February 5, 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 Kevin McConway, Emeritus Professor of Applied Statistics at The Open University, said:

“The latest bulletin on the ONS Infection Survey takes the information forward another week, to the week from 24 to 30 January. The picture, overall, looks considerably more encouraging than it did last week, though not everything in it is positive. The percentage of people in the community populations of the four UK countries who would test positive remains high, though ONS estimate that it is falling in three of the UK countries (but not in Wales, where ONS report that it remains level).

“In England, ONS estimate that about 1 in 65 people would test positive for the virus that causes Covid-19. There’s a bit of statistical uncertainty about that, as is inevitable with estimates from a survey – ONS say that the true number is very likely to be between 1 in 60 and 1 in 70. This is certainly an improvement since the estimates around the turn of the year, where ONS were estimating that about 1 in 50 people in England would test positive. But it’s still high – roughly the level of infection from the week before Christmas, when the concern was so great that tight restrictions were introduced in much of the country, followed by a national lockdown. It’s great to see, this week, that there’s clear evidence that the infection rate is falling, but it needs to fall a lot further.

“In Wales and in Northern Ireland, the proportion of people who would test positive is not hugely different from the position in England. ONS estimate that 1 in 65 would test positive in Northern Ireland, and 1 in 70 in Wales. There’s more statistical uncertainty about these figures than those for England, because they are based on smaller numbers of people because of the lower population – but the position appears to be more concerning in Wales than in Northern Ireland, because there’s no clear sign in Wales that infections are falling. In Scotland, the proportion is quite a lot smaller – 1 in 115 – and ONS report that it is falling (though rather slowly).

“I’ve pointed out in previous weeks that ONS actually produce three different sets of estimates for the positivity rate. They don’t always show the same trends, particularly if the rate of infections is turning from an increase to a decrease, or a decrease to an increase. But this week, they do all show much the same picture. In England, the number of people who would test positive is about 17% lower in the latest week than in the week before. That’s a pretty substantial fall. It is, however, not as fast as the fall in the numbers of confirmed cases on coronavirus.data.gov.uk, which was about 30% over the same period.

“I don’t think that’s a matter for concern, though. Broadly speaking there are two reasons why such a difference might happen. First, the numbers of confirmed cases on the dashboard depend on who decides to come and be tested, and how they are tested, and those things can change over time. A particular potential issue now is that an increasing number of the tests for Covid-19 are rapid tests using lateral flow devices (LFDs – the things that look vaguely like home pregnancy tests). LFDs are being used for mass testing of various kinds, often on people with no symptoms. Also they have different statistical properties (rates of false negatives and false positives) than do the lab-based PCR tests. So at a time when more LFDs are being used, it’s possible that the trends in the confirmed cases don’t track the trends in actual infections as well as they sometimes do. The ONS Infection Survey tests a reasonably representative sample of people across the country, with the sole purpose of tracking infection rates, and uses only PCR tests, so this issue does not affect these ONS results.

“But probably the second reason for a difference in trends between the ONS infection survey results and the confirmed case counts is more important. It is that they are not measuring the same thing. The confirmed cases counts are counts of new infections – they measure what is called the incidence of infection. The ONS infection survey (and the REACT-1 survey) estimate the total number who would test positive on a day in the week in question. That would include people who would have had their first positive test that day, and wouldn’t have showed up as being infected the day before, but it also includes people who might have been infected a considerable time before, but would still testing positive because the virus is still detectable in their bodies. This is called the prevalence of infection. There’s no reason why the incidence and prevalence would exactly go up or down together at the same speed – indeed it would be unusual if they did. According to one piece of research, on average people test positive for a period of 17 days, and many can test positive for a lot longer than that (while others would test positive for a shorter time). So the estimated count of people who would test positive in the most recent week of the ONS survey, 24 to 30 January, would include some people who were first infected around the turn of the year, or even before that, when the incidence of infections (as measured by the confirmed cases) was going up rather than down. Because of this, the difference between the trends in the two rates doesn’t concern me – and they are both falling substantially anyway, even if they do still have a long way to go.

