Did COVID-19 Vaccines Cause Excess Deaths?

No. The terrible science behind equally awful headlines.

Pictured: deaths. Photo by Janne Simoes on Unsplash

The COVID-19 vaccines have been, almost across the board, quite miraculous. Within a year of the start of the pandemic, against a completely novel virus, researchers developed immunizations that prevented the vast majority of deaths from the disease and even had some remarkable benefits against infection and transmission. At the beginning of the pandemic, we set the benchmark for a useful vaccine at a 50% reduction in diagnosed disease - the vaccines we ended up with surpassed this threshold by leaps and bounds.

Which makes it strange that recently there have been a slew of headlines proclaiming that the COVID-19 vaccines might have caused excess deaths across the world. According to a wide range of reporting, new research has emerged showing that the COVID-19 vaccines were associated with more death than the disease, in a stark contrast to the many thousands of research papers already published that show precisely the opposite.

Fortunately for us all, the claim that COVID-19 vaccines caused excess mortality is based on extremely shoddy science. In fact, the data shows that the immunizations have saved countless lives.

The Study

The study in question is a paper recently published in the British Medical Journal: Public Health. It’s a very curious piece of writing. The authors of this paper say that they took the Our World In Data COVID-19 dashboard on excess deaths for a group of “Western” countries, and reviewed the data for the years 2020, 2021, and 2022. They found that the rates of excess mortality were highest across most of these countries in 2021 and 2022, and then used that fact to insinuate that COVID-19 didn’t cause deaths but vaccines and non-pharmaceutical interventions might have:

“At a global level, the prevaccination Infection Fatality Rate was 0.03% for people aged <60 years and 0.07% for people aged <70 years. For children aged 0–19 years, the Infection Fatality Rate was set at 0.0003%. This implies that children are rarely harmed by the COVID-19 virus. During 2021, when not only containment measures but also COVID-19 vaccines were used to tackle virus spread and infection, the highest number of excess deaths was recorded: 1 256 942 excess deaths (P-score 13.8%)”

“The next step concerns distinguishing between the various potential contributors to excess mortality, including COVID-19 infection, indirect effects of containment measures and COVID-19 vaccination programmes.”

They also argue that the fact that there were more excess deaths in 2020 than 2022 in some places indicates that COVID-19 restrictions were ineffective or harmful, because most places had restrictions implemented in 2020 but not 2022.

There are so many errors in this short paper that it’s hard to know where to start. The first mistake is fairly simple - this isn’t new work in any meaningful sense. It’s also not taken from Our World In Data, which is itself a data aggregator and visualization website which doesn’t produce new information. The actual data in the paper was taken with some attribution from a previously published piece of work called the World Mortality Dataset from researchers Ariel Karlinsky and Dmitry Kobak. The new paper in the BMJ simply lifted the methods and part of this dataset, added their own opinions on top, and published it as “original research”. At best, this new paper is a minimal and poorly-researched commentary on Karlinsky and Kobak’s work.

Then there’s the idea of “Western” countries. This is a bizarre distinction to see this side of the Cold War. I haven’t seen the term used in a paper published this century. The authors say they define “Western” as:

“various countries in Europe and to countries in Australasia (Australia, New Zealand) and North America (the USA, Canada) that are based on European cultural heritage. The latter countries were once British colonies that acquired Christianity and the Latin alphabet and whose populations comprised numerous descendants from European colonists or migrants”

This is then cited to worldpopulationreview dot com, which seems to be an anonymous data aggregator website that mostly exists to sell ads and affiliate links. About as far from a scientific source as you can imagine, and certainly not what you’d expect to see in a paper from a leading public health journal.

You have to wonder at the scientists who saw this website and said “Ah yes, an excellent reference on which to build our entire study”

The story gets much worse. The authors of this paper argued that COVID-19 probably didn’t cause excess mortality, and then implied that vaccines and mitigation measures were likely responsible. However, had they actually bothered to read the World Mortality Dataset and associated analyses, they would’ve been confronted with direct evidence that this was wrong. Indeed, as you can see from the data, there is a remarkably strong (almost perfect) correlation between excess deaths and COVID-19 deaths across most of the world:

Source: WMD github.

