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:
“The overall picture from today’s release on the ONS Coronavirus Infection Survey (CIS) is generally one of little change compared to last week’s release. The latest release takes the data on the numbers who would test positive forward to the week 24-30 October. Actually it’s quite encouraging that the pattern of the past few weeks, which has mostly been one of increases, has been replaced by one in which there isn’t clear evidence of changes in most measurements. Perhaps the recent increases have started to turn round – though it’s really far too early to conclude that from just one week of short-term trends.
“There is one particular piece of good news, though – the estimated percentage of children in England who would test positive for the virus that causes Covid-19, for the age group including school years 7 to 11 (secondary school age), shows very clear signs of a fall. This was the group that has had, by quite a long way, the highest rate of testing positive. That has been causing some concern, partly because of potential health effects on the children themselves (such as future long Covid), and partly because of the risk of them infecting people in more vulnerable age groups. The estimated rate of testing positive in the school years 7-11 group remains the highest of all seven age groups that ONS consider in England, with about 1 in every 15 children of that age testing positive and that’s not the best position to be in. But the estimated rate of testing positive in that age groups has fallen by about 18% – getting on for a fifth – in just a week, and according to ONS’s statistical model, it’s continuing to fall.
“ONS make an important point, in this report, about numbers of infections in school-age children. They point out that it’s too early to use their results to determine the impact of the half-term school holidays on infection rates. That’s because, though the most recent week in the survey was indeed the holiday week across most of the country, there’s a time gap, or incubation period, (of about a week, on average, or maybe slightly less) between someone being infected and the infection being detectable by a swab test. So most children who tested positive in the CIS in the most recent week would have been infected the week before the half-term holiday, and we won’t be able to see the effect of the holiday (if there was one) until next week’s CIS results at least. (I have to admit than I forgot about that time gap when commenting on some other data this week.) The same point will apply to the REACT-1 survey data published yesterday, which is also based on swabs mostly taken during the half-term week, so any infections would have started before that week. It also applies to other data sources, such as the new case counts on the dashboard at coronavirus.data.gov.uk, and the estimates based on the ZOE app, because mostly they will depend on people having and reporting symptoms, or (for the dashboard) taking LFD tests, and again it takes a time after the initial infection for symptoms to develop or for the virus to become detectable by an LFD.
“Overall, across the UK as a whole, the estimated number of people who would test positive (of all ages, 2 and above) has hardly changed at all, for the week ending 30 October, compared to the previous week. Adding up the ONS official reported estimates for the four UK countries for the latest week, the total is almost exactly the same as the previous week – under 1% lower, which is a change very much within the statistical margin of error. At least it didn’t go up. For England, the ONS estimate of the rate of testing positive in the latest week is 2.02%, or about 1 in 50 people testing positive, which is exactly the same (to the second decimal place of the percentage) as the estimate for the week before. ONS say, unsurprisingly, that the trend in the most recent week is ‘uncertain’, though overall it had risen over the two weeks to 30 October. It’s always more difficult for ONS to make clear estimates of the short-term trends in the other three UK countries than in England, because a lot fewer people are swabbed for the surveys in each of those countries than in England so the statistical margins of error are inevitably wider than in England. There have been some small movements in the estimated rates. ONS say that the trends are uncertain in Wales and in Scotland, but that rates of testing positive continued to increase in Northern Ireland.
“The estimated proportions testing positive in Wales, Northern Ireland, and Scotland, for the most recent week, are respectively about 1 in 40, 1 in 65, and 1 in 80. As in England (1 in 50), these are all high rates of positivity. In England, the rate is still the highest that it has been since the CIS began there at the end of April 2020, apart from just two weeks from 27 December 2020 to 9 January 2021, at the height of the second wave. In Wales, the estimated rate in the latest week is the highest it has been since the CIS began there in July last year, apart from the previous week when it was slightly higher (though the difference with the week before is well within the statistical margin of error). In both Northern Ireland and Scotland, the estimated rates for the most recent week are considerably lower than the peak levels.
