Why statistics mislead information about the spread of the covid-19 virus
The Covid-19 pandemic has become evident to how statistics can change from being worthless to being a truth-bearer. In Sweden, it has been used to demonstrate how an entire population is being threatned, illustrated with numbers of deaths due to Covid-19 (the statistics is questionable as Covid-19 in the majority of the deaths was not the cause but a contributing cause).
There live approximately 10 million people In Sweden. When the death tolls reached 400 media spread a perception that Sweden was about to tear apart by a virus – that society was collapsing under the pressure of a virus that in a month killed more than an ordinary influenza does in a whole season. Statistically, however, the figure was modest; 0.00004%. On April 29, 2020, the death toll was 2,355, or 0.0002%. Continuing the trend in a linear curve, we could by the turn of the year 2020/2021 expect death tolls of 10,000. Statistics shows that 0.001% of the population by then would die due to Covid-19.
If you don't like statistics, you can easily dismiss it by saying that it doesn't tell anything about how many people actually die. 0.001% doesn't sound too bad, while 10,000 dead natural people – someone's family or friend – mean everything. So – statistics are worthless, numbers are everything. It's just that I have realised that statistics is experiencing some kind of renaissance. Suddenly, the statistics have become almost as important as the numbers.
Latest in the line of statistical recognition is Professor Neil Ferguson's comparison of the death tolls between the State of New York and the Stockholm region. The state of New York has had about 17,600 Covid-19 related deaths while the Stockholm region has had about 1,300. In relation to the population it amounts to almost the same percentage - 0.0009% versus 0.0006%. Numbers no longer matter – the percentage shows that the Stockholm region has as many deaths as one of the world's largest metropolises. With that said, what Ferguson is trying to say is that the Swedish strategy of not shutting down society is as unprofitable as the American shutdown strategy - or in other words; had Sweden closed down society, the death tolls would have been much lower. Arguments, however, presuppose that the preconditions are equal, which they are not. The state of New York is populated by approximately 19.5 million people spread across 141,205 km2 (138 people per km2). The Stockholm region is populated by approximately 2.4 million people spread over 6,519 km2 (368 people per km2). The state of New York has about 37,000 intensive care units (0.002 units per inhabitant, or 0.28 units per km2). The Stockholm region has 238 intensive care units (0.0001 units per inhabitant, or 0.0238 units per km2). What do these statistics tell us?
- The Stockholm region is more densely populated than the state of New York.
- The number of intensive care units is denser in the state of New York than in the Stockholm region.
Both factors indicate that the infection is more likely to spread faster in the Stockholm region than in the state of New York, while access to adequate care units is slightly better in the state of New York than in the Stockholm region (regardless of care quality). A shutdown of Swedish society might have slowed down the spread of infection. But it is also likely that a shutdown may had accelerated the spread of infection, given infected people in close relationships are stuck together in contained spaces in combination with the lack of adequate care places in the region. To me statistics are not meant for future forecasts but for showing established facts. It is possible to build forecasts on statistics, but statistics as such are not anticipatory. Therefore, it is impossible to make assumptions based on statistics about what would have happened if we had acted differently.
I guess it is up to each and everyone to decide whether the statistics are worthless or a truth-bearer. To me it means everything in combination with explanatory facts and nothing by its own.