Sweden had a different pattern of deaths over the covid era compared with the countries we have looked at.
The nation had its peak death rate in the first year of the pandemic, 2020, and had a lower death rate than expected1 in 2021.
By 2022 the death rate was basically back to normal pre-pandemic levels.
You can see this on the graph below, where we show percentage changes in death rates over the covid era (2020-2022) in the context of annual changes since 1955. The death rates are all from, or easily calculated from, data from Statistics Sweden2 (SCB).
Unlike the other countries we have analysed, the spike during the covid years in Sweden was not the highest seen in modern history. That distinction goes to 1976, when the death rate was 5.2% higher than the average of the five preceding years (1971-1975).
The graph for Sweden looks completely different from those for other countries we have examined on this project, such as Australia, New Zealand, the Philippines and Denmark, which is a very similar country to Sweden (page coming soon – see the graph below).
We will have to wait and see how the 2023 Swedish death data looks, but so far, Sweden’s graph appears consistent with the classic picture of deaths during and after an epidemic.
Normally, in the first year of an epidemic, vulnerable people die if the disease is allowed to pass through the population, uninhibited by measures such as lockdowns or quarantines. Meanwhile, the many who are not vulnerable survive the infection and gain natural immunity. Assuming there were no other significant causes of excess deaths, this appears to be what happened in Sweden in 2020, with the resulting 4.7% spike in the death rate.
But in a typical pandemic, fewer people than normal die the year after the infectious disease arrived. This is because vulnerable people who might have been expected to die in that year have already died a year earlier than expected. This is sometimes called a ‘harvesting effect’ or a ‘pull forward effect’. So, the dip in the death rate in Sweden in 2021 appears to fit a normal pattern.
How did Sweden differ from other countries in their pandemic response?
The Swedish government allowed citizens more choice on how to respond to covid than most other countries and they did not introduce compulsory lockdowns, although there were bans on public gatherings and advice was given to minimize social contact when possible.
Sweden did have a covid vaccine roll-out, but in contrast to countries like Australia and New Zealand, we can surmise that many Swedes had already been exposed to covid before they were vaccinated. By the 30th December 2021, 7,753,150 Swedes had received at least one dose of covid vaccination and 2,468,482 had received three doses.
Sweden did not have lockdowns and had far fewer other compulsory restrictions than most other countries. Sweden’s graph shows some excess deaths in 2020, but none in later years. Therefore, Sweden’s graph proves that excess deaths in other countries that did lock down their citizens were caused by those restrictions – right?
Wrong. This is oversimplistic. While the Swedish graph may be consistent with this explanation, it does not prove it. It does suggest, though, that lockdowns should be investigated as a possible cause of excess deaths in other countries.
The hypothesis that lockdowns contributed to excess deaths in another country must be tested using detailed data for deaths in that country (broken down by week, region and age-group) in conjunction with detailed regional data on the timing of lockdowns.
If the lockdown hypothesis stands up at that level of scrutiny, then we still can’t accept that lockdowns caused all the excess deaths in that country. We would need to test other hypotheses using the detailed data on deaths to show that other factors did not account for some of the excess deaths in that country.
Other factors that would need to be ruled out include: covid itself; the covid vaccines; fear; changes in health provision practices; and economic hardship. You will probably be able to think of more!
In 2021, when the covid vaccines were rolled out, there was an overall deficit of deaths in Sweden. However, this does not prove that the covid vaccines caused no excess deaths. It is possible that there could have been increased deaths from the vaccine rollout, but that these were masked by the harvesting effect following the 2020 spike in the death rate.
To prove that the covid vaccines did not cause extra deaths during the vaccine rollout, deaths in the age groups to whom the vaccines were rolled out need to be tracked using weekly data, by region. An analysis of this type has been done for Queensland in Australia, for instance, by Dr Andrew Madry.
The fact that the Swedish graph shows a deficit of deaths in the years when the covid vaccines were rolled out, has been used to argue that covid vaccines cannot have caused harm anywhere. This argument is not valid for the following reasons:
First, as explained in 2 above, weekly data by region and age-group needs to be examined to see whether there is evidence of harm caused by the vaccines as they were rolled out in Sweden. Similar analyses need to be done for other countries too. Only then will we know for each country whether the covid vaccines have been safe or whether they have caused harm.
If it turns out that there is no evidence of harm from the covid vaccines in Sweden, but there is evidence of harm from them in other countries, then we will need to understand why there is a difference.
This is rather premature, but here are three hypotheses that could be explored, (and no doubt there will be others):
I. Could the batches of covid vaccines used in Sweden have been harmless while some batches used in other countries caused excess deaths?
II. In Sweden we can assume that many people had encountered covid before they were vaccinated. In contrast, countries like Australia and New Zealand managed to keep covid ‘locked out’ with national border closures. In these countries, predominantly, people were vaccinated before they encountered covid. Is it possible that the sequence of infection and vaccination made a difference?
III. Could it be that being locked down somehow made people more susceptible to harm from the covid vaccines?
Our Sweden graph raises questions, but it cannot provide the answers. This is because the data is at a very ‘broad-brush’ level. This annual data does not enable the detailed analysis required to properly answer the questions.