“I’ve already discussed the good news that, in England, the rate of testing positive in secondary school children (school years 7 to 11) fell quite a lot in the latest week compared to the week before, and the not-so-good news that that age group still has the highest level of testing positive, about 1 in 15. Estimates for individual age groups are bound to be less precise, statistically, than for the whole population of England, simply because fewer people are swabbed for the survey in a single age group than for all age groups taken together. This means that it can be difficult to be certain about short-term trends in infection rates within age groups – though the downward short-term trend in that secondary school age group is really pretty clear statistically. In some other age groups, the news is not so good. ONS estimate that rates of testing positive increased for young adults (school year 12 to age 24 years) and for those aged 50-69, though the rates of testing positive in the most recent week in both of those age groups, about 1 in 70 and about 1 in 85 respectively, remain a long way below the 1 in 15 in younger secondary school ages. ONS say that the short-term trends in the most recent week in all the other age groups are uncertain, though the rate for those aged 70 and over did increase over the two weeks to 30 October. Positivity rates in younger children (age 2 to school year 6) remain the second highest of all age groups, at about 1 in 25. Even though ONS estimate that the positivity rate in those aged over 70 went up over the most recent two-week period, it is still relatively low at 1 in 110, though the rate for people aged 25-34 is similar (possibly slightly lower, in fact, at 1 in 120, though the margins of error are quite wide so we can’t be sure). The highest estimated rate for any age group above school year 11 is for those aged 35-49, with a rate of about 1 in 55. It’s been suggested that the relatively high level, compared to other adult groups, is because many in that age group will have school age children and could have been infected by their children, but the CIS data can’t confirm that directly.
“ONS also provide estimates for the nine regions of England. Given that, overall, the infection level in England didn’t change in the most recent week compared to the week before, it’s hardly surprising that the changes in different regions are different mostly not very clear. Overall, ONS say that the percentages testing positive “continued to fluctuate across regions in England” in the most recent week. That’s a pretty accurate summary – some up a bit, some down a bit, but not really any major changes. The overall position in the most recent week is that rates of testing positive are again highest in the South West, with about 1 in 35 testing positive, and lowest in London, about 1 in 65, but there’s quite a lot of overlap in the statistical margins of error between different regions.
“Some other sources of estimates of infection levels in the UK, or parts of the UK, have shown decreases over roughly the same period as the most recent week of CIS data. That was the case for the REACT-1 survey’s interim preprint (published yesterday just after midnight), for instance. REACT-1 covers only England – its results are also based on swab tests for a representative sample of the community population, though the list of people from whom the survey sample is drawn is different and the details of the survey procedures are quite a bit different. REACT-1’s latest figures cover swabs taken in the period 19-29 October, so overlapping with this latest CIS report, and the REACT-1 researchers presented some evidence of falling rate of infection in the later part of that period. That evidence comes from a statistical model, which is subject to quite a lot of uncertainty because the daily numbers of swabs tested is relatively small. The ONS CIS also produces daily estimated positivity rates from a (different) statistical model, and they do also show falling rates in England, from a peak around 22 or 23 October, though the fall is not very rapid. However, ONS are (rightly in my view) rather more cautious about interpreting the details of these modelled results, given the margins of error, and that’s why they say that “the trend was uncertain” in England in the most recent week. But, broadly, the two surveys are saying similar things about trends in England.
“There have also been recent falls in the numbers of new confirmed cases on the dashboard at coronavirus.data.gov.uk, and on new infections according to the researchers behind the ZOE app – though both those sources look at new infections, while the CIS (and REACT-1) instead measure the numbers who would test positive, regardless of whether their infection is new or has been going on for a few days (or longer). So they aren’t entirely comparable with the survey results.”
Prof Jim Naismith, Director of the Rosalind Franklin Institute, and Professor of Structural Biology, University of Oxford, said:
“The data from ONS cover the period to 30th October. The prevalence in England (1 in 50) does appear to have peaked as last week’s data indicated. The ONS is careful to state the trend is uncertain. I would now expect all other things being equal for the prevalence to fall.
