Press release No. 373 of 10 August 2021
WIESBADEN – According to extrapolated figures of the Federal Statistical Office (Destatis), 76,547 people died in Germany in July 2021. This is 2%, or 1.847 cases, above the median of the years 2017 to 2020 for that month. Different trends were observed during the month. At the beginning of the month, the current figure was lower than the comparative value for the first time since mid-April (26 Week from 28 June to 4 July: -1%). In the middle of the month, the figures were again higher, that is, by 9% in Week 28 (12 to 18 July). In Week 30 (26 July to 1 August) the mortality figures declined again slightly below the median of the previous years. This is shown by an ad hoc evaluation of provisional death figures. Applying an extrapolation method to incomplete reports allows first death figures for Germany to be released after roughly one week.
| Month | Total 2021 | Difference from median 2017-2020 | Relative difference from ... | ||||
|---|---|---|---|---|---|---|---|
| median 2017-2020 | 2017 | 2018 | 2019 | 2020 | |||
| number | % | ||||||
| As at 9 August 2021 | |||||||
| January | 106,379 | +21,337 | +25 | +11 | +25 | +25 | +25 |
| February | 81,895 | -1,509 | -2 | -10 | -5 | +1 | +2 |
| March | 81,541 | -5,527 | -6 | -2 | -24 | -6 | -7 |
| April | 81,300 | +2,826 | +4 | +11 | +2 | +5 | -3 |
| May | 80,264 | +4,588 | +6 | +6 | +8 | +6 | +6 |
| June | 76,079 | +5,178 | +7 | +9 | +10 | +4 | +5 |
| July | 76,547 | +1,847 | +2 | +7 | +1 | 0 | +4 |
| January to July | 584,005 | +25.213 | +5 | +4 | +1 | +5 | +5 |
Number of COVID-19 deaths is decreasing
A comparison of total deaths with the number of COVID-19 deaths by death date reported to the Robert Koch Institute (RKI) is currently possible up to and including Week 28 of 2021 (12 to 18 July 2021). In that week, there were 46 COVID-19 deaths, according to the RKI. This was a decrease of 4 cases compared with the preceding week.
| Week | Total number 2021 | Difference from median 2017-2020 | COVID-19 Deaths | Relative difference from … | ||||
|---|---|---|---|---|---|---|---|---|
| median 2017-2020 | 2017 | 2018 | 2019 | 2020 | ||||
| number | % | |||||||
| Sources: total death figures: Federal Statistical Office (as at 09 August 2021), COVID-19 deaths: Robert Koch Institute (as at 05 August 2021) | ||||||||
| Week 1-6 | 138,470 | +22,255 | 25,375 | +19 | +4 | +19 | +19 | +21 |
| Week 7-12 | 112,868 | -8,052 | 8,586 | -7 | -7 | -21 | -6 | -4 |
| Week 13-22 | 185,058 | +7,138 | 12,015 | +4 | +8 | +3 | +5 | 0 |
| Week 23 | 17,236 | +248 | 270 | +1 | +9 | +3 | -1 | 0 |
| Week 24 | 18,847 | +2,561 | 216 | +16 | +17 | +21 | +14 | +14 |
| Week 25 | 17,018 | +498 | 122 | +3 | +1 | +7 | +2 | +4 |
| Week 26 | 16,773 | -158 | 83 | -1 | +4 | +1 | -6 | -3 |
| Week 27 | 16,827 | 322 | 50 | +2 | +2 | +1 | +2 | +2 |
| Week 28 | 17,544 | 1,438 | 46 | +9 | +12 | +9 | +8 | +9 |
| Week 29 | 17,673 | 1,027 | . | +6 | +7 | +6 | +5 | +7 |
| Week 30 | 17,406 | -208 | . | -1 | +8 | -5 | -11 | +3 |
| Week 1-30 | 575,720 | +22,888 | . | +4 | +4 | +1 | +4 | +4 |
Bremen, Brandenburg and Berlin recording the largest deviations
At the Länder level, mortality figures can currently be shown up to and including Week 27 (5 to 11 July 2021). In that week, they exceeded the median of the preceding years in nine of the 16 Länder, while they were lower than the median in the remaining seven Länder. The three Länder recording the largest positive differences were Brandenburg (+17% or 96 cases), Bremen (+17% or 23 cases) and Berlin (+15% or 91 cases). The largest negative difference from the median was recorded in Hamburg (-15% or 47 cases).
A graphical overview of the development of mortality figures in all Länder is available here.
