Lies, Damn Lies and Statistics: The Media's Misuse of Numbers during the Pandemic

This week the respected German broadsheet Handelsblatt published an article claiming that the German government had leaked that the Oxford University/AstraZeneca vaccine only gave pensioners eight percent protection against COVID-19. They hastily released a clarification: the article resulted from a mishearing of a statistic: that 8% of the subjects in the AstraZeneca efficacy study were between 56 and 69 years of age. From this mishearing came a snowballing and a potentially damaging piece of misinformation

Misinformation comes in many forms. It’s easy to spot the obvious lies: that COVID-19 doesn’t exist, that it’s spread by 5G towers, that Bill Gates is trying to fill us with microchips.

Misinformation based on a misreading of statistics may be a simple mistake, as with Handelsblatt, but others may be much more cynical: that we are overreacting to a disease with only 1% mortality, that the NHS occupancy is much lower than normal and that there hasn’t been an increase in deaths compared to previous years. These lies are seductive and often used to justify more myths: that lockdowns don’t work, that lockdown has caused a rise in suicides or that wearing masks doesn’t work.  It’s time to challenge the misuse and misrepresentation of statistics.  

Let’s start with that line about how we are overreacting to a disease with only a 1% mortality rate. This is actually wrong as a statistic. The World Health Organisation COVID-19 Dashboard reports a mortality rate of 2.1%. But for the sake of argument let’s run with that figure of 1%. A  mortality rate of 1% doesn’t sound like much at all. It’s far less than the mortality rate of Ebola (50%) and another disease caused by a coronavirus, MERS (35%). 

But that’s not the point with a proportion. It’s not the percentage itself that’s important it’s what it’s a percentage of.  If you have 100 people infected that means only one person dying - not a big number at all. But if 100,000 people are infected that means 1,000 dying. Not an insignificant number. On the 8th January 1,035 people died in the UK within 28 days of testing positive. To put that in context 67 Britons died during the September 11th terrorist attacks. The right-wing media rightly admonish terriorism and honour those who died in terror atrocities. However, on 8th January COVID-19 killed 1545% the number of Britons murdered on September 11th. In total over 100,000 people in the UK have died from COVID-19. Double that killed by the Blitz.

This focus on 1% mortality also ignores other basic facts of COVID-19. Whilst ‘only’ 1% of patients with COVID-19 will die about 15% will need hospital care. About 5% of those infected will need to come to critical care (ICU). Once again, if 100 people are infected that’s 15 coming to hospital and 5 needing critical care. If 100,000 are infected that’s 15,000 needing a hospital bed and 5,000 needing ICU.  That’s without mentioning the impact of ‘long COVID’. This is not a disease where just 1% of people die and 99% have the sniffles and are fine. Painting it otherwise is a false dichotomy and a lie. 

Of course the impact of the pandemic on non-COVID-19 patients has rightly been brought up. However, this again misses the point. If a hospital’s capacity has been taken up with patients with COVID-19 then that means there is no space for patients who’ve had a heart attack or need emergency surgery. The NHS would cease to function.      

The other lie which is peddled is that because a percentage of ICU beds being used is less than the same time in a previous year this means that critical care is not as busy. 

This once again is a misunderstanding of proportions. It’s not just the percentage of critical care beds being occupied it’s the total of critical care beds available. 25% is less than 50%. But it’s a simple fact that 25% of 200 is the same as 50% of 100. Just looking at the percentage ignores the fact that ICU capacity has been increased in response to the pandemic. For example, in March 2020 Northwick Park Hospital in North West London increased their critical care capacity from 22 to 52 beds, an increase of 236%.  Even if the percentage of beds occupied is lower if the number of beds has gone up this still represents a greater demand on the NHS. 

Northwick Park is part of London North West University Healthcare NHS Trust. A look at the monthly situation report on the number of available and occupied Critical Care beds on the last Thursday of the month for 2020 tells us that at the end of January last year London North West University Healthcare NHS Trust had 33 adult ITU beds in total of which 28 were occupied. That’s 85% occupancy.

If we look at the Urgent and Emergency Care Daily Situation Reports 2020-21 we see that on 11th January 2021 London North West University Healthcare NHS Trust had 102 adult critical care beds of which 78 were occupied. That’s 76% occupancy.

So if we look at occupancy alone London North West University Healthcare NHS Trust had a smaller percentage of their adult critical care beds occupied on January 11th 2021 than at the end of January 2020. But because they had massively increased their critical care capacity the number of beds actually occupied was much larger: 78 compared to 28 or an increase of 279%. Focusing on occupancy alone ignores the true picture of the pandemic.

Another line often taken has been that “we should just use the Nightingale Hospitals”. These are seven hospitals set up in England as well as one each in Scotland, Wales and Northern Ireland  in order to create extra capacity for COVID-19 patients.  

This is misleading again. It’s all well and good to set up a hospital, another thing entirely to run it. Even in April 2020 there were warnings that the NHS could not staff the new hospitals. We could fill the hospitals with patients but without doctors, nurses and other healthcare staff they won’t be looked after. NHS staffing shortfalls pre-exist COVID-19 but the pandemic has exposed how threadbare the situation is. 

Another theme in the misuse of statistics has been to claim that mortality this year has been no higher or even less than in recent years. A rather egregious example of this was a Daily Mail article from 20th November 2020 entitled ‘What they don’t tell you about COVID’ which claimed the number of weekly deaths is currently “barely any higher” than the maximum level from the previous five years. This was wrong and torn apart both on Twitter and by Full Fact.

The figures came from a character called ‘Statistics Guy’ on Twitter who joined in April 2020. He claims to “do uk statistics for ordinary people. cutting (sic) through the waffle on your behalf”. In calculating the figures for the Daily Mail he missed out data which would have changed his conclusions as well as ‘adjusting’ figures for population growth. This was bizarre: as though the basic number of COVID-19 deaths can be negated because there are more people in the UK than in previous years. 

A more recent example on 8th January, the same day that 1,035 people died of COVID-19 in the UK, former pathologist John Lee claimed on Julia Hartley-Brewer’s talkRADIO show that:  

“We're seeing mortality that's well within the envelope of what normally happens this time of year. The last five years have been a below average number of death years if you look at the ONS data anyway compared to the last 27, which is how far their data go back...we’re below the average point of the deaths at this time of year...”

This again was torn apart by Full Fact. The five-year average is standard practice when looking for excess deaths so by not using it seems that Lee was trying to cherry-pick the data. Regardless, even adjusting for population he was wrong on both the 27-year and 5-year average. 

* * *

"Figures often beguile me...in which case the remark attributed to Disraeli would often apply with justice and force: 'There are three kinds of lies: lies, damned lies, and statistics.'"

Mark Twain

Mark Twain has been proved right many times over, especially during the COVID-19 pandemic. Statistics sound good and, superficially at least, given credence to an argument. We should be careful when we hear or read a statistic to think about what’s actually being shown. So too should those of us speaking far outside our arena of expertise. Toby Young, a right-wing journalist of no scientific or medical training, has recently had to apologise for a 'significantly misleading' column for the Daily Telegraph in which he claimed the common cold could provide "natural immunity" to COVID-19. Let’s remember Mr Twain. There is nothing more contagious than a lie. And no more dangerous a lie than a statistic.