select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
The government have released the latest estimates for the COVID-19 R value and growth rate.
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“SAGE and the Government Office for Science (GOS) have published their weekly ranges for the R number and for the growth rate of the COVID-19. The ranges for the UK are higher than they were last week, which in turn were higher than the week before. The UK R number range is 1.3 to 1.6, compared to 1.1 to 1.4 last week, so there’s no real doubt that it is now above 1 and therefore that the epidemic is continuing to increase. The growth rate range runs from +5% to +9%, indicating that GOS and SAGE believe that the number of new cases is growing by between 5% and 9% each day. (Last week’s range was from +2% to +7% each day.)
“A daily growth rate of 5% corresponds to a weekly growth rate of about 40% (higher than 7 times 5 because each day’s increase is a percentage of the previous day’s already increased figure – a bit like compound interest). A daily growth rate of 9% corresponds to a weekly growth rate of about 70%. So the range given for the growth rate is not as high as a doubling of new cases in a week, but it’s still fairly high. These growth rates correspond to a doubling time between about 14 days and about 9 days.
“Ranges for R and the growth rate are also published for England, and for the English regions. For England as a whole, the R range is 1.2 to 1.6, and the growth rate range is +4% to +8%. These are slightly below the UK figures, though we can’t be entirely confident that infections are increasing more slowly in England than in the other UK countries, because there is considerable uncertainty about the exact numbers. The ranges for the separate English regions all indicate increasing numbers of infections, though at rather different rates in different regions.
“The fact that the ranges for the R number and the growth rate are higher than last week does give a different impression than some points that were made in the interim results from Round 5 of the REACT-1 infection survey study published yesterday1 and today’s update on the ONS infection survey2. The REACT-1 report estimated R, using just data collected between 18 and 26 September, as 1.06, which is below the SAGE/GOS range. But there is a great deal of statistical uncertainty, so that the REACT-1 researchers give a range for their R estimate of 0.74 to 1.46, which does overlap with the SAGE/GOS range for England. (REACT-1 covers England only.) The range for the growth rate published in the REACT-1 report was from -4% to +6% per day, a very wide interval (including the possibility that the number of cases is decreasing), but one that overlaps with the SAGE/GOS range. These REACT-1 estimates are based on a limited amount of data from just one source, while the SAGE/ONS estimates are based on a considerably wider set of inputs. But nevertheless REACT-1 did conclude yesterday that there was some evidence that the rate of increase of infections was slowing. The ONS survey does not estimate R, but today’s ONS report suggested (tentatively) that the recent rise in the number of new infections might be levelling off.
“So how can it be that the SAGE/GOS estimates of R and the growth rate point to a possible acceleration in the speed at which the epidemic is growing in England (and indeed in the UK as a whole), while REACT-1 and the ONS infection survey are both reporting some evidence, albeit not very clear evidence, that epidemic is growing more slowly now than in previous weeks? Well, there’s quite a lot of statistical uncertainty about all these estimates. And it’s important to point out that none of these reports are saying that numbers of infections are now shrinking – only that the numbers may be growing more slowly than they were a couple of weeks ago. (To be precise, the REACT-1 interim report’s estimates do include the possibility that the number of infections is falling, but that isn’t their main conclusion, and there is a great deal of uncertainty about their estimates. We’ll possibly know more next week when REACT-1 publish their report on the whole of their latest survey round, and not just on part of it.)
“But one potentially important point about the SAGE/GOS ranges for R and the growth rate is that the models behind these estimates use a range of data, including numbers of hospital admissions, ICU admissions, and deaths. For those data sources, patterns of increase or decrease are delayed compared to changes in new infections, because it takes time after someone has been infected before they may need hospital treatment or, sadly, die from their infection. The current R and growth rate estimates cannot have fully taken changes in those figures into account, because some of the changes have not yet occurred. So it could be that the estimation process for the SAGE/GOS estimates hasn’t yet been able to catch up with very recent information from the two infection surveys on changes in case numbers. It could be that the number of new cases is increasing more slowly than it was a couple of weeks ago. But we absolutely can’t be certain of that yet, and we must remember that REACT-1 and ONS are still concluding that infections are increasing rather than decreasing. Measures to reduce the rate of increase continue to be vitally important.”
https://www.gov.uk/guidance/the-r-number-in-the-uk
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 the Advisory Committee, but my quote above is in my capacity as a professional statistician.”