“Because the infection rate is still quite low, and because the number of people tested in the CIS in the UK countries other than England is considerably smaller than in England, the trends are less certain in Wales, Scotland and Northern Ireland. In Scotland and in Northern Ireland, ONS say the trends are uncertain, though they did go up in Scotland in recent weeks. In Wales, ONS are talking of ‘early signs of an increase’. The estimated rate of testing positive in Wales is still quite a lot lower than in the other three UK countries, at about 1 in 1,050. The estimates for Northern Ireland and Scotland are roughly 1 in 800 and 1 in 680 respectively, though the margins of error around all those estimates are wide.

“The ONS estimates for regions within England, and for individual age groups, have considerably larger margins of error than the estimates for the whole of England, because the number of people swabbed for the survey in a single region or age group are a lot smaller than for the whole country. For regions, ONS estimate that the number of people who would test positive has increased in three regions (North West, East Midlands, South West), and that there may be signs of an increase in another two regions (West Midlands and London). Trends in the remaining four regions are said to be uncertain. It’s slightly depressing that ONS don’t feel the numbers are decreasing in any English region. However, all this draws attention to something the CIS can’t tell us, which is anything useful about exactly where, within in each region, cases might be going up or down. That’s because the number of people swabbed for the survey in smaller areas than whole regions is now much too small to give accurate enough results, now that infection levels are low. (ONS used to provide estimates from the CIS for smaller areas, but stopped in mid-March because the figures just can’t be accurate enough.) At this stage of the pandemic, while overall infections are low, the pattern tends to be dominated by spikes and outbreaks in relative small geographical areas. Data from routine testing and surge testing can keep an eye on those, but we still need the CIS to check that overall biases aren’t confusing the picture.

“The results for England for separate age groups show increases in most age groups, according to ONS, though they say that trends are uncertain in children of primary school age and younger, and in people aged from school year 12 (aged 16-17) and age 34. Rates of positivity in the adult age groups are all still low, however. The highest current estimated rate, and the biggest increase over the past couple of weeks, is in those of secondary school age (years 7 to 11). For that group, at the end of the most recent week, ONS are estimating that about 1 in every 180 would test positive (though the margin of error goes all the way from 1 in 90 to 1 in 380). That estimated rate went up by about two thirds just over the most recent week, and it’s now over four times as big as the rate in the age groups with the lowest estimates (50-69 and 70+). But, as the ONS bulletin reminds us frequently, trends are difficult to estimate well when infection rates are low, so the pattern might or might not look quite a lot different next week.

“What the CIS can’t tell us, not on its own at least, is how any increase in infections might feed through into increases in hospitalisations or deaths. The patterns linking infections to serious illness or death have been changed completely by vaccination, and we still don’t have good data on how the links between infection and serious illness actually look now, for some of the virus variants. So the CIS results can certainly feed into decisions on policy, such as on what might happen on 21 June in England, but they are only part of a complicated picture.”


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

“Today’s ONS figures confirm what we had suspected would be true, the number of cases of covid19 is rising.

“There are two factors at play here, once is the easing of the lockdown measures in May and the second is the delta variant (which has now become dominant).

“The trajectory of the case numbers in the North West are particularly worrying. East Midlands appears on a similar track. Other areas currently look better. In the absence of vaccines, we would expect cases to rise across the country with a delay, that is we should see the North West as the future. Localised control measures and track trace are not working, as expected from previous experience. We are seeing some evidence for an increase in hospitalisation where the case numbers are highest. Without vaccines, we would seem to be a the start of a third wave and given the nature of the delta, such a third wave could have been particularly disastrous.

“It is worth pausing to remember that if it is as transmissible and severe as early data indicate the delta variant will devastate less developed countries. I am filled with dread and sorrow for what lies ahead. Common humanity means it is urgent that we support vaccination across the globe.

“For the UK the picture is much better, complete vaccination (double jag) is effective against the delta variant. If 80 % of the population (including 12 to 18 yr olds) were vaccinated, we would be confident that this virus increase would quickly halt and the population emerge largely unscathed.

We are not there yet, there is a genuine possibility that a large number of cases could result with long covid complications in significant numbers of young people and an increase in deaths of the vulnerable.

“This will not be on the scale we saw in January.

“There are genuine concerns about waning immunity and variants against which vaccines are less effective, but these are not relevant to the decision ahead of us this month.

“It is competition between vaccination and virus. In thinking about the competition we have to remember the time lags at play. The number people with effective immunity has to be based on the number of people with two jags over two weeks ago. The number of people in hospital will have been infected over 1 week, possibly two weeks ago. A person infected today won’t end up in hospital for at least a week.

“To adopt a van Tam-ism, the cup final is decided at the final whistle not in the middle of the second half. At the moment it is end to end stuff, one moment we see data suggesting the variant may be ahead and at another the vaccine. The Prime Minister has said they will review the data and decide, this is exactly the right course. The final whistle is next week. Extending the metaphor, we can all make a difference to the result. Testing at home, wearing masks, meeting outdoors where possible and social distance when we can will slow the virus spread buying the vaccine time to score.

“I do think we might dispense the opinions of those who so confidently opposed lockdowns in autumn, November and December and do so again without any recognition that had the Prime Minister followed their advice, tens of thousands more people would have died. Too many families have already faced terrible losses, much loved relatives dying alone and frightened, gasping for air.

We may be able to open up on June 21st but we should look to the data (trends in case numbers, trends in hospital admissions, trends in vaccination coverage, knowledge about transmissibility and severity of the variant), not to those who consider so many other people’s lives so cheaply.”



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.”

Prof James Naismith: “Other roles which could give rise to conflict of interest:

Founder and shareholder of GyreOx

Non-Exec Director of GyreOx

Trustee (member of Council) Royal Society

Multiple patents (macrocycles, nanobodies)

Member of the European XFEL GmbH Council (equiv non Exec Director)

Editor of MedComm 

Scientific Advisory Board Harbin Institute of Technology  

Scientific Advisory Board Michael J Bishop Cancer Research

Scientific Advisory Board Alan Turing Institute

Advisory Board White Rose PhD training program

PhD examiner and advisor at multiple Universities

Panel D BBSRC.”

in this section

filter RoundUps by year

search by tag