John Roberts Profile picture
Actuary for far too many years. Welsh, but now live in the spiritual home of the Wurzels out west. Don’t think I’ve got the accent yet. Like zider though...
Peter Newport Profile picture Adam Kleczkowski 💙 Profile picture 2 added to My Authors
13 May
I nearly missed today's 2 weekly antibody update from @ONS, so here's a brief summary. It's fairly disappointing though, with no material increases, and the modelled curves dipping.

E increases from 68.3% to 69.3% (65.6%, 73.9%)
W is up from 61.0% to 63.2% (58.8%, 68.3%)

1/ Image
S is up from 57.8% to 59.2% (54.8%, 64.4%)
NI is up from 62.5% to 63.5% (57.6%, 71.2%)

Of note the modelled curve for NI is increasing, although the increase from the last update is no greater than the others.

2/ Image
It's interesting that even the unvaccinated age groups are showing nearly 50% prevalence. Further up, the highest age groups are now rising again, as the 2nd dose takes effect, but lower down we're now seeing the dip that we saw earlier in the highest groups.

3/ Image
Read 4 tweets
13 May
Thu vaccine update (incl England weekly data)

637k total doses today, 17% up on last week, bringing the 7D total back to 3.6m, close to the late April high.

1st doses of 184k are up 32%, and are the highest daily total since March, bringing the 7D total close to 1m.

1/5
A good day all round with 2nd doses of 452k up 12%, bringing the 7D total back up to 2.6m, and the overall total to just shy of 19m.

2/5
On take-up, England 1st doses have made more progress in the 45-49 group and look as though they should just hit 80% in due course.

That would be 10% less than the band above, which is a little worrying in terms of a downward trend as we work down the ages.

3/5
Read 5 tweets
13 May
A summary thread of the latest #REACT survey from @imperialcollege on England infectivity.

The headline is very good news, prevalence has halved since the last round, falling from 0.20% to 0.10%, with R at 0.9 between the rounds.

Round 11 is between April 15 and May 3.

1/
115 positive samples were obtained from 127k tests. Note the slightly lower number of tests this round, which the report notes is due to a lower response rate, suggesting a lower interest in participants as the situation improves.

2/
Between rounds, R is put at 0.9, though slightly higher at 0.94 within the round. This is consistent with recent SAGE views, and not unexpected as the lockdown restrictions are gradually eased.

3/
Read 11 tweets
13 May
The latest @ICNARC report on intensive care activity is out. With the situation now much improved now it enables a good comparison of the two waves (W1/2). A thread here on behalf of @COVID19actuary.

W2 patient admissions of 25,543 were well over twice that of W1 (10,951).

1/
Despite starting W2 much more slowly, London rapidly accelerated and ended up being the worst affected again in terms of population size, contrary to those who early on were suggesting that herd immunity from the first wave was protecting it.

2/
W2 developed more slowly, and extended over a longer period so the peak bed occupancy for W2 was not much greater than W1, but high volumes (and thus pressure on staff and resources) were sustained for a much greater period.

Recent levels were higher than last summer too.
3/
Read 19 tweets
12 May
Wed vaccine update (& Wales summary)

485k total doses, 26% up on last week, so the recovery from the BH continues, with the 7D total now up to 3.5m.

1st doses of 351k are just 6% up (unusually, less than the 2nd dose increase), but takes the 7D total ever closer to 1m.

1/5
2nd doses of 350k are up 36%, so another better day, albeit not at the level of 2 weeks ago.

The 7 day total is now back over 2.5m, which if maintained should be enough to keep up with the required pace for the next few days.

2/5
Wales 1st doses are settling down at 85% for the fifties, about 5% lower than the sixties. To date the forties have just hit 70% - hopefully that will continue to increase though.

Second dose progress seems to be concentrated in the 65-59 group currently.

