Methodology for the calculation of the weekly expected number of deaths in 2023 and Cause of death statistics: Survey, coding and analysis of causes of death
In recent weeks, various media have reported on the methodology used to calculate the weekly expected number of deaths in 2023 and on the survey, coding and analysis of the cause of death statistics. With this statement, the FSO wishes to contribute to a better understanding of the methodology used for both statistics. The basic principles of these statistics are described in the methodological explanations "Official statistics on deaths, excess mortality, causes of death and notifiable diseases", which were updated for the fifth time in May 2022.
Methodology for the calculation of the weekly expected number of deaths in 2023
Data set: Deaths in Switzerland are reported daily to the FSO by the civil registry offices.
In addition to the actual (observed) deaths, the weekly expected deaths (expected values) and their confidence intervals serve as the basis for calculating excess mortality. We use the term excess mortality when the observed number of weekly deaths is above the upper limit of the confidence interval. The excess mortality itself is then calculated from the difference between the observed deaths and their expected value. In this respect, excess mortality, like the expected value and its confidence interval, is a statistical estimate and not based purely on observation. These three components have been used in Switzerland since 2013 for the continuous weekly monitoring of mortality in the population (mortality monitoring) and were introduced well before the COVID-19 pandemic. Excess mortality is also estimated in other countries based on weekly observed and expected deaths. In some cases, different methods are used to calculate the expected values. However, the calculation of excess mortality is always based on the difference between the observed deaths and their expected value.
Mortality monitoring depends on an estimate made at the beginning of the year of the mortality that will occur throughout the year. This estimate must be as realistic as possible. This forecast is always based on the mortality of the previous years. The annual forecast thus reflects the mortality trend of previous years for the current year.
In the past, an increased number of deaths due to severe flu epidemics or heatwaves (such as in 2015), could be included in the estimate of the weekly expected values for the following years without having too great an influence on the multi-year trend. However, the exceptional pandemic year of 2020 could not be used directly to estimate the weekly expected values for the following year of 2021, because the greatly increased mortality in 2020 would have led to an overestimation of the weekly expected values for 2021 to such an extent that they would have been outside the expected trend. For the expected values for 2021, therefore, the expected values for 2020 were reused. As increased mortality due to the pandemic occurred again in 2021, that year could not be used either for the estimate of expected values in 2022. For this reason, the method for estimating the expected values of 2022 was adjusted. In the three periods of excess mortality in 2020 and 2021, the weekly observed number of deaths in the age group 65 and above were replaced with the expected number of deaths if the value was more than one standard deviation above the expected value. No adjustment was made for the 0-64 age group. The same procedure was also used to estimate the expected values for 2023, whereby the weekly observed number of deaths in 2022 was included unchanged in the estimate despite the increased mortality in that year.
In contrast to the causes of excess mortality in 2020 and 2021, the reasons for excess mortality in 2022 so far remain largely unknown. It is likely that excess mortality in 2022 will for the most part be due to SARS-CoV-2 infections, but COVID-19 might not be given as the main cause of death in all cases. At the moment it is not clear whether and to what extent the virus can worsen existing diseases or lead to new ones. Various theories on the excess mortality in 2022 are currently being discussed in the scientific community. The cause of death statistics for 2022 will provide further information on the causes of excess mortality in that year. The publication of the cause of death statistics 2022 is planned for the end of 2023.
Cause of death statistics: Survey, coding and analysis of causes of death
Data set: The causes of death survey is based on the reports of the civil registry offices and the causes of death reported by doctors.
Since 1995, the coding and analysis of causes of death have been based on the 10th version of the International Classification of Diseases (ICD-10) and the relevant guidelines issued by the World Health Organization (WHO). The ICD-10 and the corresponding coding guidelines are regularly revised by the WHO to incorporate current scientific findings and new diseases. Over the course of the COVID-19 pandemic, the WHO introduced new ICD-10 codes for the coding of COVID-19 deaths and also created corresponding coding guidelines. These have been implemented internationally and also by the FSO.
Neuchâtel, 14 February 2023
