Thread by @ewanbirney

Taking stock of COVID in October - for followers, in particular journalists. My scope is broadly across Europe, with a particular focus in the UK.

Context: I am an expert in one area - human genetics; I have broad data science / data analysis skills; I know experts from virus genomics, testing, infectious epidemiology and clinical trials; I am someone who deals with uncertainty by aiming to gain more knowledge.

I have one major conflict of interest in that I am a long established consultant to a company (Oxford Nanopore) that makes a new SARS_CoV_2 test. It's not so relevant in this thread, but it's worth knowing if you are reading stuff from me for the first time.

I always think it is worth reminding ourselves what we know: SARS_CoV_2 is an infectious virus that causes a severe disease, often leading to death, in a subset of people.

We know a number of things about it; it is not the most infectious virus, and it seems to infect in a bursty, context specific way (some estimates are that 80% of infections do not pass on; 20% do, but often to many people - you can call these super-spreading events if you like)

We know the disease it causes better; there is some human genetic predisposition to the severe disease (not enough to be predictive on an indivdiual level). We know being old, obese and being male are each major contributors to having severe disease.

We have new treatments due to careful clinical trials and meticulous clinical work. Rather than ~1 in 2 people dying when they enter ICU (as in March) with this disease, it is now ~1 in 4. This is excellent progress but is not a good enough reduction in mortality to tolerate.

We now know there is a long term follow-on disease, similar in some ways to other viral induced autoimmune disease, colloquially called "Long COVID". Frustratingly we don't know rates or predispositions yet beyond it is younger and less sex biased than mortality.

The presence of LongCOVID at an earlier age range further complicates guiding us out of this pandemic as any broad range of infection in the population will cause this to happen and currently we don't really know the rates of this.

Vaccines are emerging faster than ever expected; over 10 are now in the final "phase 3" trials, and there is quite an expectation that some will report by the end of this year or next year.

The bar for vaccines to be effective at to be approved is set quite modestly, because even a modest vaccine would be useful. We simply don't know until the trials are done whether the first vaccines will be amazing, reasonably good or just scrape over this modest bar.

Since July there has been a strong rise in cases across Europe - Spain, then France, then UK + Netherlands + Belgium and then Italy and, now, finally Germany. Different countries are at different stages in this second wave

It's somewhat academic what drives the variation between countries here beyond the fact that even Germany now has a rise in cases; Germany has been the stand out European success story in 1st and 2nd wave for their response, based on well provisioned public health and testing.

The fact that Germany has moved from "broadly in control in all areas" to "not all areas in control" (noting that the response is Lander based, so talking about Germany as homogenous whole is somewhat misleading) says two things.

Firstly some exogenous factor, with the simplest explanation being weather and temperature, is making the virus more infectious - similar to other coronaviruses. Secondly it says the margins are thin - one can move from control to not in control in quite a small window.

Before one gets too depressed about the future of this, it is clear that the combination processes in nearly all countries have seriously dented the reproduction rate of the virus - it is now not the hair-raising rise in infections and hospitalisations as in March.

Rather it is a far slower, but just as ultimately lethal march upwards of infections, hospitalisations and ultimately deaths. Although the rate is less, frustratingly the ultimate outcome is similar.

One might be tempted to think about separating out the "at risk" groups of society and focus effort on protect those. This is extremely unlikely to work well; firstly there are just a lot of people at risk and those people need interaction with other people to live

(this is particularly true because they are older; more support is needed from mundane aspects of living through to healthcare provision outside of COVID). Secondly the presence of LongCOVID utterly complicates the risk profiling beyond repair.

One might think the right way is to manage some steady state of "tolerated" infections, hospitalisations and deaths. Again, here the exponential nature of epidemics are against this; epidemics either grow or shrink. There is basically no middle ground.

Israel, who dealt with their 1st wave well, tried to match the infection rate to their health care capacity in the 2nd wave. It was a mess, and it is always going to be a mess as exponentials with complex lags can't be easily controlled.

Having healthcare capacity is useful - very useful - it gives times to get things wrong and fix them; it ... provides healthcare for the inevitable COVID infections which will occur - but healthcare capacity by itself is not a strategy.

So - what is a strategy? One can frame this as "lockdown" vs "Test-Trace-Isolate" but that's the wrong framing. Lockdown has many variations. TTI has many details about how it works in different scenarios. Having these as "opposing" strategies is madness. It is both.

Rather the question in my view is what are the hierarchies of regional (national) broad interventions (non-pharmaceutical interventions) that work with as good as it can be TTI to lead to contained transmission of the virus.

There are two different axes here: one is about getting the most out of TTI. I suspect there is no country with the best possible TTI; certainly many European countries (UK included) can do a lot better. This statement is *so* easy to state in a tweet and so hard to realise

(Editorial; there can be a tendency now to go to big picture, often politics of decisions about how TTI was set up in each country. Clearly some countries have done this better than others; some aspects work better in some countries vs others.

...but the question is not what can one do with a time machine, or should do in a future pandemic - there will be time for those arguments - the question in each country is what is the shortest, more certain path to improvement. It's actually a harder question answer in my view).

There are a host of things from a better COVID App (France) to using the App data better (UK) to better situational awareness + national coordination (Spain) to higher testing capacity (Czech Republic)

A common theme in Europe in my view (with the caveat I don't know anything close to all the trace systems) is still not fully using the Japanese style back-tracing with action, where known infection events that spread to >1 person are discovered and rapidly tracked down

The other axes is having a clear escalation path of population wide NPIs - ideally with good evidence that they work (though it is immensely complex to know) and the impact on other aspects of life, wellbeing and livelihoods.

This axis should be created, debated and fixed upon ideally before use (this doesn't seem to happen that cleanly in the big countries at least).

At this point the strategy becomes - add enough of your NPIs to achieve control given your current level of TTI; don't be afraid of overcompensating because it is better to be in control rather than out of it; then remove NPIs in a slightly horrible game of Jenga as TTI improves.

Always aim to improve TTI. There is basically no level of TTI which one should not aim to improve on.

Again, this is far far easier to tweet than to do.

A final point; short term it is a depressing prospect - many places across Europe have lost or are losing control of viral transmission. But mid to long term I’m more positive

I’m positive because there is clear headroom in TTI to realise; I’m positive because there are more drugs being tested; I’m positive because there is a diverse wall of vaccines in testing