COVID-19 mortality

1. COVID-19 mortality monitoring

Background

COVID-19 disease is caused by the SARS-CoV-2 virus. The most frequent symptoms are fever, cough, loss of taste and smell and shortness of breath. In 80% of cases the infections are mild. The risk of developing a severe infection increases markedly with age and with underlying conditions such as heart, lung or kidney disease, diabetes, immunosuppression or an active malignant disease.

The World Health Organization (WHO) states that “a death due to COVID-19 is defined for surveillance purposes as a death resulting from a clinically compatible illness, in a probable or confirmed COVID-19 case, unless there is a clear alternative cause of death that cannot be related to COVID disease (e.g. trauma)” [1], and ECDC later on also adopted these inclusion criteria [2]. Belgium has adopted a broader inclusion strategy for the surveillance of deaths, reporting also deaths in possible cases based on clinical symptoms. The rationale for broadening the case-definition of a COVID-19 death in Belgium was the low testing capacity during the first weeks of the epidemic, leading to a quasi-impossibility to get a lab-confirmed diagnosis for patients in LTCFs. As a result, monitoring only deaths in confirmed cases would have hidden the severity of the epidemic. Deaths occurring in hospitals, long-term care facilities (LTCFs), and in the community were reported.

In pandemic times, an almost real-time monitoring of mortality is essential. In Belgium, the cause-specific mortality registration through death certificates is a two-year process, not suitable for operational surveillance. For this reason, an ad-hoc registration in nearly real time of COVID-19 mortality was set up in emergency. Deaths in hospitals are registered through the Hospital Surge Capacity survey managed by Sciensano. Deaths outside the hospital (in nursing homes, other institutions, at home, and other places) are notified to the regional authorities via specific tools. Sciensano is in charge of compiling the information from the different data sources. COVID-19 deaths can only be fully reported by place of deaths for data availability reason (place of residence was not available for hospital deaths until the 24/04). We used place of death when place of residence was unavailable.

As the situation is continuously evolving, this page first provides links towards the continuous monitoring of the COVID-19 mortality on the Sciensano website. Afterwards, the situation during the year 2020 (10th of March – 31th of December 2020) of the COVID-19 epidemic in Belgium is presented. We use the number of COVID-19 deaths and the crude COVID-19 cumulated mortality rate per million of inhabitants from the surveillance of Sciensano.

There are important limitations in the international comparability of COVID-19 mortality rates. Because of the differences in death notification methods between countries [3,4], COVID-19 mortality comparisons are not well suited for international comparisons. Indeed, in Belgium, deaths in patients with a positive laboratory test, a suggestive chest CT scan or clinical symptoms are notified as COVID-19 death, while many other countries are much more restrictive. Here we present the crude COVID-19 cumulated mortality rate per million of inhabitants in selected countries for international comparisons (source: report of ULB-VUB [5], from ECDC original data [6]). Because of the mentioned limitations, we shall complete figures of COVID-19 mortality rates with other indicators (see the section “excess mortality”).

Continuous monitoring of the COVID-19 mortality

Since the situation is in constant evolution, values are updated everyday on the Sciensano website. More specifically, the following resources are available:

  1. Dynamic dashboard Epistat
  2. Reports around mortality:
    1. Daily reports, Mortality chapter, present the main indicators by region and their evolution
    2. Weekly reports, Mortality chapter present indicators by region and province for COVID-19 mortality and indicators around all-cause mortality monitoring.
    3. Mortality in nursing homes is presented in the weekly report about monitoring of COVID-19 in nursing homes.
  3. Open data are available. Data for COVID-19 mortality contain the number of deaths aggregated by age groups, by sex, by date of death, and by region.
  4. Frequently asked questions are answered in a specific document.

COVID-19 mortality, overview 2020

Belgium

The first COVID-19 death in Belgium occurred on March 10th 2020. On the 31th of December 2020, 19 588 deaths were notified in Belgium, representing a crude cumulative death rate of 1704 per million inhabitants.

In 2020, the epidemic presented 2 waves; the first wave was observed from the 10th of March to the 21st of June and the number of deaths peaked on April 8 with 322 COVID-19 deaths. After an interwave period, the second wave started on the 31st of August peaked on the 10th of November with 217 COVID-19 deaths, and was still ongoing by end of 2020.

Number of deaths due to COVID-19 by date of death and by region* in 2020, Belgium
Source: COVID-19 mortality database, Sciensano
* Death rates are expressed by region of residence if known (the region of death is used as proxy if unknown)

More than half of the deceased people were aged over 85 years old.

