Because of the many national and international initiatives to track COVID-19 cases and deaths, we have a relatively good understanding of the mortality impact of the pandemic. However, most of the evidence is based on deaths specifically related to COVID-19. General mortality rates may tell a different story, as COVID-19 deaths were sometimes under-reported. Particularly in the first wave, deaths may not have been recorded as COVID-19-related if testing capacity was limited, if care for other conditions was compromised by overburdened health systems, or if fear of contracting the virus led people to avoid seeking care when they needed it. We might also consider whether some groups were protected by home-centred living conditions during the pandemic, and whether deaths were "dry tinder" - deaths that were inevitable in the near future.
What is the real toll of the pandemic in Sweden? In a paper just published in the journal Demographic Research, demographers at Stockholm University (Eleonora Mussino, Sven Drefahl, Matthew Wallace, Sunnee Billingsley, Siddartha Aradhya and Gunnar Andersson) analysed excess mortality with cause-specific details to find out. They assessed whether different scenarios fit the patterns in the data, and for which groups of people.
They show an underestimation of COVID-19 mortality, especially among people with a migrant background, probably due to underreporting of COVID-19-specific deaths or collateral deaths from pandemic conditions. Even when deaths due to documented COVID-19 infection were excluded, death rates were more than 20% higher for many groups of working-age men and women who were not born in Sweden.
The group with the greatest excess mortality due to COVID-19 (about a 50% increase), and possibly a small proportion of replacement mortality, were men of pensionable age who were married, had a low level of education and were born in a low- or middle-income country.
The authors also found what they called a "sheltering" effect for some groups, with deaths from other causes declining substantially, leading to a reduction in death rates of about 15%, even when COVID-19 mortality was included. This pattern occurred mainly in the working age groups, especially among highly educated women.
Over time during the first year of the pandemic, reflecting the evolution of the pandemic (including better testing but not yet vaccination), they found that more groups joined the "protected" pattern of mortality rates. These findings provide insights into how different groups were affected by the pandemic, both directly through COVID-19 and indirectly through other societal and health care changes.