Similar to many others, over the weekend I’ve been trying to wrap my head around the strategy behind the UK government’s ‘delay’ response to the coronavirus pandemic (bit.ly/33s99b6) (1/23)
Here are some of my thoughts, but I’d be interested to hear what other people think about the strategy (particularly as my expertise is fairly far away from any relevant field) (2/23)
Firstly, it seems clear that the government believes a) that the virus cannot be practically contained before a vaccine is developed, and b) that herd immunity through infection and recovery is achievable (i.e. that such immunity is fairly stable) (3/23)
These points are by no means accepted by everyone. In particular, we don’t have much data on the persistence of immunity. Immunity to the original SARS-CoV does seem to be maintained for years (bit.ly/2UcI0op), but SARS-CoV-2 may be different (4/23)
If we accept these points though, the strategy seems to be based on two pieces of information: firstly, the demographic profile of the UK population: (5/23) 
Secondly, the mortality profile of COVID-19 infection (the below is based on data from Wuhan): (6/23)
While we can be very certain about the demography of the UK, information about the mortality rate of this new disease is much less reliable. It will be affected by lifestyle differences, how overwhelmed our healthcare system becomes, etc. (7/23)
But a consistent pattern is that mortality amongst the elderly is MUCH higher than among the young. I’ll use the above data for the sake of argument, but the basic principle is similar even if we use fairly different numbers (8/23)
To start with, we can imagine a base case where we just allow the disease to spread freely though the population. Ignoring a lot of complicating factors, let’s assume that each of these age brackets gets infected equally (9/23)
If the infection rate of each age bracket is equal to the predicted herd immunity fraction (60%), we would expect 820,000 deaths: (10/23)
If, on the other hand, we ensured that only the youngest 60% of the population catch the disease, we could substantially reduce the death toll to 78,000 (current mortality figures from Italy suggest it would be even lower than this): (11/23)
Basically, the strategy seems to be to make the final situation look as little like the base case as possible, and as much as possible like the second case. Hence the intention to ‘cocoon’ the elderly and vulnerable in the coming weeks (12/23)
However, there seems to be a bit of a catch: as I understand it, the herd immunity fraction not only indicates the lowest level of immunity required for the virus to stop spreading, it also indicates the highest level of immunity a population will ‘naturally’ achieve (13/23)
So if we have two populations, a young one that achieves herd immunity and an isolated, older one with little immunity, remixing the two populations together will dilute the immune fraction and allow the disease to start spreading again (14/23)
As toy numbers, let us keep 30% of the population isolated and allow the virus to spread through the remaining 70%. 60% of this 70% then becomes immune (reaching the herd immunity fraction) (15/23)
When we mix our two populations back together though, only 42% of the population is immune. As this is below 60%, the disease can spread again (albeit slowly). Is this an issue? (16/23)
Anyway, back to the main thread. If we accept that we can no longer contain the spread, is scenario 2 the best we can do? (17/23)
We might consider decreasing R0 (the reproduction number of the virus, i.e. the average number of people infected by an infected person) by encouraging social distancing, increasing hand washing etc. (18/23)
A herd immunity fraction of 60% is based on an R0=2.5, and in general the herd immunity fraction H is roughly equal to (R0-1)/R0 (the actual value depends on the dynamics of the spread). So pulling R0 down will decrease the population fraction that ends up infected (19/23)
But the extent to which this is achievable is limited. In particular, behavioural scientists advising the government think that the big changes that would really pull down R0 (shutdowns and isolation) can only be maintained for a short period of time (20/23)
This is why these measures aren’t being activated at the moment – given the strategy, it may even be useful for the infection rate to be as high as possible during the upcoming cocooning phase, so as to ensure it has abated after isolation ends (21/23)
So don’t necessarily expect to see us taking similar measures to other European countries over the coming weeks (22/23)
It’s turned out a long thread, apologies! Any thoughts, though? (end/23)
