Fewer Dead Souls?

By Zugzwang

In a previous essay, I speculated that somewhere in the bowels of Public Health England (PHE) there was someone who had worked out, like Pavel Ivanovich Chichikov, a method of making use of dead souls. At the time, I wondered if these dead, reported daily on the Government’s Covid dashboard, were being created literally out of nothing, as there was no trace of them in the detailed death tabulations produced by the ONS. As it turns out, that part of my speculation was incorrect.

The Chichikov in question was a Professor Newton (John, not Isaac) who had decided at an early stage of the epidemic to define death-by-Covid in such a way that he could borrow genuine deaths that were properly registered as being from other causes and claim them, for the indefinite future, provided the deceased had tested positive at some point for COVID-19. He feared, it seems, that without this resource, PHE might have been “underestimating deaths caused by the virus in the early stages”. This is a very strange explanation, as one thing that was absolutely not achieved was to avoid underestimation in the early stages. You may recall a chart (originated by Carl Heneghan and reproduced in my earlier essay) that showed clearly that PHE had underestimated Covid-related deaths during the peak of the epidemic, but had been making up ground to an ever increasing degree, more or less ever since. Under the special definition adopted, PHE was able to continue combing through registered deaths and matching them up retrospectively with the NHS England numbers of people who had, at some some time in the past, tested positive.

Why does this matter? And why is it so crazily like nineteenth-century Russian satire? It has really mattered a great deal, in all manner of ways. The main purpose of daily statistics is – how could it be otherwise? – to give the Government, and the rest of us, a daily unfolding view of the epidemic. Although death statistics tell us mainly how many people were getting infected three weeks earlier, and so are backward looking, the trend in the curve can be projected forward (within some credible interval) to tell us the direction of travel. Deaths, in this world, are relatively hard currency, compared with the results of PCR tests. Large numbers of deaths frighten people (erode confidence, if you will) whereas a well-defined trend in the right direction does the reverse. If you are not in the business of deliberately maintaining fear in the interest of some wider purpose, it makes sense to tell the truth, and even more sense to give a grounded analysis of what the numbers are saying. With that in mind, let us turn to the difference between the PHE-defined numbers of Covid deaths and the newly-cleansed series (where you have to die within 28 days of testing positive to be counted).

What we need from these numbers (and so, one would think, does Mr Hancock) is a feel for how large the numbers are now – the direction of travel if there is one – and where the process terminates (if it does). The problem with the published numbers is that they have not only failed to deliver on all three of these objectives, but have obscured and muddied our understanding of the truth.

The following chart picks up the story at a point when relations between the two narratives were just starting to break down:

From this we can see that the published deaths exaggerated the true position hugely. By the end of July, the daily total was four times higher than it should have been, and by the end it was five times. We had a report on August 11th of 100 deaths. On the new basis, the number should have been 11. Yes, 11. The numbers actually dying on an individual day have fallen more than that, and it now emerges that only on one day this month have the deaths exceeded 10. Over the last week, total deaths are actually lower than the equivalent number in Italy. So the cumulative total has encouraged us to maintain an unnecessary and unhealthy anxiety, as well as creating the false impression that we have a higher infection fatality ratio than any other country in Europe.

What about the direction of travel? Well, we should have been told weeks ago by the Government that deaths were falling fast, rather than that they were mysteriously plateauing, thus showing the difficulty of controlling the epidemic. This is the trend since the beginning of July, with a trendline that projects quite clearly (and with a very decent R-squared of 0.93) that deaths should be around zero by about now.

And that, indeed, delivers the third item, namely an end-point. Of course, there are still about 600 people in hospital in England, and about 10% of them are in ventilation, so the numbers will not actually be zero every day. But we can expect them to trickle along near that level.

I hope you will be saying to yourself that this ignores the elephant. What the Government has been doing is continuing to operate under the false assumption that our resources should be devoted to combating an epidemic that finished many weeks ago. The numbers have borrowed deaths from elsewhere and dressed them up as something they were not. Where did they come from? Chichikov does not need to balance his books, but we now know that these are genuine deaths that could reasonably be classed as part of the collateral damage of the lockdown approach. We know that there is a really frightening backlog of people who should have been screened for cancer, or taken into hospital following heart attacks or strokes. Many of these people will actually die, as some have already, and their numbers will only increase if we do not urgently go back to business as normal, at least in healthcare.