Good afternoon, children! Are you sitting comfortably?
There are four learning points today.
1. Do not jump to conclusions before you have got the facts straight.
Somehow I don’t think that journalists read my blog, because they keep making this mistake. I made it myself in an earlier blog; see if you can spot it. As I said in my last blog there was unrest at Boris Johnson’s absence from the front line, and comments along the lines of “How long will it be before he is back?”; “Who is making the decisions while he is not there?”; “Are decisions going to have to wait until he gets back?”
And, of course, it turns out that BJ’s approach has been quite hands-off, and he hasn’t actually been there quite a lot of the time. Everyone assumes he was. Fact check shows he wasn’t. So he was not essential (which is what the cabinet collectively has said repeatedly). Egg on face for all those who now have to find something else to criticise, because after all someone has to be blamed for this mess, whether it’s the politicians, the scientists, the medics, the Chinese, tourists, the bats or the pangolins. Probably Boris, on the grounds that he should have been there. What’s to blame is the coronavirus, which is new, previously never seen and nasty. Everyone reacting to it can be forgiven for caution – caution when imposing a lockdown, which would inevitably have massive economic consequences, caution in emptying hospitals in case of need, caution in taking the lockdown off in case there’s a rebound – caution in balancing the risks and benefits of saving lives and causing business disruption. What point is there in spending millions on an inquiry, which will show in all probability that the only real issues were some sloth in sorting out PPE (in mitigation because no-one really knew how bad this was at the outset, there was no real evidence that draconian measures might be needed) and some sloth in sorting out testing in care homes.
So the expectation of the scientists and medics was that a lockdown and social distancing would mitigate deaths and stop the NHS being overwhelmed. Tick. The government was under huge pressure to mitigate economic collapse and so has thrown billions at protecting businesses and jobs. Tick. If there had been no lockdown the likelihood of a high death rate was almost a certainty. There hasn’t been. Tick. So those who criticise the handling of it all must answer the question: which would have been worse? Deliberately allowing huge numbers of deaths and protecting the economy? Or reducing the death rate at an economic cost? As far as I can see every commentator has effectively rejected either option. OK, guys, if you are so clever you provide the plan. We won’t give you any help, because you obviously don’t trust any of the experts on whose advice, and evidence, the current plans are based. Indeed, you want to sack them. An analogy for you. The Grand National is about to run. You look at the form, the pundits’ analysis, the likely going, and you place a bet. Your horse falls at the second. Tough, but I hardly think you will stop buying the paper or sack your bookmaker because they “ought to have known”. How could they? How could anyone have known, with Covid-19, how infectious it was, what the fatality rate was? We might now, three months in, sort of know how infectious it is but as we don’t know how many people got it without symptoms we don’t know how fatal it really is; current estimates are only based on the known cases. So get the facts right before pontificating and reading out the “horrific” death figures like gloomy clergymen at a funeral.
2. Understand how trials work and why you can and can’t do them.
I got a bit off the lesson there and sort of onto the second, which is to teach you about trials. if you want to compare two drugs (or scenarios, or whatever) you do what is known as a controlled trial. You take drug 1, and either compare it against drug 2, or against nothing (so participants don’t know which they are having, the “nothing” is a dummy pill, capsule or injection). The trial is thus blind. If the investigators also don’t know who is having what, then the trial is double-blind. An open trial has flaws but may be reasonable if you cannot blind either investigator or participant to the intervention.
In the case of Covid-19, then, what would be interesting would be to compare lockdown against nothing. Of course that cannot be done blind. In fact it cannot be done at all. How would you do it? Let Scotland be free, and lock down England and Wales, and compare death rates? What about all the cross-border flow of goods and their deliverers? What about the naughty English that wouldn’t obey the lockdown (“They aren’t doing it in Scotland, so why should we?”) or those in Scotland who are frightened enough to keep social distancing and self-isolation anyway? The cross-contamination would be so bad as to render the trial useless. Even doing it between countries would have the same problem unless you closed all the borders. Not practical, not least because lots of PPE, or the raw materials to make it, come from overseas. So we might introduce a confounder. My head spins at the thought of trying to think up a way of doing it.
3. A model is only a model
No-one is far-sighted or prophetic enough to cover all the bases. The unexpected happens. Different assumptions and presumptions lead to different conclusions. We have seen that in the wildly different estimates of the likely numbers of UK deaths from Covid-19, ranging as high as hundreds of thousands. Modelling can be no more than a guide.
4. Pre-planning on the basis of past experience may fail if something unexpected happens.
After the First World War the French built a huge chain of fortifications along the border with Germany, known as the Maginot Line, which was pretty well impregnable. Some years back bits of it were advertised for sale, and I half thought of buying a bunker, as the ones on offer were in excellent condition, but we had other commitments. Anyway come 1940 it didn’t work, for the simple reason that the German army came round the top of it by going through Belgium. I have never been clear why the French left this bit of frontier rather weak. Perhaps because the Belgians had some pretty powerful forts themselves? Perhaps forgetting that they came that way in WW1? Anyway the Germans landed paratroopers on the Belgian forts and that was that.
The defences of Singapore in WW2 pointed out to sea, and were formidable. The Japanese rapidly conquered Malaya, and so arrived at the back door, so to speak, and there wasn’t time to turn the artillery round. No-one had thought they would do that. A little later two battleships were sent up to counter the threat and were sunk by aircraft. No-one thought they would do that either, and so the ships had no escort and no air cover.
In reverse, of course, the Japanese did not plan for an atomic bomb.
Could these scenarios have been predicted? With a liberally greased retrospectoscope, yes, but not at the time. What one knows now is far different from what one knew then. Contingency planning is sensible, but only up to a point. The facts may not be there, alternative scenarios cannot be compared, models are imagination, not fact, and the unexpected happens. That’s life.
The Wry Observer 