Calculating how many people have died because of COVID-19 is a lot more complex than one might think.

Using a measurement like "total mortality" will not likely capture a number of deaths that may be inadvertently caused by COVID-19 or counted that way.

Here is an example: the way the health systems in Russia and Spain measure virus-specific deaths are much different from each other. Russia requires a very resource-intensive autopsy to confirm a COVID-19 death. That means home deaths, deaths "likely" linked to COVID-19 like blood clots and strokes, or deaths due to conditions exacerbated by COVID-19 like heart disease or obesity, may not make the count. On the other hand, Spain's health system requires a positive polymerase chain reaction (PCR) or antibody test. Spain also only includes deaths in hospital settings, which excludes deaths from nursing homes, people without active infections at their time of death, and more.

Now let's compare Russia and Spain with Belgium's much broader definition of a COVID-19 death. In Belgium, COVID-19 death includes all suspected cases, but does not require a positive test, autopsy, or hospitalization for confirmation. Through these comparisons, it's evident how some countries would report more deaths than others. If a person dies in a snowstorm due to hypothermia but had a positive COVID-19 test in recent days, their deaths could technically be counted as a COVID-19 death in some countries but not others.

Because of these differences, most public health measurements use excess mortality as a way to more accurately determine how many people died as a result of the pandemic.

Excess mortality is the difference in the total number of deaths due to some circumstance (like COVID-19), compared to those expected during normal conditions. Another way of phrasing this is the total number of deaths from all causes during a given time period above and beyond what would normally occur in that same time frame had a crisis not occurred.

We are using COVID-19 here as the circumstance, but other circumstances can include events like civil wars, famine, natural disasters, government shutdowns, other viral outbreaks, and more.

Excess mortality is useful as it includes deaths that are directly caused by the virus as well as indirect ones like health system disruptions or a lack of health care.

This specific measurement operates under the assumption that the incidence of other diseases and deaths remain steady over time, but accounts for downward and upward trends in recent years. If we take the example of deaths due to car accidents and roadside injuries which normally increase 1.2% every year in one particular area, we would still use that historical average and 1.2% annual increase in the calculation. We use the expected number of deaths for the same region and period of time using these historical averages, but also include the total deaths that would not have occurred had it not been for COVID-19.

There are many advantages to using excess mortality as a gauge for COVID-19 deaths. Excess mortality is a fairer, more cohesive tool to make international comparisons for these reasons:

• Differences in how countries define and report COVID-19 deaths vary greatly and this causes some countries to report higher or lower numbers
• The ability of health care systems is an important calculation to consider and is factored into excess mortality estimates. Some health systems did not have the ability to treat excess patients due to COVID-19 and ran out of hospital beds, even encouraging patients not to seek care unless they were in life-or-death predicaments. People with chest pain might have assumed they were having indigestion and did not seek care at a hospital but passed away later that day due to a heart attack. People who died from preventable causes but were not able to seek care are a great example of why health care systems are so important for excess mortality measurements. A lack of staff, resources, equipment, and funding can also make autopsies nearly impossible to perform.
• Excess death counts help with miscalculations caused by international differences in population demographics. One example is comparing a country with a younger population in the global South with an older one in Europe. A measurement like total mortality due to COVID-19 does not take into account the fact that an 95-year old Finnish woman who died due to a COVID-19 infection might have already been severely ill in the hospital with terminal cancer. Excess mortality measurements incorporate those who would have likely passed away in the same time frame as part of its calculation.
• Excess mortality includes deaths from other causes that are attributable to COVID-19 but might not be counted as such in certain countries. Recent cases of these deaths can be seen in people with "long-COVID;" those who are not actively infected but have had lingering symptoms since their initial infection often months earlier. When a woman with long COVID suffered chronic kidney damage because of the virus, her death due to kidney failure should likely count as a COVID-19 death but not be tallied this way in many countries.
• Few countries have a strong enough infrastructure system to record, report, and register every death. The United Nations noted that during "normal" periods of time, only two-thirds of countries register at least 90% of all deaths that occur, with some countries only registering between 10-50% of deaths. This number might be even lower during the pandemic due to mass casualties during short time frames.
• Reporting delays also occur in many countries, sometimes months and years after a death occurs, so these deaths are often not reliably counted in total mortality numbers. Deaths are often marked according to when they are registered but not when they occurred, so national differences in reporting requirements might skew numbers lower than they should be.
• Due to COVID-19 restrictions and lockdowns in many countries, fewer deaths from other causes may be incorporated into calculations. Since there were many fewer cases of deaths due to influenza as many people were forced to stay at home and wore masks in public, excess mortality builds these reduced numbers of deaths into its formula.
• Deaths due to chronic diseases like high blood pressure, diabetes, and Alzheimer's have increased dramatically since 2020. Whether these deaths were caused directly by COVID-19 infections, unfilled prescriptions, lack of care seeking, canceled appointments, increases in stress and unhealthy eating and exercise behaviors or more, excess mortality calculations factor them into their ouput.
• Testing capacities are different in every country. Obtaining a PCR test in one nation might occur in the same day whereas finding an at-home antigen test is nearly impossible across a border. People who died as a result of a COVID-19 infection but were never tested should be counted as an excess death but may never be in total mortality calculation.
• Data gaps and weak health information systems make accurate calculations difficult using total mortality rates, but are factored into excess mortality measurements.
• Several other considerations support the use of excess mortality as a measurement but vary by country, region, and scale.

Identifying deaths caused by COVID-19 is difficult and not all COVID-19 deaths might be considered excess deaths. Since fewer people died from the flu, there may even be more confirmed COVID-19 deaths than excess deaths that would have occurred from the flu during normal years.

COVID-19 deaths confirmed through testing or autopsies most likely undercount the total impact of the pandemic on a country's mortality but these numbers are constantly changing as the pandemic undergoes several waves. Only excess mortality is able to account for these variations and as such, is a much more reliable indicator for country-to-country comparisons.

There are several limitations to using excess mortality as a way to determine how many people have died because of COVID-19. However, by comparison, total mortality rates are much less accurate and much more speculative. Excess mortality, at least when nations have similar abilities in terms of reporting death, applies a holistic context on the true impact of COVID-19.

We do not have this mortality data for every country which weakens the accuracy of any calculations. In order to monitor the pandemic's impact, many contextual considerations such at this should be included.