There are various ways to gain insight into how deadly COVID-19 exactly is. One way is through the infection fatality rate or IFR: the proportion of infected people who die. Data from different countries indicate that it is around 0.6% for SARS-CoV-2, meaning 6 deaths per 1,000 infected individuals. Another often used parameter is the case fatality ratio or CFR. It expresses the number of deaths among all people with COVID-19 symptoms. Unlike the infection fatality rate, the case fatality rate does not take into account asymptomatic infections. Both figures vary significantly based on age, which is why a general figure, considering all age groups, actually tells us very little about the impact of a disease.

It is also useful to compare the overall mortality rates. This way, we can get an idea of any excess mortality: how many more people die than the average number of deaths in the same period in other years? Depending on how data is collected, there are several reasons to believe that excess mortality provides a more reliable view of the actual COVID-19 death rates. For example, several studies showed that in Belgium, excess mortality closely corresponds to the COVID-19 death rates because the country counts not only confirmed COVID-19 cases in the hospital but also suspected COVID-19-related deaths that occur anywhere. A comparison with the Netherlands, for instance, reveals that COVID-19 deaths account for only 62% of excess mortality.

Belgium: the country with most COVID-19 deaths?

With the exception of a few microstates such as Monaco or Andorra, Belgium, at one point, had the most COVID-19 deaths in the global statistics. The Our World in Data dashboard indicates that on June 28, 2020, Belgium had 830 COVID-19 deaths per million inhabitants, more than the UK with 593, Italy with 574, Sweden with 539, France with 456, the USA with 379, and Germany with only 107 deaths per million inhabitants. The numbers appear objective as they are adjusted for the size of the population, but many other factors come into play, such as population density and the completeness of COVID-19 death figures themselves.

During the first half of 2020, the coronavirus pandemic was primarily limited to the East Coast of the USA, and the death rate, compared to the entire U.S. population, remained relatively low. However, if you specifically look at New York, the number of deaths is much higher: 1,599 deaths per million inhabitants.

The criteria for what counts as a COVID-19 death also varied from country to country. Belgium is one of the few countries with extensive criteria, including lab and radiology-confirmed COVID-19 deaths in hospitals, nursing homes, or other institutions, as well as 'possible' COVID-19 cases. Much depends on the testing strategy employed.

For all these reasons, excess mortality remains the best way to compare apples with apples.

To gain a more detailed understanding of COVID-19 mortality in Belgium and why it appears to be exceptionally high, we delved deeper into the figures from the first wave. Between March 9 and June 28, 2020, 9,621 people in our country died with or from COVID-19—4,535 men and 5,086 women. Nearly two-thirds of all those deaths occurred among nursing home residents. In people under 30, there were fewer than five deaths. In 27% of all cases (precisely 2,591 deaths), the diagnosis was 'suspected' COVID-19.

Confirmed cases of COVID-19 include a positive COVID-19 test or a confirmed diagnosis via a CT scan of the lungs. Under the category 'suspected' COVID-19 are all cases pointing to COVID-19 based on clinical criteria. The vast majority (89%) of these 'suspected' COVID-19 deaths were deaths in nursing homes.

This exceptionally high toll in nursing homes skews the numbers. Although there were many women among the deaths, in this group, there are twice as many women as men among those 80 or older. Additionally, if we calculate the number of deaths per million inhabitants without including nursing home residents, Belgium's figure drops from 837 to 438 deaths per million inhabitants. When considering only confirmed COVID-19 deaths (including those in nursing homes), Belgium records 755 deaths per million inhabitants.

The excess mortality during the study period is almost 9,000 deaths. Nearly the same as the number of COVID-19 deaths, which is an indication (not proof) that the way COVID-19 deaths are counted in our country is reliable.

Based on individual-based modeling of the number of COVID-19 infections in our country and the mortality figures, we arrived at an infection fatality ratio of 1.5% in the general population. Under the age of 40, the IFR remained almost 0 and then rose to 10% among those aged 90 and older. When further separated between non-nursing homes and nursing homes, the IFR is 0.6% versus a staggering 21%!

The higher vulnerability of the elderly can be partly explained by the presence of more underlying conditions such as high blood pressure or diabetes, two significant risk factors for a much more severe course of COVID-19. What stands out from the data in our country is that there is a greater difference between those living in a nursing home or not in the age category between 60 and 80 years old than among those aged 80 or older. It appears that the high death rates in our country are primarily due to the extremely severe situation in our nursing homes.

In addition to differences in testing strategy and reporting, there are other reasons why mortality rates may differ by country and region. An important factor is the population pyramid: since COVID-19 is much deadlier for older people, the proportion of older versus younger individuals in the population plays a major role in the number of deaths per million inhabitants. There are no extreme differences in this regard between European or other Western countries, but when compared to African or Asian countries, this certainly plays a role.

Furthermore, higher population density and more international travel contribute to more virus spread and, therefore, more deaths, in addition to the measures taken and the capacity of the healthcare system.

For all these reasons, the most reliable way to make international comparisons is based on excess mortality. Excess mortality encompasses both direct and indirect deaths resulting from the pandemic, such as deaths due to an overwhelmed healthcare system or as a result of lockdown measures. But to correctly assess this excess mortality, we must also determine what 'normal' mortality is, without a pandemic. There are various methods to do this, but each has its limitations.

