Thread by @JamesWard73

Now the kids are in bed, I can focus on updating my model to deal with the many events of today. These include 1) Boris’s plan 2) recent case trends 3) news on vaccine effectiveness vs. serious disease, and vs. transmission. Results below: mostly good news. (thread)

1) Boris’s plan is not very different from some other scenarios we have already looked at – and just a few weeks slower to unlock than our base case. Regular readers will be able to predict that it pushes the “4th wave” peak into the autumn – but no real change to outcomes.

Note I’ve assumed that the final unlocking on June 21 goes into a “near normal” state with some continuing social-distancing requirements e.g. masks in crowded spaces, limits on large indoor events, ongoing WFH etc. – which lasts until 31 Dec.

(without this final step we would get a larger peak in hospitalisations, which would challenge NHS capacity – so need to avoid)

2) Case trends in England and Scotland over the last week are rather concerning, with R seeming to increase and even heading >1 in some local areas. I’m not sure why this is, it could just be a temporary effect of the cold weather in 2nd week of Feb.

https://twitter.com/BristOliver/status/1363882541204275202?s=20

But let’s assume instead it’s a sign of the public gradually taking more risks, and reducing compliance with lockdown rules. So I’ve added a temporary boost to R over the next couple of months, which tapers out as the government controls relax (catching up with public behaviour)

This roughly counteracts the slower relaxation in the government plan, and pulls the “4th wave” peak back into the summer. (in reality I reckon seasonality would push this into autumn/winter, but it doesn’t really matter – we have to deal with it whenever it happens)

Counterintuitively, the final outcomes (i.e. total deaths and hospitalisations) are marginally improved. I think this is because by taking a bit more pain on the downswing from wave 3, we build up more immunity, so wave 4 when it comes is slower, with less overshoot beyond HIT.

3) So here comes the bit you’re really interested in: what does all the news on vaccine effectiveness today (from PHE, PHS and the SIREN study) all mean? I’ve spent a while staring at the data, and had to upgrade my model to handle age-stratified VE estimates, as follows:

Updated figures are in red: as you can see I’ve tweaked my assumption for first-dose VE vs. hospitalisation down to 80% in the older age groups (JCVI categories 1-4), in response to the PHE and PHS data. But I’ve left the second-dose figure at 93%, consistent with PHE.

In the younger age groups, I’ve left the first-dose protection vs. hospitalisation at 85% (consistent with PHS) but increased the second-dose assumption to 96% (from 93%), more in line with SIREN. And the net effect of these changes is: zilch. nada. rien.

The transmission assumptions are harder to pick: we now have decent data (from SIREN) suggesting Pfizer stops 70% of infections after 1st dose, and 85% after 2 doses. But AZ data is still sketchy, with suggestions of 67% after 1st dose, but some rather odd data for 2 doses.

So I welcome suggestions, but in the mean time I’ve gone for a compromise of 67% after dose 1, and 75% after dose 2, for all age groups other than the over-70s (which I’ve left at 50% and 65% as before). Still, this is enough to flatten the fourth wave:

In fact, on these assumptions we could get rid of the ongoing social-distancing controls in the second half of 2021, and still have an unpleasant but manageable fourth wave:

But we’re not out of the woods yet – that transmission assumption for AZ after 2 doses is still questionable, and we’re also dependent on other assumptions e.g. vaccine take-up in the under 50s. If that dropped to 60% for first dose and 50% for second, we’d get this:

...which is starting to look more challenging for the NHS again. So continued vigilance is required: however good the vaccines are, we still need very high take-up if we’re going to reach, or get close to, herd immunity without requiring a dangerous fourth wave. /end