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 latest bulletin from the ONS coronavirus infection survey (CIS) is uninterrupted good news. The estimates of the numbers of people who would test positive for a current infection with the virus that can cause Covid-19 now go up to the week of 2-8 May. ONS estimate that the numbers have fallen in each of the four UK countries, compared to the previous week, and the rates are now very low in all four countries. In England, the numbers are just slightly above the lowest level they reached in this survey, in mid-August last year. Back then, ONS estimated that 1 in 2,200 people in the community population in England would have tested positive. The latest estimate is 1 in 1,340 and falling, so not much higher than that. In Wales, the estimate for the latest week is that 1 in 4,230 would test positive. That’s a very low rate, well below the previous lowest level of 1 in 2,800 last August. I can’t make those comparisons for Scotland or Northern Ireland. The latest numbers there are really low too (1 in 1,250 and 1 in 1,430 respectively), and in each case that’s the lowest level since the survey began in those countries – but the CIS wasn’t yet running there last summer when infection levels were previously very low.
“All this does indicate that vaccination and other measures have got us into the stage of the pandemic where there will continue to be local outbreaks that need to be dealt with quickly, but that sweeping national interventions such as increases in restrictions may well no longer be needed. But that all does depend on the continued success of the vaccination programme, on no really serious new variants emerging that can’t be dealt with by local measures and vaccinations, and on generally taking care and continuing to look at the data. These latest CIS results can’t say anything about what might happen as a result of the next stage in the English roadmap on Monday, let alone the final stage that’s currently planned for 21 June. There will inevitably be bumps in the road – I just hope they don’t throw us off course.
“Because the infection levels are so low and come from a survey, the estimates can’t be very precise. We can be very confident that infection rates are low, but exactly how low is difficult to estimate, and the margins of error are very wide. The most extreme case is for Wales – the central estimate is that 1 in 4,230 would test positive, but the data are compatible with anything between 1 in 1,580 and 1 in a mammoth 22,280 testing positive. Even if it were 1 in 1,580, though, that’s still a very low infection rate. And it’s not surprising that the rates can’t be estimated really precisely – in Wales in the most recent two-week period, over 8,000 people were swabbed for this survey, but only two of them tested positive. You just can’t get very precise results from just two positive tests – but you can be certain that the infection rate is low. Estimates are a bit more precise in England because more people are swabbed for the survey – in the most recent two weeks, more than 143,000 people were swabbed in England, but only 105 of them tested positive, so there’s still quite a lot of uncertainty. The data in England are compatible with anything between 1 in 1,070 and 1 in 1,710 testing positive. But with rates this low, it really doesn’t matter that we can’t know them very precisely.
“Similar issues arise with the data on individual regions of England, and on age groups in England. Obviously the number of people tested in a single region, or a single age group, has to be smaller than for the whole country, so the estimates for a region or age group won’t be so precise. With infection levels this low, it’s difficult to be clear on short-term trends for regions or age groups, and the ONS report has to say in several places that trends are uncertain. But in every region and every age group they estimate either that numbers testing positive are decreasing or that the trend is uncertain. They don’t estimate increases in any English regions or age groups.
“The ONS CIS estimate, that 1 in 1,340 people in England would test positive in the week 2-8 May, is a bit lower than the estimate published by Imperial College’s REACT-1 survey* earlier this week of about 1 in 1,000. However, the REACT-1 estimate is based on data collected between 15 April and 3 May, so rather earlier than the latest ONS figures. The REACT-1 estimates are close to the ONS CIS estimates for the same period, and since all the evidence is that infection rates are continuing to fall, you’d expect the latest CIS estimate to be lower than the latest REACT-1 estimate.
“ONS usually also estimate the rate at which new infections are occurring. (The figures I’ve mentioned so far are for all infections, whether they have just begun or not.) However, they have not published the estimates for new infections this week, because of difficulties in making these estimates when infections are so low. Obviously the number of new infections is less than the number of all infections, so the estimation is even harder for new infections.”
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.”