To do that, as mentioned above, weekly data on deaths by age group and region will be needed, together with information on timings of vaccine rollouts and covid infections in those regions3.
Understanding exactly how and why Sweden is different from other countries in its pattern of excess deaths may be very important to learn lessons for dealing with future epidemics. And this Swedish data is already providing valuable information that can be applied to other countries – as you will see in the next section.
To answer this question properly, data on all causes of death for each age group in Sweden in 2020 needs to be examined to see whether any causes of death showed a significant increase (or decrease) compared with normal pre-pandemic levels.
Clearly the covid vaccines cannot be a cause of the 2020 excess deaths – they were still under development.
Even at this broad-brush level, the data from Sweden’s year of excess deaths is valuable because it informs the debate on excess deaths worldwide.
Many governments, including Australia, are assuming that every death from or with covid is an excess death. But is this correct? The data from Sweden’s peak year of excess deaths in 2020 suggests that it is not.
Using data from the SCB, we calculate that the 4.7% increase in the death rate equates to 4369 unexpected or extra deaths in 20204.
Wait a minute! There were more covid deaths than the total number of excess deaths? This makes sense if we remember that many people who died from covid were old and frail and would have died in 2020 anyway. If covid had not killed them in 2020, something else would have done6.
This Swedish data suggests that, at most, just 43% of deaths from or with covid are likely to be excess deaths.
The World Health Organisation is working towards an agreement on pandemic prevention, preparedness and response that will increase central control of signatory nations in their approach to future pandemics.
Had this been in force during the covid era, Sweden would likely not have been able to implement its own response. We might never have seen the Sweden graph above, that clearly shows a better performance – measured by per cent excess death rate – than Australia, New Zealand and the Philippines (and no doubt many other countries).
With no comparators that do anything different, nothing can be improved and flawed policies will not come to light.
We owe a great deal to Dr Anders Tegnell, Sweden’s State Epidemiologist until March 2022, who put Sweden on a different course from other nations.
Thank you to Swedish volunteer Mattias Johansson for his work in providing the information, without which this page would not exist. Thank you also to other members of the Nordic excess deaths team of volunteers: Jakob Hansen; Rasmus O Firla-Holme; and Tore Gulbrandsen for their contributions.
1 The ‘expected’ death rate is here defined as the average over the five pre-pandemic years (2015-2019)
2 Statistiska Centralbyrån
3 Ideally, we would also have access to the depersonalized data on age, sex, prior health conditions, vaccination status and date, and date of death for each person.
4 The calculation is explained in the notes below and is also in the downloadable spreadsheet.
5 It appears from this Cato Institute report that some of these deaths were deaths with covid rather than deaths from covid, as deaths from any cause within 30 days of a positive covid test were classified as covid deaths in Sweden.
6 This statement is in no way intended to diminish the loss of these individuals to their family and friends.
All the data has been obtained from the Swedish Statistical Office (Statistiska Centralbyrån or SCB).
In general, the blue and black lines show the percentage excess death rate for a year calculated by comparing that year’s death rate with a five-year average.
If the blue or black lines are above zero it means the death rate was higher in that year than the average of the previous five years.
Conversely, when the blue line goes below zero it means the death rate that year was lower than the average over the preceding five years.
For 1955-2019, to calculate excess deaths in any year the death rate in that year is compared with the average of the death rates for the five preceding years (the ‘baseline comparator’). Because we expect death rates to have been abnormal during the covid pandemic, we have not included pandemic years (2020-2022) in our baseline comparators. Hence, we have compared death rates in a pandemic year (2020, 2021 and 2022) with the average of death rates in the five years from 2015-2019 – a period that can be regarded as ‘normal’.
The death rate used in the calculations is deaths per 1000 population. For the blue line portion of the graph, the death rate was calculated from figures for deaths and population provided by the SCB. Population figures are reported for the 31st December in Sweden. To get an estimate of the population on 30th June for a calendar year we took the average of the December 31st population figures of that year and the year before. For example, for 1980, we took the average of the figures for December 1979 and December 1980.
For calculating the black line portion of the graph (from 2000 onwards), figures for death rates were available directly from the SCB.
In 2020 the SCB reported death rate was 9.48 deaths per thousand of the population.
The average death rate over the five pre-pandemic years (2015-2019) was 9.06 deaths per thousand of the population.
Hence the excess in the death rate in 2020 was 9.48-9.06 = 0.42 deaths per thousand.
Using SCB estimates of population on 31st December 2019 and 31st December 2020 of 10,327,589 and 10,379,295 respectively we assume that the population on 30th June was the average of these two numbers, being 10,353,442.
Hence the increase in deaths (excess deaths) is 0.42 * 10,353,442/1000 = 4,369
Public Health Agency of Sweden (Folkälsomyndigheten). Statistik för vaccination mot covid-19 as published on 30th December 2021 (obtained using https://archive.org/web) Accessed 8th December 2023
https://www.statistikdatabasen.scb.se/pxweb/en/ssd/START__BE__BE0101__BE0101G/BefUtvKon1749/ – Accessed 7th December 2023