In the Dark

Preface: I wrote this on Monday 16th March, before the release of a report from Imperial College admitting that the previous modelling was based on incorrect assumptions. Most of what I argued still stands but I have updated a few points.

–o–

The current Coronavirus outbreak is posing a great many questions not only about how governments should act but also about how they should communicate with the public. One aspect of this issue that came up last week was an open letter (now closed) asking the UK Government to release the data and models underpinning its COVID-19 strategy. In the interest of full disclosure, I didn’t sign it but only because I’m no longer based in the UK.

Although this letter received many signatures, I was very surprised by the negativity with which it was greeted from some corners of the scientific community. Those of us who work in astrophysics are of course used to open sharing of data and models being the norm. Some of us even see it as an essential component of the scientific method, so I was a bit shocked to see hostility from some other scientists. I think the reason was largely that it wouldn’t help to people without expert knowledge playing around with the data, getting hold of the wrong end of the stick, and jumping to erroneous conclusions. There is of course a danger of that, but in the absence of openness people are jumping to conclusions anyway and conspiracy theories are rife.

For what it’s worth, my view is that if governments can’t get those with scientific training on board then it has no chance with the general population. Astrophysicists, for example, at least understand what an exponential curve really means. Those of us who have a scientific background will not stop asking questions – nor, I think, should we. That’s how we view the world and for many equations and numbers are how we make sense of things.

So, undaunted by the calls that I should shut up because I’m not an expert, but prefaced by a clear admission that I am not an expert, I’m going to comment on a question that a lot of people are asking: why is the UK Government’s Coronavirus strategy so different from that adopted in other countries?

I didn’t watch the press conference last week that ignited this question, but I have listed to clips. The controversial issue is that of so-called herd immunity. Here is a quote by Sir Patrick Vallance, the UK’s chief scientific advisor:

Our aim is to try to reduce the peak, broaden the peak, not suppress it completely; also, because the vast majority of people get a mild illness, to build up some kind of herd immunity.

Now in the absence of a vaccine there isn’t going to be herd immunity in the sense that I understand it, but (as I have already said) I am not an expert. I think the key words are `some kind of’ in the above quote. What is envisaged is a large number of people getting infected, hopefully only contracting mild symptoms, but in any case subsequently acquiring immunity. It would be bad news for this line of thought if it turns out that people can be reinfected, as indeed seems to be the case.

After listening to the press briefings, however, it seems to me that this idea isn’t a key driver of the science policy and that Vallance simply used the phrase `herd immunity’ inadvisedly.

So if that’s not the reason why would the UK’s approach be so different from other countries? Again I preface this by admitting that I’m not an expert.

At the core of a public health strategy to combat a pandemic will be mathematical models of the spread of infection. I only know a little bit about these but I’d guess that most government agencies will have similar models (though there might be different choices of parameters reflecting different populations). But that’s not all the strategy will be based on. Among the other factors are:

1. the resources available for treating infected persons; and
2. the likely behaviour of the population (and hence the infection rate) as a result of any measures taken.
3. A decision about what it means for a strategy to be ‘optimal’.

In the first of these the UK is clearly in a very different situation from most of the rest of the world: the National Health Service has (per capita) far fewer hospital beds and, most importantly, far fewer intensive care unit facilities than other developed nations. The latter, in any case, run at close to capacity even at normal times so the resource available is severely limited. The need to `flatten the curve’ would therefore seem to be even more pressing for the United Kingdom than in many other nations.

Update: the Imperial College report explains that previous models made unrealistically optimistic assumptions about the number of infected persons requiring critical treatment. The old strategy would have led to upwards of 250,000 deaths as the NHS would have been swamped. This was exactly what was being pointed out by ‘inexpert’ commenters on social media.

Here is a dramatic confirmation of this:

The red line represents UK current critical care capacity. No amount of ‘flattening’ will be enough to avoid the NHS being overwhelmed.

That is a difference in input, but it doesn’t explain why the UK is not taking more stringent measures on social distancing. Quite the opposite, in fact. Apparently the Government has already accepted that hospitals are going to be overrun and that things are going to be very grim indeed for a long time.

By way of support for this interpretation, Boris Johnson recently announced that the elections scheduled for May 2020 will be postponed for a whole year, rather than the six months recommended by the Electoral Commission. It is a reasonably inference that the Government does not believe that this will be anywhere near over by the end of 2020. That signals that it won’t be able to put extra resources in place on the timescale needed to deal effectively with COVID-19 as China and South Korea seem to have done.

It seems, then, that the reason for not enforcing stricter policies now is item 2 above, and it is a judgment based on behavioural psychology: that severe social distancing measures would not be effective because people would get bored or there would be widespread social unrest if folk were asked to endure them for many months. That very pessimistic view of the likely behaviour of the UK population may well be realistic but assuming it has serious implications for mortality.

My interpretation of this is that the Government thinks people won’t really pay enough attention to social distancing instructions until the body count starts to become very scary indeed which, with exponential growth leading to a doubling of cases, every 2-3 days, won’t take very long.

So that brings me the reason why I think there is no way the UK Government is going to release its modelling calculations, namely that they contain numbers for how many people are going to die over the next few months. It won’t do that because it thinks the numbers would just cause people to panic. That may be a correct call too. Those of us who work in subjects like astrophysics don’t have to worry that releasing our data and models will terrify people.

There’s also point (3) about what defines an optimal strategy, as constrained by (1) & (2). The criterion could be overall mortality, but one can imagine that a government might decide to include economic cost as well or instead. One can certainly imagine the UK Government making such a choice.

I’ll add one final comment.

Here in Ireland the HSE has increased the level of testing in recognition of the evidence that community transmission seems to be more probable than previously thought. A surge in the number of confirmed cases of COVID-19 is expected.

Meanwhile, in the United Kingdom, the National Health Service is no longer carrying out any community testing:

I suspect the reason for this is a combination of (1) and (2). Counting deaths rather than infections is arguably a more reliable indicator of the growth of the epidemic and it is certainly cheaper. Moreover, one way of keeping the numbers down to avoid frightening people is to stop counting them…

Update: As of yesterday Germany had 5813 COVID cases and had 13 deaths; Norway had 1356 cases and only 3 deaths. The UK claimed 1391 cases but 35 deaths. These numbers provide drastic evidence of undercounting cases in the UK.