  • Number of deaths
  • Mortality rates

Number of COVID-19 deaths by age group and gender in 2020, Belgium
Source: COVID-19 mortality database, Sciensano

Age-specific COVID-19 mortality rates (per million inhabitants) by gender in 2020, Belgium
Source: COVID-19 mortality database, Sciensano

Regional specificities

There are moderate regional differences in COVID-19 mortality rates. During the first wave, the cumulative rates expressed per million inhabitants were respectively of 1.171 in Brussels, 927 in Wallonia and 719 in Flanders. The higher mortality rate in Brussels can probably be somehow explained by a higher concentration of hospitals (as the region of residence was often unknown and replaced by the region of death during the first wave). The Brussels region is also exclusively made up of a large city, and the high density of the population is known to enable the circulation of the virus.

During the second wave in 2020, the cumulative rates expressed per million inhabitants were respectively of 772 in Brussels, 1.107 in Wallonia and 714 in Flanders. Wallonia was severely hit in the second wave.

International comparisons in the first wave (selected countries)

Belgium has attracted attention internationally due to a high COVID-19 related mortality during the first wave. Belgium had indeed the highest cumulated death rate in Europe during the first wave. However, due to the methodological limitations mentioned, those figures have to be interpreted with caution. Other indicators are best suited for cross-country comparison (see the section “excess mortality”). A comparison of COVID-19 mortality and excess mortality was made in a ULB-VUB report on a selection of countries close to Belgium. Currently, this comparison has only been published for the first wave.

Cumulative mortality rate by COVID-19 (deaths per million), 15/02 to 29/06, selected countries
Source: ULB-VUB report on mortality [5], mortality data extracted from ECDC [6]

Epidemiological International data can be consulted on several platforms:

2. Excess mortality

Background

In the context of measuring the impact of the COVID-19 on the mortality, it has been recommended [5,7] to use the excess all-cause mortality to:

  • Measure the global impact of the COVID-19 on the mortality burden
  • Allow international comparisons that are less biased than the COVID-19 mortality
  • Assess the reporting of the COVID-19 mortality resulting from the ad-hoc surveillance system

The number of deaths in Belgium is registered by the Belgian National Registry nearly in real-time (95% of deaths in 3 weeks). However, this information does not contain the cause of death, which will become available after 2 years. The information is transmitted to Sciensano via the Be-MOMO project.

The Be-MOMO project (Belgian mortality monitoring) is a surveillance of all-cause mortality carried out by Sciensano on a weekly basis. The mortality monitoring model is a tool for rapid detection and quantification of unusual mortality (from disease epidemics such as influenza or from extreme environmental conditions such as heat/cold waves or environmental pollution) to guide public health measures, e.g. vaccinations for influenza and the national heat action plan [8].

The model predicts the daily expected number of deaths along with a prediction interval, by modeling the past 5-year mortality data and past epidemic seasons and extreme environmental events (cold and heat waves, and air pollution). When the number of observed deaths exceeds the upper prediction limit there is a significant mortality excess for this day, allowing to visualize correlations between mortality deviance and those events [8,9]. The weekly updated figures of Be-MOMO can be found on Epistat.

In 2020, both Be-MOMO and Statbel calculate the excess mortality, with slightly different methods but leading to very similar results [5,10]. Also, contrarily to Statbel, Be-MOMO does not include deaths of residents having occurred abroad.

In this report, we will use the following indicators and data sources:

For Belgium:

  • Excess of all-cause mortality for the first and second wave, from Be-MOMO expressed as a “P-score
  • Comparison between the excess mortality and the COVID-19 deaths.

For international comparisons:

Currently, those comparisons have only be published for the 1st wave. We use selected international and regional comparisons from the mortality report of the ULB-VUB [5]. The researchers have selected some EU countries based on the relevance of the comparison with Belgium. The original data came from ECDC [6] and ONS-UK [10].

Excess all-cause mortality

Be-MOMO estimates 17 955 excess deaths in 2020. This can be explained by 19 588 COVID-19 deaths in 2020 and excess mortality during the Augustus’ heat wave with 1 503 excess deaths [11]. The difference between excess deaths on the one hand and mortality due to COVID-19 and the heat wave on the other is explained by the fact that a proportion of the COVID-19 deaths occurred in people who would have died within the year anyway, and a small proportion of the deaths were probably prevented by lockdown measures. This analysis can only be done when information on the causes of death is available.