3/5
Read 6 tweets
4 May
How soon to first dose all adults? We can look back around 11 weeks to D1 volumes to predict how many D2's are needed, and deduct that from an assumed supply - we'll say 3.7m for now.

The chart shows future D1 volumes - apart from 1 week it should gradually increase.

1/6
If we assume c85% of under 50s take-up the jab then we need to reach around 47m to have given everyone who wants it a first dose.

That will take until around 28th June. Any doses after that (above the grey line) can be used to rapidly accelerate second doses if appropriate.

2/6
If supply is less than 3.7m pw what difference does it make? Even dropping it to 3.3m only pushes it back two weeks, to July 12th, still well ahead of the govt's target of the end of July.
3/6
Read 6 tweets
3 May
Monday vaccine update: 50m passed!

We've now passed 50m total jabs, although today's total (excl Wales) of 255k is another disappointment, down a quarter on last week.

The 7D total continues to fall, to 3.4m.

As yesterday, the shortfall is all in the second doses.

1/5
171k for 2nd doses is down a third, and we are now barely ahead of my 11 week line. The 7D total of 2.6m is the lowest for a fortnight.

What's the prognosis for the next couple of weeks though?

2/5
The yellow line of 1st doses in Feb is now starting to fall, and will drop gradually from c3.1m pw to 1.9m over the next 4 weeks. That should mean it's a bit easier to keep up with the 2nd dose requirement, and hopefully give a bit more leeway for 1st doses too.

3/5
Read 5 tweets
1 May
A paper for SAGE released yesterday analyses hospital admissions post vaccination. Reinforcing the point that one dose does not give full protection, 400 people entered hospital and subsequently died who displayed symptoms after vaccination.
1/6 assets.publishing.service.gov.uk/government/upl…
The majority of these (72%) showed symptoms within 21 days of the jab, so probably were infected prior to the jab, or before immunity develops. However, 113 people only showed symptoms after 21 days or more (likely "vaccine failure").

2/6
You can see also that the "over 21 day" cases weren't confined to the few days after, but are spread over a longer period. (Note these graphs relate to admissions, but the point is the same.)

The decline in the green zone will be at least partly due to declining prevalence.

3/6
Read 6 tweets
30 Apr
The latest ONS infection survey shows dramatic falls across the UK, with England falling by over 40% and Wales by 50% in just one week. (Data is to 24th April.)

E falls from 0.17% to 0.10% or 1 in 1,010.
W falls from 0.12% to 0.06% or 1 in 1,570

1/6
There are slightly less dramatic falls in Scotland and Northern Ireland, though still very encouraging.

NI falls from 0.15% to 0.11% or 1 in 940
S falls from 0.18% to 0.16% or 1 in 640.

2/6
Regionally, only Yorkshire and Humber now rounds to 0.2% (actually 0.19%).

In the opposite corner the SW is now shown as 0.0%, though is actually 0.03% or 1 in 2,980.

3/6
Read 6 tweets
30 Apr
The latest @ONS survey on vaccine sentiment shows continuing strong support to have (or have had) the vaccine. There's still the downward shape by age, but even below 30 it's put at 88%.

We have started to see a downward trend in reality too, as we work through the ages.

1/5
As noted yesterday, for England the 50-54 group is not quite yet at 90%, which suggests the actual picture is slightly lower than the survey (which puts 50-69 at 96%). Whether this is due to hesitancy or practical issues with attending is not obvious.

2/5
Other data in the survey shows higher proportions of younger people not maintaining social distancing with friends, whereas for older ages it's mixing with family members that is more the reason for not following the guidance.

3/5
Read 5 tweets
29 Apr
Thu vaccine update (take 2 - apologies)

597k total doses today, up 12%, so the 7D total is now up to 3.7m.

First doses of 134k are up 14%, so we are now giving first jabs to nearly 850k each week.

1/5 ImageImage
2nd doses of 463k are up a healthy 11%, and bring the total to over 14m. That's broadly equivalent to the most vulnerable groups 1 to 4, though as we've seen throughout the distribution isn't quite that simple.