We used a statistical model that goes beyond the limitations of simply taking the average over five years but corrects for distortions caused by correlations and outliers, without the need for additional data. We applied it to Belgium and the Netherlands and found that this model could better predict excess mortality in both 2014 and 2016 compared to just using the weekly average of the previous five years. This gave us confidence to proceed in the same manner for 2020.

For Belgium, we calculated excess mortality to be between 20,467 and 21,000 deaths. If we subtract the 1,460 additional deaths due to the summer heatwave in 2020, we arrive at 19,007 to 19,548 deaths as a direct or indirect result of the pandemic. The official figure for COVID-19-related deaths in our country in 2020 stands at 19,288, aligning perfectly with excess mortality.

In the Netherlands, this situation is quite different. In exactly the same way, we estimate excess mortality for 2020 and arrive at 20,585 to 22,796 additional deaths—much more than the officially reported 11,527 COVID-19 deaths. According to our calculations, that official figure underestimates by about 50%. Later, the relevant authorities in The Neterhalnds would indeed publish a corrected figure of slightly more than 20,000 COVID-19 deaths.

Although the first wave was much more intense in both Belgium and the Netherlands, the second wave lasted longer, resulting in approximately the same number of deaths in the spring and fall. Given that the timing of excess mortality coincides with the actual deaths, it suggests that excess mortality is mainly attributable to the direct consequences of the virus and not the indirect consequences of the measures.

Worse than common flu?

While there are some parallels between COVID-19 and seasonal flu, it's not straightforward to compare the mortality of both. The problems already start with the availability of data. As mentioned earlier, the definition of a COVID-19 death is not the same everywhere. The figures for the flu contain even more uncertainties because they are largely estimates.

In the case of the flu, we are dealing with an endemic virus with different subtypes circulating and against which a portion of the population has built immunity through prior infection or annual flu vaccination. In the case of the coronavirus, it's a completely new virus for which initially, nobody had immunity.

In a sense, it would be more logical to compare a pandemic with a pandemic, such as the Spanish flu pandemic from a century ago. The case fatality ratio for that pandemic is currently estimated at 2%, which is similar to COVID-19. But back then, there were no lab tests or self-tests, so the infection fatality rate remains uncertain. A more recent flu pandemic is the H1N1 or swine flu in 2009, which caused almost 500,000 confirmed cases, including over 18,000 deaths. This gives us a case fatality ratio of 3.75%, relatively high compared to COVID-19. But there are also questions about these numbers because we know that many infections were never detected.

In other words, assessing the lethality of COVID-19 in comparison to the flu is comparing apples to oranges, and it is also of limited value. Both infectious diseases can have far-reaching consequences and overwhelm healthcare, temporarily or long-term, locally or globally.

Did vaccines help?

We've already discussed the effect of the vaccination campaign on viral transmission and hospitalizations. Based on data until October 31, 2021, Sciensano reported that the chance of dying from COVID-19 was more than 85% lower for those who were fully vaccinated. During the second wave in the fall of 2020, for every 1,000 COVID-19 cases, there were still 30 deaths to mourn. This number decreased to only 2 deaths per 1,000 cases from July 2021.

Scientists from London calculated, based on official death statistics from 185 different countries, that more than 14 million lives were saved between December 8, 2020, and December 8, 2021, thanks to the deployment of various vaccines. When they relied on excess mortality rather than official COVID-19 death rates, that estimate rose to almost 20 million avoided deaths. Our colleagues arrived at an estimate for the European Economic Area (the European Union plus Norway, Iceland, and Liechtenstein) of around 220,000 averted deaths. Detailed conclusions about the effect of the vaccination campaign—by region, age group, virus variant—are actively being researched for Belgium. Still, the overall effect is clear: excess mortality dropped to almost zero again in 2021.

And what about vaccine side effects? In 2021 and 2022, approximately 30 million doses were administered in our country. In just over 17,000 cases (about 0.06%), side effects were recorded with temporary health concerns as a result, for example, due to fever or muscle pain. In 297 deaths, a link to recent vaccination was investigated, and in four of those cases, doctors did indeed establish a likely causal link. Furthermore, the relevant authorities continue long-term monitoring to identify any potential rare vaccine side effects; just as they do for all other new medications that are rolled out.

What about long COVID?

In the section about our compartmental models, we discussed the four letters S E I R: susceptible, exposed, infected, and recovered. We have already mentioned that, in addition to the recovered compartment, there is also a compartment for those who die from COVID-19. We have extensively discussed the death toll in this section. However, it turns out that the course of a coronavirus infection has a third possible outcome: long-term physical and mental consequences after the acute phase of the infection. This syndrome is described as long COVID, but there is still no precise clinical definition. It involves symptoms that arise during a coronavirus infection but persist for more than three months, with no other explanation found.

Now that the most pressing phase of the pandemic is behind us, there is increasing attention and research into these long-term effects, both for those who suffer from them and for an assessment of the consequences for our society. So far, from a mathematical modelling perspective, we have conducted little research on this, but undoubtedly, we will have to integrate new insights into the extent and impact of long-COVID into our calculations of viral transmission and the clinical and economic consequences of the crisis in the future