“There are other encouraging results the prevalence in teens shows a trend downwards from very high levels. This age group had very alarming levels of infection and this was driving case numbers.
“The decision to allow such high levels is quite rightly taken by elected representatives not scientists. In time, the cost (health complications) benefit (less restrictions) ratio of the this policy will be evaluated.
“The combination of vaccine roll out to this age group and the high level of infection, will drive down the numbers of newly infected teens rapidly from this point.
“Some areas of the UK, notably London and East Midlands appear to have peaked at lower levels than others.
“In the most vulnerable group 70 + prevalence is very low, the booster program may be having an effect.
“I will be taking my booster as soon as I possibly can, data suggest this will give me almost complete protection against serious disease.
“There may also be some encouraging news from Scotland, the prevalence at 1 in 80 is really the same as last week. The ONS points out the trend is uncertain, but there is no evidence of a climb. I am relieved at this good news, I would therefore expect prevalence in Scotland to continue to fall and England to follow behind. Data show that universal masking and social distancing will accelerate the fall in prevalence. The comparison of the case trajectory between Scotland and England would indicate that subtle differences in protocols have no effect on the virus, despite their political appeal.
“It should be borne in the mind the UK has new daily hospitalisations of over a thousand and a thousand people a week dying.
“These deaths will have hit people hard, the virus remains a terrible killer and I send my condolences for the loss.
“As a matter of the policy, the UK’s nations have since the summer operated with a high level of prevalence since the summer.
“Consequently we have had a much higher burden on the health care system than almost any other country, with the notable exception of the USA.
“The high level of prevalence has of course fed through to a higher death toll when compared to similarly vaccinated countries. I do not know whether this strain on the NHS will have any consequence in January.
“There is no doubt that this Christmas will be transformationally different than last. Last year we faced a peak around 1300 deaths a day by mid January.
“We only avoided a much higher death toll by locking down in early January and in effect partial lockdown mid December.
“There is simply no better evidence for the vaccine than this change our fortunes.
“I am encouraged by the licensing of the new drugs and treatments.
“I am not sure celebration is appropriate; 164 000 people are dead from covid19 in the UK.
“This is around half the British military dead of WW2 who we rightly remember next week.
“It is over twice the number civilian dead from that conflict.”
Dr Simon Clarke, Associate Professor in Cellular Microbiology, University of Reading, said:
“Today’s weekly data from the Office for National Statistics show that coronavirus infections remain at one of the highest levels they have been across the UK, since the start of the pandemic.
“The ONS survey is a much more reliable source of information than many apps and other sources of infection data, due to its high sample numbers and consistent statistical approach.
“It shows that in the past week well in excess of a million people have been infected, many of whom will have been walking around in their daily lives, spreading the disease without realising it before developing symptoms, or without being affected at all.
“While there may be a slight downturn in the figures compared to the previous week, it is too early to know if this will be the peak of the latest wave of infections or just a temporary week to week reduction. The slightly lower numbers overall are likely due to a reduction in cases among children, which may be an impact of the autumn half term school holiday, and the rising numbers of vaccinated teenagers.
“It is concerning that infection levels in older people could be on the rise. This may be a result of waning immunity, which would emphasise the importance of the current booster jab campaign.”
Prof Paul Hunter, Professor in Medicine, UEA, said:
“The latest report from the ONS provides some reassurance that covid infections in England may have indeed peaked and supports what has been seen in the daily case reports which have been falling for a couple of weeks now. Because this report presents prevalence and people often remain positive for more than a week after becoming infected it will always look like data here are falling more slowly than seen in daily case reports which are incidence data.
“Of particular interest is the decline in children in school years 7 to 11. Estimated prevalence in this group peaked on the 21st October and so started falling before most schools broke up for half term which would give hope that cases in this age group may not start to increase again now that schools are back. We will have to wait another one or possibly two weeks before we can be certain.
“It is too soon to see any effect that the booster vaccination campaign may have on case numbers, we may have to wait another couple of weeks before we can see any clear impact of booster vaccinations.”
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.