for those of you too distracted by Eurovision to read the PHE's technical briefings, here's a quick summary thread. first, the bad news: 1) as per press reports earlier today, B.1.617.2 appears to have a moderate degree of vaccine 'escape' particularly after 1 dose 1/n
2) looking at secondary attack rates, B.1.617.2 appears to have a transmission gain of around 50-60%; some of this will be due to the vaccine escape, but as I noted earlier, that's not enough to explain 50-60%, so there's probably an uplift in R0 also 2/n
but there's also quite a bit of good news, some obvious and some less so: 1) the vaccine escape is much smaller after 2 doses than after 1 (so once we're fully vaxxed, much less of an issue) 2) there's no sign yet of large numbers of re-infections, suggesting that immunity.. 3/n
... from prior infection is mostly unaffected by the variant 3) also no sign of any change in hospitalisation rates, although it's early days on this data 4) since B.1.617.2 is already more than 50% of cases in the UK, and yet the national case rates are still flattish... 4/n
... this suggests that B.1.1.7 must be shrinking fast (squashed by the impact of controls + vaccine). So while R may increase a bit in the next 2 weeks as B.1.617.2 takes over completely, and due to the Step 3 relaxations on Monday, this will be offset by more vaccine and... 5/n
...better weather. So it feels unlikely R will get above 1.3, and should then come back down (helped by the half term week). The trickier question is what happens when we open on June 21st. My gut feel is this is still in the "just about manageable" category but it's... 6/n
... not a slam dunk, and I will need to run some new model scenarios over the next day or two before giving a clearer view on that. More soon (but not tonight!). /end
• • •
Missing some Tweet in this thread? You can try to
force a refresh
When I tweeted my initial reaction to the PHE data release last night, I promised you some model scenarios to help understand the impact of B.1.617.2 on the roadmap, and in particular whether opening up on 21st June still looked possible. 1/
I’ve done some analysis, and I’ll warn you in advance it’s a bit of a mixed bag: at the more optimistic end of assumptions, things look not too bad. But at the more pessimistic end, we’re back to facing a mid-sized exit wave, which calls the timing of Step 4 into question. 2/
Before that, a quick word on assumptions: while the PHE data last night was very helpful, it still leaves quite a lot of questions unanswered, and some of the data is confusing, or appears to contradict things we thought we knew. See for example: 3/
having played around with some numbers, I think it is plausible that a moderate immunity escape by B.1.617.2 (as per JBM thread) could account for a ~15-30% apparent gain in transmissibility for the new variant, given the UK's current distribution of immunity.
some workings for those following at home: I'm assuming (rough numbers, allowing for lags) we have about 35% of total population with an effective 2nd dose, and another 25% with an effective 1st dose. and about 30% of the unvaxxed will have a prior infection.
I'm then applying transmission reduction factors of 80% for a first dose, 90% for 2 doses, and 85% for a prior infection. note these factors include the VE vs. infection, with an assumed 40-60% reduction in transmissibility of breakthrough infections (more after 2nd dose).
My summary thoughts on B.1.617.2: 1. The data on increased transmissibility is concerning, but it’s good that a significant vaccine escape looks unlikely 2. My model suggests that we can probably still exit with only a small wave if we’re prepared to extend baseline controls,
… and vaccinate teenagers in July. 3. We need to keep watching the data closely, and be prepared to take further action, but at present I don’t see a compelling case to alter the roadmap.
A long thread follows, with model graphs to illustrate the various scenarios/actions.
Sadly I’m as confused as everyone about the data on transmissibility – but after playing around with some regression analyses, I have no reason to doubt the SAGE conclusion that a ~50% increase in transmission is a “realistic possibility”. 3/ assets.publishing.service.gov.uk/government/upl…
HI thread Mark II: after I tweeted last week about Herd Immunity (HI), quite a lot of people got in touch to point out that my model was either over-estimating or under-estimating the HI threshold, because it treated people as all being the same. 1/n
And the thing is, they’re right, in both directions. Differences between groups of people, and between individuals within those groups (so-called “heterogeneity”), make it likely that I’m over-estimating the effect of vaccines, and under-estimating the effect of infections. 2/n
Let’s see if we can correct for both of these effects, or at least estimate how wrong I might be. We’ll start with the age group effect on vaccines, because that’s easier to sort out. The problem here, as @AdamJKucharski noted in his reply, 3/n
A thread on herd immunity. In summary:
- We’re not there yet
- Full herd immunity will probably arrive in late June, once we’ve vaxed all 18+
- From early June we’ll likely be in a partial-HI state where R will stay <1 with just baseline controls and cautious behaviour. 1/n
There’s been a lot of wild claims about herd immunity in recent weeks, varying from “we’re already there” (fil.ion.ucl.ac.uk/spm/covid-19/f…) to “it’s impossible” (nature.com/articles/d4158…). Both are demonstrably wrong, and new data this week (from ONS and PHE) makes this clear. 2/n
First, the ONS-based analysis (ox.ac.uk/news/2021-04-2… ) showed us that a single dose of either AZ or Pfizer reduces covid infections by 65%. (and after 2 doses by 70%, although that’s probably an underestimate). 3/n
Reflecting on this week’s “lockdown vs. vaccines” debate, while the binary logic of “it's one or the other – which is it?” is clearly wrong, my initial reaction of “it’s a bit of both, the balance changes over time” was also slightly too reductive. 1/
Surely the bigger strategic point is this: to get to the relatively good place we’re in now, we needed both lockdown AND vaccines. Essentially, lockdown kept the virus pinned down until the vaccines were ready to come and knock it on the head. 2/
This wouldn’t be a proper thread without some numbers from the model to illustrate it, so here goes. First, let’s start with a world where there is no lockdown, and no vaccine. The tricky bit here is what to do instead of lockdown: I think the most realistic alternative… 3/