High excess mortality in Spain, moderate excess mortality in Belgium, low excess mortality in Italy
The EuroMOMO network for mortality monitoring compares findings about excess mortality across Europe using another approach. Based on its own extrapolation of incomplete data reports and its excess mortality concept, the network also has first figures up to Week 30, which may however change due to late reporting Currently, EuroMOMO does not report excess mortality for Germany in that week. High excess mortality is reported for Spain, moderate excess mortality for Belgium, and low excess mortality for Italy.
New benchmark: median instead of arithmetic mean
As of the reference month of July 2021 a new basis has been used for the comparison of death figures with those of previous years. The so-called arithmetic mean has been replaced by the median. Here the results of the previous four years are divided into two halves; so the median is in the middle between the second highest and third highest value of the previous four years. Neither the lowest nor the highest number of deaths of the preceding four years has an impact on the result of the median. Its advantage compared to the arithmetic mean is that it is less prone to specific one-off developments and outliers. Otherwise the second coronavirus wave would lift the average from October 2020 to a level at which the current values would be compared with an “excessive” mortality caused by extraordinary effects rather than a “normal” mortality. Using the median allows a more meaningful evaluation of the development in the further course of the year. If there is no substantial one-off development in the comparative period, the results of the comparison with the median and the comparison with the arithmetic mean are very similar. For example, the mortality figure for Week 30 is 1% below the median and 2% below the arithmetic mean.
Methodological notes on the mortality figures for Germany:
The 2021 ad hoc evaluation is based on first provisional data (raw data). These are mere counts of the cases of death reported by the registrar's offices; the usual data plausibility and completeness checks have not been carried out. Due to legal regulations concerning the reporting of deaths to the registrar’s offices and differences in the routines of registrar’s offices submitting data for official statistics, these data are still incomplete.
Because of the high relevance of timely mortality figures during the coronavirus pandemic, the Federal Statistical Office has developed an estimation model for the extrapolation of incomplete data. The model allows nationwide mortality figures to be provided already after roughly one week. The mortality figures of the last nine weeks are extrapolated using the figures so far submitted by the registrar’s offices. Therefore the figures can be slightly higher or lower at a later time. The estimate is based on the patterns observed in past reporting delays, some of which differ considerably between regions. Comparable results for the Länder are therefore only available after roughly four weeks.
The relevant notes in the ad hoc evaluation "Sterbefälle – Fallzahlen nach Tagen, Wochen, Monaten, Altersgruppen, Geschlecht und Bundesländern für Deutschland 2016 bis 202" (Deaths – Number of cases by day, week, month, age group, sex and Land for Germany, 2016 to 2021) provide more detailed information on the extrapolation procedure. The tables also permit analyses of your own. The ad hoc evaluation is updated every week on the Deaths, life expectancy theme page. New results are released every Tuesday.
Periods of excess mortality in the course of a year can be identified on the basis of the provisional death figures. This reveals direct and indirect effects of the COVID-19 pandemic on death figures at an early point in time. For that purpose, we compare the figures with the median of several previous years to account for the varying impact of recurring seasonal effects (for instance due to flu or heat waves). The impact of the rising life expectancy and the increasing proportion of older people on the number of deaths to be expected cannot be considered by this comparison.
From March 2020, the figures can only be interpreted in the light of the measures taken to contain the coronavirus pandemic. In addition to preventing COVID-19 deaths, these measures and the change of behaviour may have contributed to reducing the number of deaths from other infectious diseases such as the flu, which also has an impact on the difference from the median. Decreases or increases in the number of deaths from other causes may also have an effect on total deaths. The mortality figures, however, do not provide information on the incidence of individual causes of death.
To give a final evaluation of the mortality development, the number of deaths is then placed in relation to the actual population in order to consider, for instance, the ageing process of the population. The final results which are required for this, including all deaths reported late, are regularly available in the middle of the subsequent year. Information on the relevant results for 2020 is given in the press release of 9 July 2021.
The provisional mortality figures refer to the date of death, not the date on which a death was registered. As the Robert Koch Institute (RKI) currently publishes the reported COVID-19 deaths by death date until Week 28 of 2021, these can presently be compared until that week with the provisional total death figures. This does not include cases for which no information or implausible information on the date of death was reported. The results have not been corrected yet for late reporting and are expected to increase due to late reports. More background information on these data can be found on the RKI website.
More information:
The Federal Statistical Office provides more information on the ad hoc evaluation of mortality figures on its Deaths, life expectancy theme page and its Corona statistics webpage.
For background information on the calculation of excess mortality, current methods and results please refer to an article in “WISTA – Wirtschaft und Statistik“ on mortality figures during the coronavirus pandemic.