An excess mortality was observed in Be-MOMO during the first wave of the COVID-19 epidemic from week 12 of 2020 (starting on 16/3/2020) to week 19 (ending on 10/05/2020) and during the second wave from week 43 (starting on 19/10/2020) to week 52 (ending on 27/12/2020). This is a good indicator of the global impact of the COVID-19 on the mortality.

  • First wave
  • Second wave

Number of all-cause deaths by week, first wave, Belgium
Source: Be-MOMO [8], Sciensano
ExcessMort_1wave_EN.PNG

Number of all-cause deaths by week, second wave, Belgium
Source: Be-MOMO [8], Sciensano
ExcessMort 2wave EN

Comparison between excess mortality and COVID-19 mortality

During the first wave, an excellent correlation (91%) was found between the all-cause mortality and the COVID-19 mortality. This validated the inclusion criteria for the COVID-19 related deaths [7].

Expected number of deaths and confidence intervals, observed number of deaths, and COVID-19 deaths in 2020, Belgium
Source: Be-MOMO [8] and COVID-19 mortality database, Sciensano

 

Graph created by Mathias Leroy
For best performance, please use Firefox, Chrome or Microsoft Edge browsers

International comparisons in the first wave (selected countries)

Due to the methodological issues mentioned before, the excess all-cause mortality during the first wave of the COVID-19 is a more suited indicator than the COVID-19 specific mortality for international comparison of the impact of the COVID-19. Using this indicator, the UK, Spain, and Italy had the highest cumulative excess mortality rates in the selection of countries. Currently, this comparison has not been done yet for the second wave.

Cumulative excess mortality rate (deaths per million), during the respective period of excess mortality of the countries, selected countries
Source: ULB-VUB report on mortality [5], mortality data extracted from ECDC [6]
More information

Definition

P-score
The p-score is a measure of excess mortality. It is calculated as the number of excess deaths (observed deaths – expected deaths) divided by the number of expected deaths *100.

References

  1. WHO. International guidelines for certification and classification (coding) of COVID-19 as cause of death. WHO; 2020 Apr. Report No.: WHO/HQ/DDI/DNA/CAT.
  2. ECDC. Case definition for coronavirus disease 2019 (COVID-19), as of 5 May 2020. European Centre for Disease Prevention and Control. https://www.ecdc.europa.eu/en/covid-19/surveillance/case-definition
  3. European Centre for Disease Prevention and Control (ECDC). Weekly surveillance report on COVID-19, Week 24, 2020. https://covid19-surveillance-report.ecdc.europa.eu/#4_severity
  4. Institut national d’études démographiques (Ined). The demography of COVID-19 deaths - Seven data-related key issues. Ined - Institut national d’études démographiques. https://dc-covid.site.ined.fr/en/presentation/
  5. Lagasse R, Deboosere P. Évaluation épidémiologique de l’impact du Covid-19 en Belgique à la date du 15 juillet 2020 - Page 48. Brussels; 2020. https://esp.ulb.be/fr/les-actus/l-esp-dans-les-medias/rapport-d-analyse-de-l-epidemie-covid-19-n-ii
  6. ECDC. Daily number of new reported cases of COVID-19 by country worldwide. European Centre for Disease Prevention and Control. https://www.ecdc.europa.eu/en/publications-data/download-todays-data-geographic-distribution-covid-19-cases-worldwide
  7. Bustos Sierra N, Bossuyt N, Braeye T, Leroy M, Moyersoen I, Peeters I, et al. All-cause mortality supports the COVID-19 mortality Belgium: comparison with major fatal events of the last century. Submitted.
  8. Leroy M, Dupont Y, Braeye T, Bossuyt N, Bustos Sierra N. Epistat – Belgian Mortality Monitoring (Be-MOMO). https://epistat.wiv-isp.be/momo/
  9. Cox B, Wuillaume F, Van Oyen H, Maes S. Monitoring of all-cause mortality in Belgium (Be-MOMO): a new and automated system for the early detection and quantification of the mortality impact of public health events. Int J Public Health. 2010 Aug 1;55(4):251–9.
  10. Campbell A, Morgan E. Comparisons of all-cause mortality between European countries and regions.
  11. Sciensano. Analyse de la surmortalité liée au COVID-19 en 2020. https://www.sciensano.be/fr/coin-presse/analyse-de-la-surmortalite-liee-au-covid-19-en-2020