“As always, the ONS survey has provided evidence of quite large differences in infection rates between different regions of England, and indeed different trends in different regions. Across the regions, the estimated positivity rate is highest in London, where an estimated 1 in 50 people would test positive in the most recent week – but that rate continues to fall quite rapidly from its peak around Christmas. Rates are also fairly high – 1 in 55 or 1 in 65 – in several other regions – North West, East Midlands, West Midlands, and the East of England. They are lower, but still at levels that would have caused considerable concern a couple of months ago, in the North East, Yorkshire & The Humber, the South East, and the South West. ONS estimate that positivity rates are decreasing in five of the nine regions (London, the South East, North West, North East, and the South West), roughly level in another three (West and East Midlands, and Yorkshire & The Humber), but worryingly they estimate that the rate is increasing in the East of England. The rate of increase there is not rapid, but it’s concerning nevertheless.

In terms of age groups, ONS estimate that (on average across England) infection rates are decreasing in all the age groups they look at, except for the 70+ group where the rate seems to be level. But the rate for the 70+ group is the lowest across all the age groups they look at, with about 1 in 115 testing positive, and the people in that group will very soon be protected from serious illness by vaccination, so the fact that the rate isn’t decreasing for that group doesn’t concern me.”

Dr Simon Clarke, Associate Professor in Cellular Microbiology, University of Reading, said:

“As tempting as it is to find encouragement in infection rates that are falling for England as a whole, we still see very high levels of Covid-19 in every region.  The Office for National Statistics estimates that, across the UK, nearly 1 million people in the community, which does not include people in care homes or hospitals, had Covid-19 in the week ending 30 January 2021.

“The lockdowns across the UK appear to be arresting the spread of Covid-19, but with seemingly mixed success.  Infection numbers decreased in England, Scotland and Northern Ireland but remained stable in Wales.  In the East of England, infections actually increased and remained stable in the East and West Midlands, with decreasing numbers in every other region.

“In England, infection numbers decreased in all age groups except the over 70s, who’ve been the target of the vaccination campaign.  It will be both worrying and disappointing if this is because people are letting their guard down as soon as they’re vaccinated.  It remains important to leave enough time for immunity to develop, and to have the second dose.  But if it’s not the case that they’re behaving any differently, we can expect to hear calls for more studies into the protection that Covid-19 vaccines offer older people.

“It will be essential to have this sort of full understanding of the nationwide picture of infection before considering how and when to lift lockdown.  If an overall national average masks high numbers in some parts of the country, the virus will spread causing another wave amongst those who have not yet been vaccinated and for those in whom the vaccine fails to work.  We may be over the peak of this wave, but there is still far too much coronavirus out there to loosen restrictions too quickly, or we will be right back to where we started again.

“The ONS study has recently increased in sample size and the data continue to be reliable for comparative purposes and represent a more comprehensive picture than data which relies on self-selection by its participants.”

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

“With thousands of deaths a week, I would like to express my sympathy for the families and friends left bereft.

“Today’s ONS survey results are somewhat better news.  The incidence rate has resumed its decline: between 1 in 60 and 1 in 70 people are infected in England, while Scotland has notably lower prevalence of 1 in 115 to 1 in 135.  These still represent very high rates of prevalence.

“The roll out of the vaccine will not yet have impacted upon these figures – the decrease is due to lockdown. However, the incredible progress being made in the vaccine roll out is to be celebrated as this will greatly reduce serious illness and deaths.

“The new variant looks to be responsible for the majority of infections across the UK, and the speed at which it has spread is striking.  I greatly fear for countries that lack the UK’s resources.  The UK support of CoVax deserves praise, as we will only be safe when everyone is safe.  Allowing the virus to rage unchecked anywhere or in any population risks new mutants developing.  We now know how fast such variants can spread.”

All our previous output on this subject can be seen at this weblink:

www.sciencemediacentre.org/tag/covid-19

Declared interests

Prof 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.