2.85m over 7 days is another new record.

2/5 Image
England take-up figures show first doses making slow progress, (though this predates the opening of 42-44).

It's disappointing that the 50-54 group has not reached 90%, so we can see a clear trend downwards from 65 to 50. It's too early to form a view on 45-49 yet though.

3/5 Image
Read 5 tweets
28 Apr
The PHE study reported widely today shows that those who are vaccinated who still get infected are 40% to 50% less likely to pass it on to a household contact than those unvaccinated.

1/7
khub.net/documents/1359…
This is an important finding, as household setting transmission is more difficult to suppress through the interventions we are all familiar with, particularly in large multi-generational households.

2/7
This benefit is on top of the lower risk in becoming infected, which is typically around 65% after a first dose (there isn't enough data on any further protection from a 2nd dose).

So the overall first dose effect in terms of protecting your family becomes around 85%.
3/7
Read 7 tweets
28 Apr
The latest ONS antibody survey is out. The last one (2 weeks ago) showed a surprising pause in the increase in the proportion with antibodies, but this has now resumed.

By country the movement is
E: Up from 55% to 68%
W: 49% to 61%
NI: 55% to 63%
S: 46% to 58%

1/5
Looking at England, by age we can see the 2nd dose effect kicking in at older ages, which had started to fall away. eg Over 80s is back up from 78% to 87%.

We also have modelled estimates for those vaccinated - these appear slightly higher than other data, eg NHS.
2/5
By region London is put at around 10% behind the rest in terms of having had the first dose, but is "in the pack" in terms of antibody prevalence - maybe due to higher levels of antibodies acquired through infection.

3/5
Read 5 tweets
14 Apr
The latest antibody survey from @ONS has been published, and show that levels have flattened off in England and are falling in the oldest age groups. Similar patterns are seen in W and S, though not in NI.

The overall % in England is 54.9%, compared with 54.7% two weeks ago.
1/5
In England the first sign of a fall in older ages was seen two weeks ago, but is now clearer and extends down to all groups over 65. Younger age groups are continuing to rise.

The fall over the 2 weeks for 80+ is from 86% to 78%, and is 8-9% in the three upper groups.

2/5
Visually you can see it here at the latest point in time, or follow the link to have the interactive version. ons.gov.uk/peoplepopulati…
3/5
Read 5 tweets
13 Apr
A 40m Tue vaccine update:

261k total doses today, around 150% higher than the last day of the Easter BH data.

With all the BH now out of the 7D total, it's just tipped back over 3m, consistent with govt messaging for April supply.

2nd doses and vaxing the 40s next.

1/4
Just on 200k 2nd doses today, so another new high for the 7D total of 2.4m. Over the last week the split has been around 80/20 in favour of second doses.

We remain 3 days ahead of my 11 week target line, so are still keeping to the required pace.

2/4
With bookings opened for the 45-49 cohort, how long will these take to jab? There are around 4.4m in total (UK), but we know around 27% of the 16-49 group have already been jabbed (Eng). That's likely to take the number down to no more than 3m, maybe even less.

3/4
Read 5 tweets
8 Apr
The latest @imperialcollege #REACT study shows a sharp fall since the last report, with infection levels falling by 60% from 0.49% to 0.20% (0.17%, 0.23%).

It suggests that within the round the decline has levelled, and puts R at 1.0 for the period.
spiral.imperial.ac.uk/bitstream/1004…

1/9
Round 10 (11/3 to 30/3) saw 227 positive tests out of 141k swabs, and the central estimate is now at its lowest since early September. It's slightly below the latest ONS figure for England too, of 0.27% (for the week to 27th March).

2/9
The movement between Rd 9 and 10 suggests an R of 0.84 , but samples collected within the round indicate that the fall has levelled off.

