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 James Naismith, Director of the Rosalind Franklin Institute, and University of Oxford, said:
“The ONS survey today is good news. England with roughly 2 in 1000 people positive appears now to be following Scotland, which also has dropped again.
“Encouragingly the prevalence appears to have fallen everywhere except perhaps the North East where it may have levelled off but this uncertain.
“There is other good news, prevalence in school children is falling. In the most vulnerable age group over 70, less than 1 in 1000 are infected.
“The lockdown has worked as expected as has the vaccination campaign.
“Vaccination is highly effective against the most common UK strain and will give some protection against any of the rare new variants, therefore more people who are vaccinated here and overseas, the less chance we give the virus to mutate.
“The UK sequencing efforts are actually world leading, these are helping to keep us safe.
“I would hope the public will support continued investment in scientific research to develop new therapies. Vaccines have been enormously successful but having an effective collection of therapeutics ready to deploy is an insurance policy for the world.
“As new variants enter or arise in the UK, a robust testing system is vital. Thus far the new variants (like the original strain) are relatively harmless to and asymptomatic in the young; this makes management much harder. The problem is compounded by the increased speed of spread of the new variants.
“As prevalence continues to fall, the at home testing regime with lateral flow will increasingly identify false positives. Some thought needs given to how this problem can be solved. The key as has been true from the start is to rapidly and cheaply identify true positives, support their effective isolation and break the transmission chain. True negatives (but false positives) need to be allowed to resume normal life as quickly as possible.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This week’s regular bulletin from the ONS Infection Survey is pretty well unremittingly good news. The latest data on prevalence, that is, the numbers of people in the community who would test positive if tested for an infection with the virus that can cause Covid-19, is for the week ending 10 April, and it shows falls in the numbers infected in all four of the UK’s countries, compared to the previous week. It’s several weeks since that last happened – of course there have been very large falls in all four countries since the peak of the most recent wave at the start of the year, but for the past five or six weeks, the decrease in infections had slowed, and generally one or more of the four UK countries was showing signs of a small increase compared to the week before. But for this latest week, the ONS’s estimates show decreases in all four. The estimated size of the latest decrease varies from one UK country to another, but ONS rightly warn about over-interpreting small differences. All estimates based on surveys have a margin of error, and as the infection rate gets lower and lower, the width of that margin of error can be quite large relative to the estimated rate. Now, in England, Wales and Northern Ireland, the percentage of people who would test positive is down to the level from roughly mid-September. I can’t make such a comparison for Scotland, because the ONS survey was not yet publishing estimates in mid-September. Really the only tiny cloud in these data, however, is that rate of testing positive in England is still about four times what it was in August last year, and the rate now is about double the August rate in Wales. (The Northern Ireland and Scotland data from this survey don’t go back that far.)
“This margin of error does mean that we can’t be entirely sure that the national infection rate varies from one country to another. ONS’s central estimate is that about 1 in every 480 people would test positive in England, about 1 in 920 in Wales, about 1 in 710 in Northern Ireland, and about 1 in 500 in Scotland. But the margins of error around those estimates are quite wide, and particularly so in Wales and Northern Ireland where the populations are smallest, and the number of people swabbed for the survey are smallest. So ONS report that, for England, the number infected could plausibly be anywhere between 1 in 420 and 1 in 560, while in Wales it could plausibly be between 1 in 520 and 1 in 2,080. Those ranges overlap, so although it’s pretty likely that the infection rate in Wales is considerably lower than in England, we can’t be sure. Such are the limitations of survey data.
“What I can say, though, is that for the most recent week, ending 10 April, the ONS estimates for the four countries add up to a total of 129,000 people testing positive in the whole UK. The week before, that number was 184,800 – so the number has gone down by almost a third (30%) in a week. There is statistical uncertainty about the exact size of that decrease, but it is certainly very substantial. This is for the week after Easter, after some loosening of lockdowns had occurred and that might have risked an increase in infections – but, potentially working in the other direction, schools were on holiday so that there would have been fewer social contacts there.
“The issue of statistical uncertainty applies also to trends and comparisons in the English regions, and for different age groups. Obviously the number of people swabbed for the survey in a single region, or a single age group, is considerably less than for the whole country, so the statistical uncertainty is greater than for the whole country. This means that ONS, rightly, have to be cautious about interpreting these trends and differences. They report that the percentage testing positive did decrease in two regions, Yorkshire and the Humber and the South East, in the week ending 10 April compared to the week before, but that the trend is uncertain in the other regions. There do remain some quite large differences between the infection rates in the different regions, and some of those differences are large enough so that they can’t be explained away by statistical variability issues. For instance, ONS estimate that the positivity rates in the southern regions of England apart from London (South East, South West, East of England) are all roughly 0.1%, meaning that roughly 1 person in every 1,000 would test positive. In the North West, the estimate is that about 1 in 260 would test positive, so about four times the rate in those southern regions, and there’s pretty clear evidence that the rate is really higher in the North West despite any statistical uncertainty. The rate in the North East, estimates at 1 in 310, also definitely looks to be higher than in most of the regions in the south. On trends and differences between age groups in England, ONS report that rates have decreased in all age groups except in people of secondary school age, and those aged 50-69, where the trends are uncertain. It remains the case that positivity rates look lower in most older age groups than in school age children, but the margins of error are wide.
“There’s more good news from the ONS estimates of incidence – that is, of the daily numbers of new infections (as opposed to the prevalence estimates, which cover all people who would test positive regardless of whether the infection had just happened). The incidence estimates are, as always, a week behind the prevalence estimates, so the latest week is the week ending 3 April. For that week, ONS report that the rate of new infections was lower than in the week before in three of the UK countries. The exception is Wales – the incidence estimate did fall there too, but not by as much as in the other three countries, and the margins of error are wide, so ONS say that the incidence rate in Wales may well be level.”
Further information
“ONS have stopped producing estimates of the infection rates in areas of the UK smaller than the four countries and the English regions. These haven’t appeared for three weeks now, and ONS have said that they need to revise their estimation methods because the infection rates are so low now. Up till this week, the data sets release with these bulletins said that new estimates for smaller areas would appear today, but they have not appeared today, so it appears that more work needs to be done on this part of the statistical modelling. In any case, even before ONS last published these estimates on 26 March, the margins of error on the estimates for the smallest areas and the areas where the estimated positivity rates were lowest were very wide, and really didn’t tie the numbers down enough to be useful.”
All our previous output on this subject can be seen at this weblink:
www.sciencemediacentre.org/tag/covid-19
Declared interests
Prof James Naismith: I have had the first dose of AZ/Oxford and am taking part in the test at home scheme.
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.