Lies, damned lies and, er, statistics
Updated: Jan 22, 2021
A topic which should be a simple matter of counting has become a hotly debated political issue. There is even a left-right split: the left says Covid-19 is a world-wide scourge and we must all suffer to get rid of it; the right says it's just a case of the sniffles and it will go away soon. Why the right would allow itself to be painted into this mindless corner is a mystery to me. If someone you know is dead from Covid-19, or if your local hospital is bursting at the seams and the nightly TV news features tearful nurses saying they can't cope, then you might come to the view that the left's version of events is closer to the truth.
But the disagreement over facts v. alternative facts continues, and the most hotly contested facts are the numbers of deaths from Covid-19. Enter David Spiegelhalter, chair of the Winton Centre for Risk and Evidence Communication in Cambridge, England. Professor Spiegelhalter has a set of facts which make perfect sense.
If you simply take the number of deaths as being the number of people who died and whose death certificate gives Covid-19 as the cause of death, then in the final quarter of 2020 some 29,000 people died of Covid-19 in the whole of the UK. But this underestimates the Covid-related death count because it doesn't include a significant number of people who suicided during lockdown, and it doesn't include those who died because they couldn't get the right treatment for non-Covid medical problems in Covid-crowded hospitals.
A better way to look at it is to take the average number of deaths in a non-Covid year, and ask how that number has increased in 2020. Using that statistical base, Professor Spiegelhalter found that the number of deaths from all causes in the final quarter of 2020 was up 21,000 from the UK national average. Ergo, Covid-19 was responsible for 21,000 deaths. But that is 8,000 fewer than the 29,000 people who had Covid-19 on their death certificate as the cause of death. So is there a giant left-wing conspiracy to over-cook the Covid-19 figures?
Not so, according to Professor Spiegelhalter (who was writing in The Guardian, so this is the moment for right-wing conspiracy theorists to say I-told-you-so.) I quote Professor Spiegelhalter: First, around half of this 8,000 will be people whose primary cause of death was something else: they died “with”, rather than”‘from”, Covid. Second, there has been almost no flu. But sadly, we are seeing the absence of the many elderly people who were among the 59,000 excess deaths between March and June, but would otherwise have survived until later in the year.
So the learned professor is telling us that about half of the 8,000 extra deaths had Covid-19 all right, but they also had another condition which contributed significantly to their death. Also, some 59,000 people had jumped the gun and died already, in the second quarter of 2020. To which I respond: fair enough, but where did the 'around half' come from? Sounds like pure guesswork to me. Nevertheless Professor Spiegelhalter's numbers make statistical sense, and I commend them to one and all.
Which is more than I can say for some numbers quoted by Dr Paul Kelly (pictured above), the Chief Medical Officer for the Commonwealth of Australia. There are three anti-Covid vaccines in current use around the world: Moderna, Pfizer and AstraZeneca. In large-scale tests both the Moderna and Pfizer vaccines were shown to be about 90-95% effective. But they are expensive - about seven times the price of the AstraZeneca vaccine - and the clinical trials gave the AstraZeneca vaccine about a 70% success rate. Furthermore the AstraZeneca vaccine, as well as being cheaper, is much easier to handle. It doesn't have to be stored at exotic sub-zero temperatures. So it is a tempting choice for governments, including Dr Kelly's Australian government.
Dr Kelly then defended this option by saying that in clinical trials the AstraZeneca vaccine had prevented death from Covid-19 100% of the time. He also gave it a 100% score for 'preventing severe illness' from Covid-19. To say that the statistical case for this assertion is flimsy is probably an understatement. Here are the numbers: the AstraZeneca vaccine was tested on 12,021 people in the UK and Brazil. There was then a control group of 11,724 who were not given the vaccine (but were presumably given a placebo and then exposed to the disease.) We don't have the numbers for how many who received the vaccine actually contracted Covid-19 but to get a 70% effective result as many as 3500 of the 12,021 is a plausible number. However it's fair to say that of those who received the vaccine and subsequently contracted Covid-19, no one was hospitalised, and no one died. Hence Dr Kelly's 100% numbers.
What of those in the control group? Again, we don't know how many actually contracted Covid-19, but we do know that some 10 of them were hospitalised, two of them with severe symptoms, one of whom subsequently died. So Dr Kelly's figures are based on a sample of 10 people from 11,724, and his death rate figures are based on a sample of one. I leave it to you to judge how much reliance to place on them.