select search filters
briefings
roundups & rapid reactions
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:
“Today’s bulletin on the ONS Infection Survey takes the information up to the week ending 6 March. Since that’s just before the reopening of English schools on 8 March, it can’t yet tell us anything about any impact of that (or of the other loosenings of lockdown this week) on infections. Overall, the news on the latest trends is still reasonably good, though not quite as encouraging as last week.
“In England and in Wales, the estimated percentage of the community population that would test positive for the virus that can cause Covid-19 continued to fall in the most recent week, compared to the week before. The trend over that week in Northern Ireland and in Scotland is less certain, and ONS say that the decline, that had been taking place since peak infection levels around the start of the year, has levelled off in those countries. That doesn’t sound so good, and in fact in England and in Wales there are some signs that the decrease in the most recent week was a bit slower than the week before – but still pretty clearly a decrease in those two countries. It’s important not to over-interpret these signs of a slowing in the downward trend, though. First, the decline since the peak of infections has still been very marked in all four countries. In England, Wales, and Northern Ireland, the rate of testing positive is now between a seventh and a fifth of the peak rate around the start of the year. In Scotland, the decline is rather smaller in those terms – it’s now between a quarter and a third of the peak rate – but the peak level of infection was lower in Scotland than in the other three UK countries, so there wasn’t so much room for a decline. Second, there are quite wide margins of statistical uncertainty about these estimates in the case of Scotland and, particularly, Wales and Northern Ireland. That’s because the numbers of people who provide swabs to be tested for the survey are considerably smaller in the other countries than in England, which is simply because their populations are smaller. So the trends are less clear in countries other than England.
“The upshot of all this is that the rates of testing positive are more similar across the four UK countries than they have been in the recent past. In England, ONS estimate that 1 in 270 people in the community would test positive for the virus in the week ending 6 March, down from 1 in 220 the previous week. But allowing for the statistical uncertainty, the England data are compatible with an estimate for the latest week of between 1 in 245 and 1 in 305. For Wales, the range is 1 in 255 to 1 in 560; for Northern Ireland, 1 in 200 to 1 in 540; and for Scotland, 1 in 235 to 1 in 450. These ranges all overlap quite a lot. So, although in terms of the central ONS estimates, the infection rate is lowest in Wales, a little higher in Scotland and Northern Ireland, and a little higher again in England, these estimates are now so close together that we can’t really be certain the infection rates differ between countries at all.
“What we can be pretty certain of is that the rates are now roughly at the level where they were at some point in the second half of September or the start of October (depending on the country), but still a good deal higher than they were in England and in Wales in July and August last year. (We can’t make comparisons so far back in Scotland or Northern Ireland, because the ONS survey was not running in those countries then.) We do have vaccines now, and we didn’t back then, so infections will be less likely to lead to severe Covid-19 illness or death. But not everyone can be vaccinated, and the effect of vaccines on things like long Covid is not yet known, so we do still need to be cautious. ONS estimate, for the most recent week, that nearly a quarter of a million people in the community population of the UK (231,400 people) would test positive for the virus. That’s very roughly the population of Southampton or Derby, and not much less than the population of Newcastle. There will be more infected people who are not counted in these estimates because they live in communal places such as care homes or prisons (though those numbers will not be large compared to quarter of a million). So there’s still quite a lot of infection about.
“For the English regions, the position varies. ONS conclude that infections are continuing to decrease in five of the nine regions, and in London the decrease has been pretty rapid. If I don’t concern myself with the margins of error, the rate of testing positive is now lower in London than in any of the other regions, even though six weeks ago the infection rate was highest in London. But ONS say there are some signs of a possible increase in the rate of testing positive in the South East and the South West, and the trend is unclear in the West Midlands and in Yorkshire and the Humber. The margin of statistical uncertainty is quite wide for the regional estimates, though, because obviously fewer people are tested in each region than for the whole country. The intervals that ONS use to show the margin of uncertainty all overlap, so the actual order in which the different regions come in this depressing league is far from clear. It doesn’t trouble me too much that the trends in different regions are different, though. As infection rates come down, the infection rates for a region can be changed quite a lot by one or two fairly local outbreaks, that might be quickly contained, so I’d expect some variability.
“In terms of the positivity rates in different age groups in England, ONS report that they have decreased in all age groups except in children of primary school age, where the trend is uncertain. It would be good to see some effect of vaccination on the rates for different age groups, but I don’t see one. There’s no evidence that infection rates have fallen faster in the older age groups where more people have been vaccinated. The rates of decrease in those aged 25-34, 35-49, 50-69 or 70+ do not differ very much, even though only a small proportion of the people in the younger two of these age groups would have been vaccinated. In recent weeks, infections have fallen more slowly in people up to age 24 than in these other groups, but that can’t have anything to do with vaccination. But we must remember that these ONS results say nothing about how ill the infected people are, if indeed they have symptoms at all – older people’s health may well be much better, on average, because of vaccination.
“One final note – this week ONS have added a very nice animation feature to their so-called sub-regional estimates for smaller areas of the UK. I could play with that for hours…”
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.”