Excess mortality during the Coronavirus pandemic (COVID-19)
Excess mortality during COVID-19
What is “excess mortality”?
Excess mortality is a statistical term that refers to the additional number of deaths, from all causes, during a crisis, above the level we expect to see in “normal” conditions.1
In this case, we’re interested in comparing the number of deaths during the COVID-19 pandemic to the number we would have expected had the pandemic not occurred. This is a very important statistic that cannot be known but can be estimated in several ways.
Excess mortality is a more comprehensive measure of the total mortality impact of the pandemic, compared to the number of confirmed COVID-19 deaths. This is because it captures not only confirmed deaths, but also COVID-19 deaths that were not accurately diagnosed and reported2 as well as deaths from other causes that are attributable to disruptions and overall crisis conditions.3
In the section “Excess mortality during COVID-19: background”, we discuss the relationship between confirmed COVID-19 deaths and excess mortality in further detail.
How is excess mortality measured?
Excess mortality is measured as the difference between the number of deaths actually reported during a given period, versus the “baseline” number of deaths expected during that period if the COVID-19 pandemic had not occurred.
Excess deaths = Number of reported deaths – Number of expected deaths
The baseline number of expected deaths can be estimated in several different ways.
We use estimates produced by Ariel Karlinsky and Dmitry Kobak as part of their World Mortality Dataset (WMD).4 To produce this estimate, they first fit a regression model for each region using historical death data from 2015–2019.5 They then use the model to project the number of deaths we might normally have expected in 2020–2024.6 Their model captures both seasonal variation and year-to-year trends in mortality.
For more details on this method, see the study by Karlinsky and Kobak (2021).7
Previously, we had used a different baseline for expected deaths: the average number of deaths over the years 2015–2019.8
We made this change because using a simple five-year average for the baseline has important limitations. A major limitation is that it does not account for year-to-year trends in mortality. For example, increases in the annual number of deaths due to a growing and aging population would lead to an increasing number of baseline deaths over time. Therefore, using the previous five-year average could lead to misestimating excess mortality.
On the other hand, the WMD projection accounts for year-to-year trends and does not suffer from this limitation. Our charts using the five-year average are still accessible in the links in the sections below.
For reported deaths, we have sourced our data from both WMD and the Human Mortality Database.
The P-score: a measure of excess mortality that is more comparable across countries
The raw number of excess deaths gives us a sense of scale, but it is less comparable between countries because they can have large differences in population sizes and annual expected deaths.
To help make better comparisons between countries, we measure excess mortality as the percentage difference between the reported and projected number of deaths.
This metric is called the “P-score”, and we calculate it as:9
P-score = [ (Reported deaths – Projected deaths) / Projected deaths ] x 100
For example, if a country had a P-score of 100% in a given week in 2020, it would mean that the country had a death count for that week that was 100% higher than the projected death count for that week, in other words, twice as high.
Excess mortality P-scores
The chart here shows excess mortality during the pandemic for all ages using the P-score.
Important points about excess mortality figures to keep in mind
The reported number of deaths might not count all deaths that occurred. This is the case for two reasons:
- First, not all countries have the capacity to register and report all deaths. In many low- and middle-income countries, a large share of deaths are not recorded in vital registries. This is often because of a lack of hospitals, healthcare staff, and infrastructure to register deaths. These problems may be worsened during the pandemic.10
- Second, there are delays in death reporting, which means mortality data can be provisional and incomplete in the weeks, months, or even years after a death occurs — even in richer countries with high-quality mortality reporting systems.11 The extent of the delay varies by country. For some, the most recent data points are clearly very incomplete and therefore inaccurate — we do not show these clearly incomplete data points.12
The date associated with a death may refer to when the death occurred or when it was registered. This varies by country. Death counts by date of registration can vary day-by-day even if there is no actual variation in deaths, for example, due to registration delays or the closure of registration offices on weekends and holidays. It can also be the case that deaths are registered, but the date of death is unknown — this is the case for Sweden.13
The dates of any particular reporting week might differ slightly between countries. This is because countries that report weekly data define the start and end days of the week differently. Most follow the international standard ISO 8601, which defines the week as from Monday to Sunday, but not all countries follow this standard.14 In the charts on this page we use the ISO 8601 week end dates from 2020–2024.15
Weekly reported deaths might not be directly comparable to monthly reported deaths. For instance, because excess mortality calculated from monthly data tends to be lower than the excess calculated from weekly data.16
For more discussion and detail on these points, see our article with John Muellbauer and Janine Aron as well as the methods and metadata from the Human Mortality Database and World Mortality Dataset.
Related charts:
Excess mortality P-scores by age group
The chart here shows P-scores broken down by age groups. The mortality risk from COVID-19 and other respiratory diseases increases exponentially with age.
Countries whose data on reported deaths is sourced from the World Mortality Dataset are not included in this chart because the data is not broken down by age. However, WMD does provide the projected baselines used for calculating P-scores by age in this chart.17