select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
The government have released the latest estimates for the COVID-19 R value and growth rates.
Comment also on other estimates of COVID-19 R value and growth rates:
Prof James Naismith, Director of the Rosalind Franklin Institute, and University of Oxford, said:
“Over the last week, the UK has averaged over 1000 deaths per day. I wish to extend my deepest sympathy to those left bereft.
“It is important not to become numb to the scale of the human tragedy these numbers represent.
“The last two days have seen releases of R value from government (1.2 to 1.3), Nowcasting from Cambridge (60, 000 infections every day, 95 % CI 46, 000 to 80, 000) and ZOE/KCL (49, 000 new cases).
“What can we conclude?
“There is still a lot of people in the UK becoming infected. Some of these people will get sick, of those who get sick some will end up in hospital and of those who end up in hospital some will die. Around 5 to 8 people for every 1000 people who are infected die.
“The spread of the virus in London has slowed, the number of new cases each day has reduced. This will reduce the pressure on the hospitals in time. Even if we have passed the peak number of new admissions, stability only comes when more people who have recovered from covid19 leave hospital than newly infected people enter hospital. We are not there yet.
“The national lockdown is working in that rate of increase in the virus is slowing across the UK but we have to redouble our efforts on social distancing and hygiene. The virus is far from beaten.
“The November lockdown and the cancelling of Christmas in London, were seen at the time and more obviously in retrospect as having saved many thousands of lives. In a democracy, elected politicians have to balance the cost in lives of not locking down against the cost incurred by lockdown.
“Had we listened to those with an established track record of opposing lockdowns, who at every turn downplayed the risks of covid19 and are yet to concede their errors, we would have overwhelmed the NHS in London.
“As a result the thousands of deaths each day that would have occurred would include not only people infected with covid19 and by the cruel arithmetic of the disease been predicted to die, but added to them would be other covid19 sufferers dead only because health care would have to have been prioritised and further include those who would have died from other acute diseases simply because there was not enough doctors, nurses or beds.
“Arguing against lockdown whilst accepting the real costs in lives of not locking down is a proper debate for a democracy; pretending the costs in lives does not exist is deluded or dishonest.
“The good news is the development and distributions of vaccines.
“The emergence of new strains is to be expected when the virus is spreading rapidly. We have surely learned of the potential for the virus to travel the globe.
“We should want to reduce the opportunity for new variants to emerge because there is a danger a variant may at least partially escape vaccine. I am confident new vaccines can be made to any variant, however it’s better not to start a fire in the first place than rely on a good fire extinguisher. .
“This means as close as possible to universal vaccination, letting the virus spread unchecked in a population (be it in one country or in an age group) is to play with fire.
“For the UK, there is an important debate to be had now, as to what happens when the most vulnerable are vaccinated.
“A sudden return to normal when the virus is widespread in young people could lead to emergence of new strains.
“In terms of reducing the threat to the UK, we also have to consider the situation in poor countries and those without the infrastructure to vaccinate efficiently. By helping them, we help ourselves.
“It is at least plausible that vaccination may need to repeated or even tweaked over time, there is a strong case for continuing to invest in the study of coronaviruses, to develop more vaccines and new medicines.”
Comment also on other estimates of COVID-19 R value and growth rates:
Prof Keith Neal, Emeritus Professor of the Epidemiology of Infectious Diseases, University of Nottingham, said:
“COVID 19 is spreading in a variety of different settings and there are different implications for control between community cases and care home cases where infection control is important and in community transmission where we use social distancing.
“What is driving the pandemic are community cases. Care home cases are an issue for health care services as these are many of the people requiring hospital care.
“Once infection is in a care home there is limited opportunity for spread further outside the care home as the residents don’t leave or have visitors and staff are tested weekly to see if they have infection, if positive they then isolate. Also the control measure in care homes are infection control measures and not social distancing. We have now added vaccination to this but it is a little too early to see any effect of vaccination but hopefully in a few weeks this will be apparent.
“The Zoe App data (https://covid.joinzoe.com/data) is a measure of spread in the community as these people are the participants in the study. It will also include health and social care workers but these people are part of the community.
“One of the problems with models reporting results is that the crucial assumptions inherent in the model are not always apparent or published. Without knowing which assumptions have been made it is difficult to interpret any model.
“I do not know what the current calculations from the government include, but in the past care home and hospital acquired cases (important for the impact of COVID-19) have been included in these growth rates, artificially increasing the community R value by not separating out where infections occurred. Large numbers of cases in care homes may not be a true reflection of what is actually happening in the wider community.
“Large numbers of new cases in care homes are very important because of the health impact but if added to community cases inflate the picture of the community spread which is where the spread of COVID-19 is maintained.
