Thread by @BillHanage

Trying to write a couple talks for next week based on this new preprint from us on how we expect vaccines to perform against variants with enhanced transmissibility, some immune evasion, or both. A🧵may follow as I sort my thoughts out
medrxiv.org/content/10.110… 1/?

One important thing - it is always important to compare impacts of vaccination against the alternative, so we've estimated infections averted for a simple model, and varied stuff like time and pace of vaccine introduction and the point at which variants emerge 2/?

I called it a simple model, but it has quite a lot of compartments. The arrows are the possible ways people move from S (susceptible) through being I (infected) or V (vaxxed). Resistant (thanks to immunity from infection R) and various breakthroughs 3/?

It is simple though in that it doesn't include a whole load of other things like illness or even death, let alone how they vary by age, because the major outcome of interest is like I said, infections averted 4/?

We have 4 variants in the contest.
V0 - which is exactly the same as the original virus. This should behave like the original virus if we've not built any weird stuff into our model structure (for more on that and how to avoid it, see sciencedirect.com/science/articl…) 5/?

V1 - has no immune escape, but it *is* 60% more transmissible. Think Alpha.
V2 - prior immunity only 60% as protective. Think Beta.
V3 - both. This was originally considered as Gamma, but it can be taken as similar to Delta too (was a bit pessimistic for gamma) 6/?

Sanity check - V0 behaves as expected. Good! (note these are log scales so they don't look like pretty epidemic curves) 7/?

This is more complicated. V1 arises, rapidly reaches a higher peak and declines. V2 can reinfect people in the absence of vaxx but gets stopped in the lower panel. V3 though is harder to handle - but needs enhanced transmissibility along w mild evasion 8/?

However, and perhaps hard to wrap your head around, vaccination still averts MORE infections, even with V3. Because the increased transmissibility means we'd expect more of them in the absence of vaccine 9/?

I am just going to point out that we think breakthrough infections, while they *can* be serious, are in general milder than they would be in a host with no immunity. Just leaving that here while we think about how useful vaccines are even w variants... 10/?

Here are the infections averted by timing and pace of vaccine rollout. Note the variants are always showing up ~9 months in 11/?

Key conclusion - Reinfections and breakthrough infections remain rare with moderate immune escape unless aided by transmissibility. It's the transmissibility! 12/?

Now we are hopefully convinced that transmissibility matters, let's ask what the impact of relaxing all other interventions part way through vaccination would be. That will ⬆️transmission more. We'll assume it happens when 50% are vaccinated 13/?

Oh dear. Note that this is excess in comparison with previous slides. Other interventions continue to help as vaccines are rolled out, and this is *especially* important for more transmissible variants 14/?

There is much more in the paper including varying vaccine efficacy, and allowing more immune escape. I urge you to go look at it if you like. I have a few closing comments 15?/

This is a modeling paper and as such contains a lot of assumptions. However the headline conclusions that vaccine evasion is less important than transmissibility is remarkably robust. There are some important implications for immune evasion and delta 16?/

Not sure I can put it better than the paper 17/?

I may add to this thread as other thoughts occur to me, but for now thanks to @bradfordptaylor @rebeccajk13 and @mlipsitch for contributing, but especially to @maryebushman for leading the work 18/end