The government have published the latest estimates for the COVID-19 R number and growth rates.
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This week’s revision from SAGE of the R number for England is, as usual, expressed as a range of possible values, and this week, for the first time since mid-January, that range goes above the threshold value of 1.0. It runs from 0.8 to 1.1 – last week the range was 0.8 to 1.0. This does not mean that R is definitely now above 1, which would mean that numbers of infections would tend to increase. Instead it means that SAGE say that it’s possible that it is above 1. The fact that the range is wider than it was last week means the SAGE are less certain that they were last week about the actual value of R. And, though the whole interval is no longer below 1, more of it is below 1 than above 1, which broadly means that SAGE believe it’s more likely to be below 1 than above 1.
“Though this isn’t the best news, I think there are several reasons why it’s not hugely concerning. First, as SAGE say explicitly, this range (and the range for the growth rate) “represent the transmission of COVID-19 2 to 3 weeks ago, due to the time delay between someone being infected, developing symptoms, and needing healthcare.” That delay is because the estimates are based on a range of data, including test results, hospitalisations, deaths, contact surveys, infection surveys, and more, and changes in many of these take time to occur after someone is infected. More recent data on some aspects have become available since then, most recently from the ONS infection survey results today, which showed infections generally falling. I’m not saying that these survey results trump the R range revision, but they were falling 2 to 3 weeks ago and are still falling, and indeed falling more rapidly than they were two or three weeks ago. Second, even though it’s still possible that infections may rise (even if they aren’t rising just yet), the impact of any rise on healthcare and on people who are unfortunate enough to be infected will be much less than it was, because of the success of the vaccination programme and the strong evidence that vaccines substantially reduce serious illness, hospitalisation and deaths. Third, though I’m not personally involved in producing estimates like these, I do wonder whether the models that are used have all been able to keep up well with changes in the relationship between numbers of infections, hospitalisations, deaths and other measures, given the changes in patterns resulting from vaccination. We’ll find out more about that in the near future. I very much hope I’m not being inappropriately optimistic here, but I’d need to see more data and wait a bit longer before I’d be really concerned about the upper end of the R range going above 1. Anyway, it’s a only a possibility that R might be above 1 (or that it was above 1 two or three weeks ago), not a certainty. No cause for panic.
“The growth rate range is, in my view, easier to understand anyway. This week’s range for England runs from -4% to -1% per day. A negative growth rate is actually indicates that new infections are decreasing, so this week’s range means that for every 100 new infections on one day, the number on the next day will be somewhere between 96 and 99. Last week’s range was -5% to -1% a day, and this week’s is slightly narrower, indicating that SAGE are slightly less uncertain about the exact growth rate. But, again, the range does not go above 0 into positive numbers, so SAGE are saying that currently (or at least 2-3 weeks ago), new infections are decreasing, on average across England. That is in accord with what we’ve seen from the ONS infection survey and with confirmed cases on the dashboard at coronavirus.data.gov.uk. A growth rate of -4%, if it continued for a time, would see the number of new infections halving in about two and a half weeks, while a rate of -1% would see infection halving in about ten weeks. Today’s ONS Covid-19 Infection Survey results for England estimated that the number of new infection in England roughly halved between the week ending 9 April and the week ending 17 April (though there’s quite a lot of statistical uncertainty about that figure) – that’s a daily decrease of something like 9% or 10%, so a bigger rate of decrease than the lower end of SAGE’s growth rate range. I’m not saying that the ONS figure is more accurate – the SAGE figures are based on a wider range of data – but this does illustrate that these quantities are not easy to estimate.
“As always, SAGE also provide ranges for R and the growth rate for the NHS regions of England. Several of them have changed since last week. The R ranges for three of the seven NHS regions now have an upper end above 1.0, and two of the growth rate ranges (for the East if England, and the South West) now go above 0, indicating that SAGE think it’s possible that new infections were increasing there (2-3 weeks ago). But in both those cases, the growth rate ranges go from -4% to +1%, so the bulk of the range is below 0, indicating that a falling number of new infections is more likely than a rising number. So again, this isn’t the best news, but I don’t find it particularly concerning, for the same reasons that I don’t find the revised ranges for England particularly concerning.
“A technical point about the regional ranges is that SAGE have changed their rule for deciding when estimates might not be reliable enough. It used to be based on numbers of deaths; now it’s based on numbers of new Covid-19 cases in hospitals. In recent weeks, SAGE have been flagging most of the regional estimates as unreliable – this week, none of them are flagged, though I suspect that many would have been flagged under the old rule. I think it’s probably sensible to have changed this rule – though it does again draw attention to the fact that the relationship between numbers of infections and numbers of deaths has been profoundly changed by the vaccines.”
All our previous output on this subject can be seen at this weblink:
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.”