Notice though that there is a wide CI (0.81% to 1.21%), reflecting uncertainty as to whether there is still some fall or even a rise.
3/9
Read 9 tweets
7 Apr
Wed vaccine update:

For the first working day post Easter, a disappointing 266k total doses, only just over half last week's figure, so the 7 day total drops further to just 2.35m, with 2nd doses making up 2/3rds of that figure.

2nd doses and Wales next.

1/4
With only 180k second doses, down a third on last week, the 7D total continues to fall back, now 1.55m.

The two red lines move ever closer, with only 2 days leeway now over the 11 week benchmark with first dose progress.

A national comparison is of interest though...

2/4
HT to @RhonddaBryant for prompting this, but if we compare current 2nd dose totals to the equivalent nearest 1st dose day we see that:-
E is slowest, only on 20th Jan (76 days)
S/NI are both on 26th Jan (70 days)
W is well ahead on 2nd Feb (63 days).

More on Wales next.

3/4
Read 4 tweets
6 Apr
Tue vaccine update: (Spoiler alert - still v slow)

Only 105k total doses today, down 3/4 from last week, and similar falls to those reported yesterday. Hopefully this is the last day of the Easter effect, although we will probably continue to see low 1st dose numbers.

1/4
It's a similar story for 2nd doses, at just 65k, although the fall is not so marked. But there's more urgency here to recover lost ground, to keep a comfortable margin ahead of the 12 week guide (and my prudent 11W benchmark too.)

Next, what's the prospect from now on?

2/4
The SPI-M papers published yesterday show clearly that the govt has reduced its expectation of the roll-out recently, asking the modellers to slow it from 3.2m to 2.7m for England until the end of July. Beyond that there is an even steeper fall, halved from 3.9m to 2m.

3/4
Read 4 tweets
5 Apr
The latest SPI-M modelling by Imperial and Warwick has been published. Both suggest that the current roadmap would result in a third wave of some degree over the summer. The Warwick central scenario is more pessimistic with around 250 deaths pd in Aug.
assets.publishing.service.gov.uk/government/upl…
1/4 Image
One interesting chart by Warwick models who might become ill by vaccine status and age. Of note, during the peak, a high proportion is for those who will have had both doses. The assumption of 90% efficacy against death after 2D drives this figure, but that feels low to me.

2/4 Image
Note that vaccine roll-out in England is assumed at 2.7m pw until end of July and 2m thereafter, which suggests the 4m (UK) we saw in March is unlikely to be repeated. This is a substantial reduction from the previous assumption, suggesting a slower roll-out is now expected.
3/4 Image
Read 4 tweets
16 Mar
The latest ONS data show further proportionate falls in the oldest age groups since mid-Jan, benchmarking against the under 70s.

Age 80+ deaths are now 45% lower, age 70-79 are 37% lower.

What might that mean in terms of lives saved?

1/4
Overall COVID deaths in the latest week were around a third lower than they would have been without the vaccine. The total saved by this estimate is now over 4,000 lives.

It's only an estimate of course, we shall probably never know the exact number.

2/4
The rapid fall in infections remains the greatest influence in the overall reduction in deaths we've seen. But we can now clearly see the relative additional impact on those vaccinated first in the faster falls of those age groups.

3/4
Read 4 tweets
14 Mar
With volumes now expected to increase considerably, what possible end dates are there for the adult vaccination programme?

Assuming 4m a week 1st doses could be around three days after the lifting of all restrictions (21 June) - nearly a month later if we only average 3m.
1/4 Image
Once 1st doses are complete, there should be a rapid completion of 2nd doses, with no need to stick to the longer interval. So end of August at the latest, but maybe by mid July if we can sustain the higher level.

2/4
I've assumed a slightly lower take-up for Groups 10-12. You could argue that my assumptions are a little high, but then the aim should be to get as many as possible to be vaccinated. And the latest ONS estimate is that only 6% are hesitant across all age groups.

3/4
Read 4 tweets