“The Zoe data and the government figures are measuring different things.
“The Zoe data still shows an R of 0.9 so social distancing measures cannot yet be relaxed.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“Probably the most important change in the range for the UK R number this week, compared to last, isn’t that it’s gone up or down, but that it has narrowed a lot. This week’s interval is 1.2 to 1.3, which is as narrow as it ever gets. That shows that SAGE and its advisers are much more confident than they were last week about the national R value. Last week the range was 1.0 to 1.4, which is the widest it’s been at national (UK) level. In terms of R, the uncertainty has become much less than it was a week ago. I don’t know the reason for this, but I’ll speculate it could be because some of the usual data sources underlying the estimate weren’t available over Christmas, or perhaps that the estimation teams have learned more about modelling the new variant that continues to spread across the country. Unlike last week, it’s now much clearer than the national R value is above 1, indicating that the number of cases is increasing. You might wonder how that can be, given that the data on confirmed cases is now showing a decrease in three over the four UK countries (and probably indeed all four, though the position isn’t so clear in Scotland), and in all of the English regions (though much more clearly in some than in others). One reason is that the R estimation uses a lot of data sources, and some of them lag behind the current infection numbers. That applies particularly to data on hospital admissions, people in ICU, and deaths, because those events don’t happen immediately when someone is newly infected. Therefore this week’s R number and growth rate estimates can’t completely take into account the position in very recent days. Provided case numbers do keep moving in the direction that they have been moving recently, the R number and growth rate should fall over coming weeks – but that might not happen unless people continue to comply with the restrictions.
“The growth rate range for the UK is +2% to +5% per day. If it’s +5%, for instance, that would mean that the number of new infections tomorrow is 5% higher than the number today, and the day after is 5% higher again, and so on. If that rate continued, the number of cases would double in about two weeks – but if confirmed cases do continue to fall, there won’t be a 5% per day increase and that doubling won’t happen. A +2% growth rate implies a doubling time of around 5 weeks, if it continues.
“Again, though, SAGE point out clearly that the trends in the pandemic are quite a lot different in different places. This means that concentrating on the national UK R number and growth rate only gives a kind of average picture, that hides the differences between the UK countries and regions. For England as a whole, this week, the R range is 1.1 to 1.3 – a wider interval than for the whole UK, and with a slightly lower value for the lower end (1.1 instead of 1.2). A difference that small isn’t very important, given the uncertainties, but a possible reason why the English situation could be slightly better than the national average is that the R number for Scotland, the second largest UK country, seems to be rather worse than in England. The English growth rate range is +1% to +4% per day, corresponding to a doubling time of between ten weeks and two and a half weeks, if it continued over time.
“But even an average over the whole of England disguises a lot of differences between regions of the country. The direction of the epidemic is most favourable, according to these SAGE figures, in the regions which were first and (initially) worst affected by the new variant of the virus . For London, the R range is 0.9 to 1.2. Part of that is below 1, which would indicate a shrinking number of new cases, though most of the range is above 1 and that would indicate that it’s most likely that R is still above 1. But an interval going below 1 is good news, particularly given the dire pressure on many London hospitals. In the East of England and the South East, the lower end of the R range is now exactly 1.0, which gives room for hope in these hard-hit areas. In London the growth rate range is -2% to +3%, and if the growth rate is negative, that means that cases are falling (though the range goes above 0 too, which would mean that cases are increasing – the forecasters can’t be sure of the direction). The growth rate range for the South East also goes from below zero to above zero, which is relatively good news given how things have been there, and for the East of England, the lower end of the growth rate is 0 (meaning no change over time in cases), though the upper end is still well above zero (+4%). The position on the R number and the growth rate is now worse in the rest of England, that is, in the regions where infection rates weren’t increasing so much in the past month or so. That includes the regions of the North and Midlands, which have gone through month after month of high case rates but where all the restrictions had been having a clear effect towards the end of 2020. It also includes the South West, where infection rates have generally been relatively low. In the South West, the R range is 1.2 to 1.5, and the growth rate range is +4% to +7%. A +7% daily growth rate, if it continued, would produce a doubling of cases in about a week and a half. A possible reason for higher R and growth rates in the English regions outside the south-east corner of the country may be that the more infective new variant of the virus has become more common there very recently. Although the new variant was already present in those regions, and was accounting for reasonably large numbers of cases in most of them, it had not become as dominant as in London, the South East and the East. Maybe that’s no longer the case in those other regions. I can’t be sure, though – an excellent check would have been the data on virus variants that I’d expect to see in the weekly results of the ONS Infection Survey, but their publication is delayed this week.”
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.”
None others received.