The number of Swedes who died in April and May was 19,149, according to preliminary figures from Statistics authority of Sweden (SCB). It’s a lot, but not historically much. At the turn of the month December-January, more people died in 1995, 1993 and 1988. Relatively the population is not even one of our top 100 deadliest monthly pairs.
But it should not be used to shrug off covid-19. The excess mortality during April-May is uniquely large.
How does that work?
Every week, SCB reports on how many people die in Sweden. I have previously written that April 2020 was one of our deadliest months ever, but that several months during the 90s were more deadly, because the flu was unusually severe. When May now have been reported, we can compare patterns also over a two-month period.
From December 1993 to January 1994, a total of 20,245 people died in Sweden, which is over a thousand more deaths than April / May 2020. There are two more severe flu periods that have caused more deaths than the Swedish covid-19-months:
In the 1990s, Åke Nilsson was responsible for SCB:s population statistics. He remembers the high death tolls during the flu months of the 90’s and he remembers the silence that surrounded them. When he now looks back to compare the mortality of our two covid-19 months, he sees that the flu still killed more than covid-19, when counting absolute numbers. He even sees hundreds of monthly pairs where Sweden had more deaths, if you measure the death toll relative to population size.
In 2009, mortality was higher in the months of January and February, when 1.86 per thousand inhabitants died, compared to 1.85 in april/may 2020. Between the years 2000-2009 alone, there are as many as nine monthly pairs with higher relative mortality-numbers.
How can that be?
– This is because Sweden have had such success in reducing mortality every year. Life expectancy is increasing and living conditions are improving. And on top of that our medical advances, says Åke Nilsson.
During his years at SCB Nilsson could use his spreadsheet to follow how Swedens development affected the over all health. The new medicines in the 1980s – which caused cardiovascular diseases to suddenly decrease – meant that men’s average life expectancy increased faster than women’s. And the new life expectancy, in turn, meant that society had a growing group of seniors, which meant that the pension system was replaced. And then the flu waves. Which several times each decade meant that a few thousand elderly people just died – disappeared! – without anyone reacting. Åke Nilsson, now 80, thinks this is still strange.
– Yes, it’s strange. The statistics have been around for all these years, but still nothing has been done about it. It is a nonchalance shown towards elderly care. But this is nothing new and should be difficult to change, says Nilsson.
When covid-19 came to Sweden, it was during a period when Sweden usually does not see very many dead. We also came from a winter that was unusually mild, which in turn meant that we had unusually few deaths. And the winter before that, 2018/2019 was similar. From a purely statistical point of view, Sweden thus had citizens who, in an ordinary year, would have died of an infection, virus or influenza, but had now survived. It also means that Sweden 2020 have an aging population. In ten years, the proportion of men over the age of 90 has increased by almost 40 percent, women by almost 20 percent. In just one year, between 2018 and 2019, the group older than 90 years increased by 1,324 people. That was when covid-19 struck. Which makes comparisons complex.
Tomas Johansson today has Åke Nilsson’s job at Statistics Sweden. He has read Åke Nilsson’s calculations, confirms them, but emphasizes that they do not automatically show excess mortality. He himself has made calculations and compared the monthly pair April-May back in history. He has to go back to 1927 to find the same mortality – if using the same monthly couple. In April this year, 2,894 more people died than normal – that is, based on how many people usually die during an April month in previous years (the excess figure is called excess mortality). In May, the excess mortality rate was 1603.
– In total, the excess mortality during April and May will be almost 4,500 dead, Johansson says.
When comparing april/may to months like january and february, you need to understand the differences in conditions. Åke Nilsson also points this out and makes a comparison if the covid-19 pandemic had come in December. Since the groups that die from influenza are similar to those that die with covid-19, it would have been more difficult to distinguish what was the pandemic and what was a normal mortality.
– Then we would have underestimated the excess mortality.
That would not have meant that the covid-19 was harmless, but that comparison shows how statistics can be misunderstood, or interpreted differently depending on the context. A comparison with something worse back in time does not mean that anyone can relax or believe that covid-19 is the same as a flu. Or that 2020 has not been particularly deadly.
Statistics can be used to push evidence and theses in different directions, so it should be interpreted with caution. Above all, we can use it to see historical patterns and learn ahead. In our case, there seems to be several lessons to be learned about how we handled our elders.
This article is free and accessible with the help of people who contribute financially every month for journalism to be open.
You can also contribute, through the Patreon service.
The article is licensed under the CC-BY license, free to share and republish if you give appropriate credit to this post.