Estimating COVID-19-Related Excess Mortality Excluding Seasonal Phenomena in Belgium

Abstract

Excess mortality, rather than reported COVID-19 deaths has been suggested to evaluate the impact of the SARS-CoV-2 induced Corona Virus Disease (COVID-19) pandemic on mortality. However, the relationship between excess mortality and COVID-19 mortality is perturbed by seasonal phenomena, such as extreme temperatures and seasonal influenza. Models used to estimate excess mortality often ignore these underlying patterns. We propose a dynamic linear state-space model to estimate all-cause mortality, which accounts for extreme temperatures above 25°C and seasonal influenza via the Goldstein index. The state-space model prediction of the excess mortality that is not explained by heat waves and seasonal influenza coincides with the reported COVID-19 mortality in the year 2020 in Belgium.