Age-Adjusted Excess Mortality During the Pandemic in Sweden

One of the most startling facts of the pandemic is that, between January of 2020 and June of 2021, Sweden had negative excess mortality – the country’s age-standardised mortality rate was below the five-year average.

Despite all the warnings about “disaster” and “folly”, Sweden actually saw fewer deaths than usual over this eighteen month period. (The ONS data on which this conclusion is based only go up to June of 2021. But it’s a good bet the picture hasn’t changed much since then.)

Even lockdown sceptics might find this difficult to believe. Sure, Sweden didn’t turn out to be the disaster that lockdown proponents claimed it would be. But can it really be true that there were fewer deaths than usual? What about those reports of Swedish hospitals “filling up with patients” in the winter of 2020?

It can really be true. And those reports are easy to reconcile with the mortality data. You just have to remember two phenomena: the ‘dry tinder’ effect, and mortality displacement.

For those who’ve forgotten, here’s a brief reminder. The ‘dry tinder’ effect refers to the tendency for mortality to increase following a period of unusually low mortality, due to the presence of large numbers of very frail elderly people, who ended up living longer than expected.

Mortality displacement is just the opposite. It refers to the tendency for mortality to decrease following a period of unusually high mortality – for deaths to be ‘brought forward’ by some deadly event, such as a pandemic, heat wave or unusually cold winter.

Once you’re familiar with these two phenomena, you realise it’s perfectly possible for hospitals to fill up temporarily without the mortality rate going up at all in the medium term (over a period of several months or a year, say.)

The chart below shows the weekly age-standardised mortality rate in Sweden as a percentage of the five-year average, from January of 2020 to June of 2021. (I made the chart using data published by the ONS.)

In the winter of 2020, mortality was below average, leading to the build up of ‘dry tinder’. When the pandemic hit in March, mortality increased substantially. It then fell back below average during the summer of 2020, giving rise to mortality displacement. This pattern then repeated itself over the subsequent 12 months.

Aside from the obvious fact that Sweden’s approach didn’t lead to disaster (far from it), what’s the major lesson here? It’s that you won’t get a reliable picture of the pandemic if you only follow the news. After all, ‘Hospitals Fill Up With Patients’ grabs your attention, whereas ‘Another Week of Below-Average Admissions’ doesn’t pack the same punch.