README.md

Data on excess mortality during the COVID-19 pandemic by Our World in Data

For more general information on our COVID-19 data, see our main README file in /public/data.

Data sources

The all-cause mortality data is from the Human Mortality Database (HMD) Short-term Mortality Fluctuations project and the World Mortality Dataset (WMD). Both sources are updated weekly.

WMD sources some of its data from HMD, but we use the data from HMD directly. We do not use the data from some countries in WMD because they fail to meet the following data quality criteria: 1) at least three years of historical data; and 2) data published either weekly or monthly. The full list of excluded countries and reasons for exclusion can be found in this spreadsheet.

See here for a full list of source information (i.e., HMD or WMD) country by country.

We source our baseline of projected deaths in 2020 from WMD.

We calculate the number of weekly deaths for the United Kingdom by summing the weekly deaths from England & Wales, Scotland, and Northern Ireland.

For a more detailed description of the HMD data, including week date definitions, the coverage (of individuals, locations, and time), whether dates are for death occurrence or registration, the original national source information, and important caveats, see the HMD metadata file.

For a more detailed description of the WMD data, including original source information, see their GitHub page.

Excess mortality data

Stored in excess_mortality.csv.

As of 20 September 2021, the data columns are:

• location: name of the country or region
• date: date on which a reporting week or month ended in 2020 and 2021 only (week dates according to ISO 8601). These dates do not apply to other years, such as "deaths_2019_all_ages"; instead, the deaths data across years is organized according to the week or month number in that year — see the "time" and "time_unit" columns below.
• p_proj_all_ages: P-scores using projected baseline for all ages; see note below for the definition of the P-score, how we calculate it, and changes implemented on 20 September 2021.
• p_proj_0_14: P-scores using projected baseline for ages 0–14
• p_proj_15_64: P-scores using projected baseline for ages 15–64
• p_proj_65_74: P-scores using projected baseline for ages 65–74
• p_proj_75_84: P-scores using projected baseline for ages 75–84
• p_proj_85p: P-scores using projected baseline for ages 85 and above
• p_avg_all_ages: P-scores using 5-year average baseline for all ages
• p_avg_0_14: P-scores using 5-year average baseline for ages 0–14
• p_avg_15_64: P-scores using 5-year average baseline for ages 15–64
• p_avg_65_74: P-scores using 5-year average baseline for ages 65–74
• p_avg_75_84: P-scores using 5-year average baseline for ages 75–84
• p_avg_85p: P-scores using 5-year average baseline for ages 85 and above
• deaths_2021_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2021
• deaths_2020_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2020
• projected_deaths_2020_all_ages: projected number of weekly or monthly deaths from all causes for all ages for 2020
• average_deaths_2015_2019_all_ages: average number of weekly or monthly deaths from all causes for all ages over the years 2015–2019
• deaths_2019_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2019
• deaths_2018_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2018
• deaths_2017_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2017
• deaths_2016_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2016
• deaths_2015_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2015
• deaths_2014_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2014
• deaths_2013_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2013
• deaths_2012_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2012
• deaths_2011_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2011
• deaths_2010_all_ages: reported number of weekly or monthly deaths from all causes for all ages in 2010
• time: week or month number in the year
• time_unit: denotes whether the “time” column values are weekly or monthly
• excess_proj_all_ages: number of excess deaths; calculated as reported deaths minus projected deaths
• cum_excess_proj_all_ages: cumulative number of excess deaths; cumulated starting 1 January 2020
• cum_proj_deaths_all_ages: cumulative number of projected deaths; cumulated starting 1 January 2020
• cum_p_proj_all_ages: cumulative P-scores using projected baseline for all ages

How P-scores are defined and calculated

As of 20 September 2021, we calculate P-scores using the reported deaths data from HMD and WMD and the projected deaths for 2020 from WMD, as an estimate of expected deaths. The P-score is the percentage difference between the reported number of weekly or monthly deaths in 2020–2021 and the projected number of deaths for the same period based on previous years.

Before 20 September 2021, we calculated P-scores using a different estimate of expected deaths: the five-year average from 2015–2019. We made this change because using the five-year average has an important limitation — it does not account for year-to-year trends in mortality and thus can misestimate excess mortality. The WMD projection we now use, on the other hand, does not suffer from this limitation because it accounts for these year-to-year trends.

Important points about the data

For more details see our page on Excess mortality during the Coronavirus pandemic (COVID-19).

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 infrastructure and capacity to register and report all deaths. In richer countries with high-quality mortality reporting systems, nearly 100% of deaths are registered; but in many low- and middle-income countries, undercounting of mortality is a serious issue. The UN estimates that only two-thirds of countries register at least 90% of all deaths that occur, and some countries register less than 50% — or even under 10% — of deaths.
• Second, there are delays in death reporting that make mortality data provisional and incomplete in the weeks, months, and even years after a death occurs — even in richer countries with high-quality mortality reporting systems. 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. (For a detailed list of the data we exclude for each country see this spreadsheet.)

The date associated with a death might refer to when the death occurred or to when it was registered. This varies by country. Death counts by date of registration can vary significantly irrespectively of any actual variation in deaths, such as from registration delays or the closure of registration offices on weekends and holidays. It can also happen that deaths are registered, but the date of death is unknown — those deaths are not included in the weekly or monthly data here.

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 international standard ISO 8601, which defines the week as from Monday to Sunday, but not all countries follow this standard. We use the ISO 8601 week end dates from 2020-2021.

Deaths reported weekly might not be directly comparable to deaths reported monthly. For instance, because excess mortality calculated from monthly data tends to be lower than the excess calculated from weekly data.