A preprint, an unpublished non-peer reviewed study, reports on the latest data from the REACT-1 study on COVID-19 spread across England.
This Roundup accompanied an SMC Briefing.
Dr Simon Clarke, Associate Professor in Cellular Microbiology at the University of Reading, said:
“The latest REACT-1 survey brings news of broadly sustained numbers of infections across England between 6th to 22nd January: the national prevalence was 1.57%, the estimated R number of 0.98, indicating that the situation remained in stasis. There was still a fairly wide variation in prevalence between areas with infections still the highest in London.
“Intriguingly, during this time, hospital admissions began to fall around the 14th January, which would normally indicate an overall decrease in infections. This apparent contradiction might be because infections were highest in the age groups from 13-24 years, who are least likely to be hospitalised. With that in mind, merely looking at hospital admission numbers will not be a reliable way of gauging the risk to society of lifting restrictions and opening schools. While there is still so much virus circulating in the population, there will be a significant risk of filling intensive care units with unvaccinated members of the working age population and those for whom the vaccine fails to give any protection against Covid-19.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“Last week’s interim preprint from the REACT-1 survey from Imperial College covered only the first part of their round 8. This new preprint adds more data, based on swabs that they tested between 17 and 22 January, so it covers the whole of Round 8, from 6 to 22 January. However, of the almost 168,000 people that provided data for Round 8, 85% (about 143,000) were included in the interim preprint last week. Therefore most of the conclusions are little changed. Some of the estimates have moved a bit, the statistical uncertainty has decreased somewhat, but on the whole the picture is mostly the same as in last week’s preprint.
“As always, the REACT-1 survey estimates the pattern of current infection with SARS-CoV-2 by carrying out swab PCR tests on a reasonably representative sample of the community population of England. It is thus unaffected by changes in the reasons why people might be tested in the routine systems used to produce the daily numbers of confirmed cases on the dashboard at coronavirus.data.gov.uk, and one of its important uses is to see whether the trends in those confirmed case numbers do really represent what’s going on in the pandemic. But it also provides other useful data, because the people who take part in the REACT-1 survey provide more information about themselves than is generally available from routine testing data. Again as always, the REACT-1 estimates are for the numbers and proportions of the English community who would test positive, if they could all be tested at about the same time. Because any test of this general kind can produce false positive and false negative results, and because there are likely to be a lot more false negatives than false positives at this stage of the pandemic, the true numbers who are infected are likely to be rather higher than the REACT-1 estimates, but it’s difficult to say exactly how much higher.
“As last week, the clearest conclusion from REACT-1’s full round 8 is that the percentage of people who would test positive in the period covered by Round 8, 6-22 January, is much higher than in the previous REACT-1 round in November and the first few days of December. The Round 8 estimate, that 1.57% would test positive, is 73% higher than the estimate for 25 November-3 December. (The estimated percentage testing positive in last week’s interim preprint was 1.58% – very little different.) But that increase is already known from many other sources, including last week’s interim report. A more important question, right now, is what has happened since the confirmed case numbers peaked at the turn of the year. Last week’s interim report found no strong evidence that infections were either rising or falling during the period covered (6-15 January) – but the researchers said there was a slight suggestion that infections were rising towards the end of that period. I remain rather dubious about the validity of estimates of change over these very short periods, because they are based on relatively small numbers of swab tests from individual days during the period, and the statistical signal is not very strong. This time, again the researchers again report that there is no strong evidence of a rise or a fall over the period from 6 to 22 January, but they now say that there is a ‘suggestion’ of a fall towards the end of the period. I think that’s a reasonable thing to say, but in my view the ‘suggestion’ is inevitably pretty weak. A fall of any kind in the number of infected people is good news, and a fall would match what we’ve seen recently in numbers of confirmed cases, but I’d want to see more data to be surer of this trend. We’ll get some more data, covering the end part of this REACT-1 round, in the next ONS weekly infection survey results this Friday. I very much hope (and expect) that they also show evidence of falling numbers.
“This REACT-1 preprint points out again that the pattern of change is different between different regions. In most regions of England, the central estimate is that the number of cases was decreasing over the round (6-22 January), though there is very considerable statistical uncertainty about that. In every region except the South West, the REACT-1 data are consistent with either a fairly slowly increasing trend or a fairly slowly decreasing trend, or anything in between, though in the South West the statistical analysis is more conclusive that infections are decreasing. The picture in the South West is not entirely consistent with the latest report from the ONS Infection Survey, which did not show a decline in the South West, and across other regions, there are other differences in the trends between the two surveys. But the latest ONS results go only as far as 16 January, not to 22 January which is the end of the REACT-1 round, and this week’s ONS results might show somewhat different trends. There is so much unavoidable uncertainty in some of these short-term trend estimates for parts of the English population that one should not expect every conclusion to match.
“Both last week’s preprint and this new one present interesting and important information on factors that are related to the chance that a person will test positive. The conclusions this week are generally broadly the same as reported last week, but the extra data has made some of them even clearer. Having Asian or Black ethnicity makes it more likely that someone will be infected, compared to the position with people of White ethnicity. The more deprived the area where someone lives, the more likely they are to test positive. And living in larger households is associated with higher risk of infection. None of those patterns is new, of course, but it’s depressing that they have been so consistent throughout the pandemic.”
Preprint (not a paper): ‘REACT-1 round 8 final report: high average prevalence with regional heterogeneity of trends in SARS-CoV-2 infection in the community in England during January 2021’ by Steven Riley et al. was posted online at 00:01 UK time on Thursday 28 January 2021. This work is not peer-reviewed.
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.”
None others received.