,The latest #REACT study (Round 9b) on infectivity from @imperialcollege suggests continuing falls but at a slower rate in the second half of Feb, with R around 0.86 by the end of the period. (4th to 23rd Feb), having been 0.73 earlier a few weeks earlier.
A thread....
1/12
A thread....
1/12
Starting with the summary details.
Prevalence is put at 0.49% (0.44%, 0.55%) for Rd 9, although slightly lower in the second half at 0.47%. These rates are less than a third of those seen in Rd 8 just a month earlier. That's great news!
2/12
Prevalence is put at 0.49% (0.44%, 0.55%) for Rd 9, although slightly lower in the second half at 0.47%. These rates are less than a third of those seen in Rd 8 just a month earlier. That's great news!
2/12
The results suggest that R was around 0.73 during the fastest period of the fall (halving time 15 days), but has now eased off to 0.86 (31days), with notable regional variations causing the increase in R.
3/12
3/12
You can see here that the position in London and the SE appears much more uncertain, with R possibly above 1. These regions fell fastest in Jan from very high levels, as can be seen from the lower half of the table.
4/12
4/12
Visually it's certainly clear in London, maybe less so in the South East. Other areas still appear to be falling however. Within London the growth appears to be concentrated in the West and North West.
5/12
5/12
By age group the picture looks much flatter at employment ages 25-54, with a variable picture below, and falls at the upper, more vulnerable ages. It's maybe surprising the lowest group has continued to fall, given reported numbers attending primary schools.
6/12
6/12
There's more granular data on ethnicity in this report, although given numbers testing positive for some groups, there's a limit to how much can be inferred from the detailed data. Though at the top, Asian and Black levels of infectivity remain double those for whites.
7/12
7/12
Odds ratio data shows that the deprivation variation looks to be widening. It also shows that over 65s appears to be improving - a vaccine effect maybe?
The ethnicity odds appears to be at odds (pardon the pun) with those shown in the previous table. Why could that be?
8/12
The ethnicity odds appears to be at odds (pardon the pun) with those shown in the previous table. Why could that be?
8/12
I would interpret it as showing that the absolute differences seen (the "doubling") can largely be attributed to other factors, such as employment, deprivation etc. Once these are allowed for, the remaining differences are lower.
9/12
9/12
As a BBC programme reported earlier in the week, that doesn't mean there isn't an issue here, which actually goes well beyond COVID to wider issues of health inequalities. COVID has just highlighted the issue very starkly.
10/12
bbc.co.uk/iplayer/episod…
10/12
bbc.co.uk/iplayer/episod…
The usual caveats apply regarding REACT being just one source of data to help understand the pandemic's course, along with ONS and ZOE. This round completed 9 days ago, which also needs to be borne in mind.
Report here: spiral.imperial.ac.uk/bitstream/1004…
11/12
Report here: spiral.imperial.ac.uk/bitstream/1004…
11/12
This latest extensive survey tested 165,456 people with 689 positive results between 4th and 23rd Feb.
Thanks to all at @ImperialCollege involved in the survey, along with its partners @Ipsosmori.
12/12 END
Thanks to all at @ImperialCollege involved in the survey, along with its partners @Ipsosmori.
12/12 END
• • •
Missing some Tweet in this thread? You can try to
force a refresh
