This article originally featured on the Conservative Home website.
Liam Fox is a former Secretary of State for International Trade, GP and army medic. He is MP for North Somerset.
One of the features of reporting of the Coronavirus emergency has been the tendency to construct league tables that tell us which countries have the highest number of deaths or the highest number of hospital beds or any other number of variables.
Often these are not only hugely inaccurate, but paint a damagingly distorted picture which can influence public opinion and, by doing so, public policy.
One well-established example is to talk about the number of cases in a country, when what is actually meant is the number of cases who have been tested positive for the coronavirus. Often, we have no idea how many cases may genuinely exist, since a significant proportion of the population will not have not been tested at all.
Another is to report the number of total deaths without taking into account the population of the country concerned – which makes comparison almost meaningless.
For example, the United States, with a population of over 300 million people, is obviously going to have more cases than a European country such as Spain, Belgium or Italy with populations of 47 million, 11.5 million and 60 million respectively.
Yet, if we take a standardised measure – deaths per million of the population – we find that it is in fact Belgium that has the highest rate of the major nations with 496, Spain second with 440 and Italy third with 385. France has 288 and the United Kingdom currently has a rate of 232.
The United States, by comparison, has had a death rate of 119 per million population, well below that of most major European nations. This, of course, doesn’t stop those who are innately hostile to Donald Trump’s administration from ignoring the standardised data. “Attack first, verify later” isn’t a trait confined to American politics.
Even these more standardised comparisons are flawed because of the different stages of the pandemic cycle in each country. The true extent of our real differences will only become apparent with time. Rushing to instant judgements on how any country has “performed” runs the risk of both undermining the truth and public confidence.
For example, the low US figure disguises massive discrepancies between New York, the worst affected area, and some of the more rural states – so how is “America” to be judged? In fact, talking about nation states as though they are homogeneous entities often misses crucial differences within those states themselves.
Perhaps the most striking example is Italy. Although the overall death rate for the country is 385 per million, the Lombardy region, which includes Milan and is the epicentre of the European outbreak, has a rate of 1178 while Lazio, which contains the city of Rome, has a rate of only 57.
To talk about whether Italy, treating it as a single entity, has done well or badly is to utterly fail to look at reasons for the underlying differences. These may be related to population density, the age and health of the population or underlying social conditions.
These regional differences are found in most countries, including the United Kingdom – where London has a death rate per million of 417 (higher than Italy’s overall figure) while the South-West has a rate of only 116 (lower than the US).
Most regions sit in the middle, with the Midlands at 265 and the North West at 262. Scotland and Wales sit at the lower end of the spectrum at 164 and 170 respectively. This may be due to any of the factors already mentioned, or the fact that other parts of the UK may simply be lagging behind London in terms of the stage of infection.
Another area of oversimplification is comparing different countries’ health systems. One major newspaper ran a piece recently entitled “Oversupply of hospital beds helps Germany to fight virus”. It may well be true that German capacity removed the need to alter normal health priorities in order to accommodate the surge of Covid-19 patients. But this can only be a part of the story.
The “oversupply” of beds in the German healthcare system is a controversial subject, with many claiming that, because insurance companies have been paid for bed occupancy, there has been a perverse incentive to keep patients in bed longer than necessary with the consequent need for higher bed numbers.
The key question is whether there is a direct relationship between the number of acute beds (generally measured as per thousand population by OECD) and the number and rate of Coronavirus-related deaths.
It is certainly true that Germany has the highest number of acute beds in Europe with eight beds per thousand population, but Belgium at 5.6 and France at 6.0 respectively have substantially higher death rates than the UK, with our lower bed numbers at 2.5, or the Netherlands at 3.3. Again, far more complex factors must be at work to create this discrepancy.
One statistic that is worth looking at is excess mortality ie the number of deaths above the longer term average for any given time period. Looking at the peak winter months of December to February, the average from 2015-2019 is 146,954. For the three months from December 2019 to February 2020, the last three months available, the number is 147,828.
In other words, the excess mortality over the same period a year ago is 874 deaths. We will need to see what the April figures show to give a comparison with the first quarter average over the past five years, which stands at 151,932.
If the figure is low, the Government will claim that its measures were successful and opponents will claim that it shows the lockdown and the consequent economic cost was unnecessary.
If the figures are high, the Government will claim that the measures taken stop the people from being even higher, but opponents (probably the same ones) will claim that the Government’s actions were too little and too late.
The bottom line is that we cannot declare success or failure from a single dataset. To simply choose one factor and imply that it is the key determinant is to create a potentially misleading picture. To be fair to the paper in this case, the story was much more nuanced than the headline.
If we are to have a rational debate about how we deal with the current pandemic and how we learn lessons from it then it must be based on a sound assessment of standardised data. Supposition, insinuation and misrepresentation will not help our understanding of what needs to be done now and in the future and everyone – the media, politicians and any other commentators, including members of the public on social media – need to understand our responsibility in that process.
