June 25, 2021

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:

“Today’s ONS figures (measuring the period up to 19th June)  show the prevalence of the delta variant have once again risen. In England more than 2 people in every thousand are infected.

“The prevalence in Scotland is significantly higher, I assume Scottish government science advisors will comment on the advice given and followed.

“The rate of increase in Delta is thankfully moderate and it would suggest that current restrictions will give the time to allow the vaccination campaign to do its work. If we can all keep to the current restrictions for a bit longer, we will defeat the delta variant in the UK. Had we removed restrictions, I would be less confident of success. What we are seeing in real time is the power of vaccination at scale to hold back a wave.

“It is important to understand the Delta variant is more transmissible, possibly more severe but it is not a new virus. Those most vulnerable, elderly and those with health complications, are at a much higher risk of serious disease. Young and healthy citizens are at much much lower but not zero risk. As we have seen, long covid in younger people is a serious problem. The vaccination campaign has saved thousands of lives particularly in over 50’s, the majority of whom are now doubly vaccinated. Vaccination has reduced the hospitalisation and burden on the NHS. We should pause to consider the value of the science investments made by governments in life science here in the UK and worldwide over the last 20 years. Without these, we would be helpless against the Delta variant.

It would helpful to remember this when considering future government priorities.

“As the campaign reaches the over 18’s, we will further reduce the potential of virus to spread and cause long covid. The threshold is around 80 % doubly vaccinated of the population who can spread the disease. We do not know the age cut off for spread of the virus is but it does include under 18’s. 

“Whilst adults can give informed consent, children cannot. Vaccination of children is only ethical if on balance it is beneficial to the child concerned. It is not appropriate to vaccinate children solely to protect others because any medical intervention, even something as safe as a vaccine, does have a very small risk. The younger the child the lower the risk to them from covid19, it will take some time to establish what age groups of children might be offered vaccination.

“For over 18’s the picture is clear, the risks from covid19 far outweigh any risk from the vaccine. I would urge all who can to get vaccinated quickly. 

“Whilst the UK is not out of the woods yet, it appears that we will win the race against Delta (just). Other countries may not be so lucky, we should do all we can to ensure vaccines are made available to fellow humans in poorer countries.

“Whilst much is made of failure to prevent import of cases, we can see from the UK and now the EU just how difficult this is to achieve in practice. In reality, the best solution is to vaccinate as quickly as possible.”

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

“This week’s bulletin from ONS on their Covid-19 Infection Survey (CIS) takes the data up to the week 13-19 June. Overall the picture isn’t very encouraging, particularly about the situation in Scotland. Infection levels across the UK as a whole, measured by the estimated numbers of people who would test positive on a swab test for the virus, are up again, compared to the previous week. There’s more evidence that the Delta variant has become dominant in Great Britain, though not (yet) in Northern Ireland. In most parts of the country, positivity rates are still well below what they were at the height of the pandemic wave in winter and early spring this year, however, and this means that ONS have to repeat their warnings that the estimates are subject to a lot of statistical uncertainty, particularly in countries other than England and in subgroups of the English population (by region or by age). That’s because the numbers of people swabbed for the survey in those areas are relatively small, and relatively small sample sizes together with low positivity rates inevitably lead to wide margins of error. But that doesn’t hide the position that, overall, things are still moving in the wrong direction.

“Overall, ONS’s estimate of the number of people who would test positive across the UK, in the most recent week, is about 29% higher than the figure for the previous week. That’s quite an increase. But it’s not spread evenly across the UK. For England, ONS estimate that about 1 in every 440 people in the community population would test positive (with margin of error from 1 in 380 to 1 in 520). ONS consider that the most appropriate measure of the rate of change in the positivity rate comes from what they call the ‘modelled daily rates’. In England, that rate in the middle of the most recent week was 9% higher than a week before. That’s about the same as they are now estimating for the change from the week before to the week before that. I’d be happier if the rate weren’t increasing, but things could be worse. But something interesting and important is that the increase is much less than the increase over the same period in new confirmed cases (weekly averages) on the dashboard at coronavirus.data.gov.uk. They went up by about a third, 35%, over that week. Now we’ve got to bear in mind that the ONS CIS measures all positive tests, even if the infection began some days earlier, while the dashboard new cases figures count people who have just newly tested positive, so they aren’t quite measuring the same thing. But the dashboard cases figures can be affected by changes in the numbers of people being tested, and in the reasons why people are tested. The CIS figures aren’t affected by biases of that sort, because they are derived from a survey of a representative sample of people who are tested only to provide estimates of the infection rate. The upshot is that, while it’s clear that infection levels in England are definitely increasing, they are not increasing as fast as the impression given by the dashboard new cases data. If the infection rate is increasing by 9% a week, as the ONS figures indicate for England, and if that rate of increase continues, the number of people testing positive would double every 8 weeks or so. That’s not wonderful, but it could be a lot worse.