It is quite rare to see such close correlations between unrelated data. While this does not prove that COVID-19 was responsible for all excess mortality since 2020, it does show that most of these deaths were almost certainly related to the virus. We also have extensive, detailed investigations from national authorities that tie the virus directly to these deaths. In other words, we know that most excess mortality during the pandemic was caused by COVID-19.

Conversely, the exact same data shows that there is almost no correlation between vaccination campaigns and excess mortality except where vaccinations were done during a COVID-19 wave. Correlation doesn’t equal causation, but in this case there isn’t even a correlation - there’s simply no reason to even suggest that vaccines or COVID-19 interventions were responsible for excess mortality since 2020.

This is even identifiable from the data that the authors plotted themselves - here is their graph of excess deaths for various countries. Look at Australia - the dark blue line. I’ve added lines for when the main vaccination campaign in Australia started and reached the majority of the population:

Source: BMJ Public Health

As you can see, there were very few excess deaths in Australia during the vaccination campaign rollout. This is because there weren’t many cases of COVID-19. When the major Aussie Omicron outbreak started, in late 2021, the excess deaths increased substantially. This data is exactly the opposite of what you’d expect to see if vaccines were really related to excess deaths. In a similar vein, during the peak restriction periods in Australia - 2020 and mid 2021 - there were no excess deaths (in fact, there were negative excess deaths) - meaning that restrictions could not have been the primary cause of excess mortality in the country.

Somehow, it gets even worse. The article was just published now, in 2024, but it only contains data up until the end of 2022. There’s no explanation given as to why the authors excluded the last 18 months of information. As Stuart Mcdonald pointed out in an excellent thread on Twitter, 2023 showed dramatically lower excess deaths than 2022, despite relatively similar vaccination rates (but much lower COVID-19 burden). In the same thread, he pointed out examples of direct copying from Karlinsky and Kobak in the BMJ Public Health paper. The authors also rely in many parts of their study on very questionable references, some of which could be reasonably described as pseudoscientific. The list of issues with this paper is almost endless.

Vaccines Save Lives

The bottom line here is quite simple: COVID-19 vaccines have saved countless lives. This is not a matter of belief - it is a readily demonstrable fact with an enormous amount of evidence behind it. We can put fairly specific estimates on the number of deaths caused by vaccines, because they are sufficiently rare that we know of almost all of them.

The recent BMJ Public Health paper that has caused all of this noise is quite simply awful. It is a remarkably poor work of science which makes numerous basic errors, and it’s quite hard to understand how such a problematic study got published in the first place. Aside from everything else, it is astonishing to see a study that reportedly copy+pasted large segments of the methods section from another published paper in a BMJ journal.

On the journal’s part, they have since released a public statement that argues that the paper did not show that vaccines have caused excess deaths. This is true, which begs the question of how in the world the editorial team allowed the authors to make the problematic claims in their published paper. In addition, the statement reads:

“The message of the research is that understanding overall excess mortality since the COVID-19 pandemic is crucial for future health policy, but that identifying specific causes is complex due to varying national data quality and reporting methods.”

This is false. In fact, the World Mortality Dataset, which was where all this data came from in the first place, spent quite a lot of time showing that excess mortality during the pandemic - or at least between 2020-20221 - was most likely caused by the virus itself. One of the interesting findings of Karlinsky and Kobak’s work is that COVID-19 restrictions were associated with negative short-term excess mortality. While there are many complexities to the implementation of non-pharmaceutical interventions, particularly from an economic standpoint, one of the things we have very good evidence for is that they have almost certainly not caused excess deaths.

Bad science aside, the evidence is clear - COVID-19 vaccines saved lives. Excess deaths during the pandemic were almost all caused by the virus itself. While there were a very small number of deaths caused by the vaccines, the figure is dwarfed by the many millions of lives that our incredible immunizations saved.