I’m new to Twitter. Have mostly tried the academic route on this. The letter of corr + systematic review I sent to LC&AH on this were rejected, and now my univ’s Research Gov Office is working with UKRIO to organise an audit.
@ingridjohanna66@threadreaderapp I originally worked alone on this, since didn’t want to disrupt med researchers at a time like this.
I know journals are doing the best they can with an avalanche of submitted articles that could influence policy that saves/jeopardises lives.
Difficult to know how hard to push.
@ingridjohanna66@threadreaderapp (To clarify, what I sent was rejected by the LC&AH editor without ever being sent to peer review.)
New @UKHSA Omicron update, including new growth estimates.
From data up to 5-6 Dec, they estimated a .35 continuous daily growth advantage of Omicron over Delta in England (so times exp(.35) = 1.42 per day). That would mean doubling every 2.0 days. 1/ assets.publishing.service.gov.uk/government/upl…
To estimate this growth advantage,…
Their step 1 was estimating what proportion of S-gene dropout cases in England was due to Omicron over time, using genomic sequencing to check some of their SGTF samples.
It wasn’t until ~28 Nov that most SGTF in England was Omicron. 2/
And indeed, if you start with @UKHSA SGTF proportion data from England (numerically extracted from a graph shared by @AlastairGrant4), and you then fit a line to log odds of SGTF for the period from 28 Nov to 5 Dec, you get a slope of .35/day for continuous growth advantage. 3/
Some preliminary transmission-advantage estimates (by others) for Omicron in South Africa and England 🧵
I’ll try to give an accessible account of some main methodologies for estimating total transmission advantage, some key limitations, and preliminary results. 1/
Total transmission advantage, the fraction
R_t(new variant)/R_t(old variant),
is roughly the factor by which
Omicron_new_daily_cases/Delta_new_daily_cases
changes during one generation interval (GI). (Often 5 days is used for Covid GI. Can also depend some on variant.) 2/
Eg, England had
R_t(alpha)/R_t(wild) ≈ 1.5,
R_t(delta)/R_t(alpha) ≈ 1.7.
Other countries saw results reasonably similar to these.
For Omicron, however, we might see larger variation by country, depending on its differences in response to different sources of immunity. 3/
If there’s one thing worse than hoarding vaccines, it’s hoarding AND wasting them.
If jabs are rolled out so slowly that the target group nearly all get infected pre-jab, that’s a big decrease in overall vaccine benefit—a serious waste.
Out of the 811 non-“see results” respondents about their own 12-15 school in England,
9% (75) jabbed in Sept,
31% (250) scheduled for Oct pre half-term,
60% (486) scheduled for later, or not scheduled at all.
(A few wrote in to revise votes, but not enough to change %s) 2/
But even for scheduled roll-outs, for the past week alone (+ 2 from end Sept), 10 wrote that their school’s 12-15 covid jabs that week had either been cancelled last-minute or had only been able to jab a limited number, eg only Years 10-11, due to inadequate staff or supplies. 3/
In case anecdote is helpful re this NIMS unvaxxed count,
during my 2020-2021 visit to Princeton, the subletter of our 2-bed Cambridge UK flat received vax invitations for *4 adults* besides herself—2 for folks who’d moved out more than 10 yrs ago, and 2 for my own household. 1/
For the 2 who’d moved out > a decade ago, we asked her to send back the envelope and tick “not at this address.”
But we’ve repeatedly sent back their NHS post ticked “not at this address” for years with no effect, so I don’t know if that actually removed them from any list. 2/
There’s no way to use the invite letter’s suggested website to announce absence from the UK, and you can’t phone 119 from outside the UK. Besides, if you take the invite literally, it says ignore the letter if you already have an appointment to be vaxxed, which we did (in US). 3/
TL;DR:
0-19-year-olds tend to make up a larger share of total cases for B.1.1.7 than for wild type, and likely experience a larger relative increase in infectiousness than other age groups. 1/
This finding isn’t really new. Here’s Figure 4B from an earlier Imperial College modelling group paper medrxiv.org/content/10.110…
It’s just these earlier findings were later partly discounted as likely due to environmental effects alone. My paper argues against the latter. 2/
The idea is down to something called “dominant eigenvectors”:
For a linear operation satisfying certain conditions, if you act on a system repeatedly, the system converges to a particular configuration. In this case, the configuration is a certain age-distribution of cases. 3/