“The trends in Wales and in Northern Ireland aren’t so clear. In Wales, ONS consider that the rate of testing positive did increase over the two weeks to 19 June, but the trend is uncertain during the most recent week. They now estimate that about 1 in 830 people in Wales would test positive. In Northern Ireland, the trend is uncertain, and ONS estimate that about 1 in 720 would test positive in the most recent week.

“However, in Scotland, there’s very clear evidence of an increasing trend. The ONS ‘official’ estimate of the rate of testing positive in the most recent week is about 1 in 220 people, a higher rate than in any of the other three countries, even allowing for the inevitable statistical uncertainty in estimates based on a survey. The estimated number in Scotland who would test positive in the most recent week is 24,400 (with margin of error from 16,800 to 33,500). The central estimate is not all that much short of three times last week’s estimate (which was 8,800). The ONS modelled daily rate for the middle of the latest week is almost exactly double the figure for a week previously – those rates are the ones that ONS regard as most appropriate for assessing trends. Obviously that corresponds to the number of people testing positive doubling in a week, compared to doubling in eight weeks in England. Putting it another way, the estimate of the rate of increase of positivity in England is about 9% a week; in Scotland, it’s roughly 10% a day, on these ONS figures. And another comparison: In England the latest official estimate from ONS puts the rate of testing positive about where it was in early April this year (or in mid-September last year), whereas in Scotland the estimate of the rate for the latest week is about the same as it was in mid-February this year when the pandemic was in a much worse position than in April. Obviously these ONS figures can’t give the reasons for this alarming trend in Scotland from these data, and nor can I. It could possibly be something to do with the rise to prominence of the Delta variant in Scotland having started rather later than in England and accelerated faster, though you’d have to ask a virologist whether that is a reasonable explanation. I suppose I can’t entirely rule out that it’s some sort of statistical anomaly (and I should point out that the confirmed cases on the dashboard, for Scotland, rose by much less, about 25%, over the week in question), but I still find the trend in these ONS figures for Scotland very concerning.

“I’ve already explained that the positivity estimates for the regions of England, and for age groups within the English population, are less precise than for the whole country, because the number of people swabbed for the CIS in an individual region or age group is smaller. That means that the trends are also less certain. Positivity rates are highest in the North West region, with an estimated 1 in 150 people testing positive. ONS estimate that infections are increasing there and in the North East, though currently in the North East the positive rate is still a lot lower than in the North West (at about 1 in 330). ONS report some signs of increase, at slightly different times, in the East of England, the South West, and the West Midlands, but say that trends elsewhere are uncertain. But infection rates in those regions are considerably less than in the North West, and in the South East the rate is still very low (1 in 1,130, but with quite a wide margin of error).

“In terms of age groups in England, ONS estimate an increasing trend in the age group from school year 12 (16 or 17 years of age) to age 24. They say that there was an increase in the age group 25-34 in the two weeks to 19 June, but that the trend in that group is uncertain in the second of those weeks. It’s impossible to say from these data, but if that increasing trend in the 25-34 age group has in fact stopped or even reversed, that might have something to do with vaccination, as at least some of that group will now have had a vaccine dose. That’s less likely to be true for the 16/17 to 24 age group – they have only very recently got access to vaccines, and although they appear to be being jabbed at a high rate now, the vaccine wouldn’t have had time to have an effect that would show up in these data. The rate of testing positive is highest in the 16/17 to 24 age group, with an estimated 1 in 150 testing positive (margin of error from 1 in 100 to 1 in 230). The infection rates in the age group one below that (secondary school age) and one above that (25-34) are only about half the rate in the 16/17 to 24 group, and positivity rates in age groups 35 and over are much lower still (e.g. for 70+, the estimate is 1 in 1,430), reflecting (among other things) the high level of vaccination in those age groups.”

https://www.ons.gov.uk/releases/coronaviruscovid19infectionsurveyuk25june2021

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.