Excess Mortality (2020-2023) as Proxy of COVID-19 Deaths?

Abstract

Importance: Concerns regarding excess mortality estimates and the subjective nature of diverse models utilized have emerged. We examined its theoretical underpinning by exploring two popular excess mortality models based on regression and time series analyses that highlight their weaknesses in forecasting excess deaths during COVID-19 emergency.
Observations: Excess mortality estimates are errors/residuals of prediction models increasingly used to determine the number of unreported deaths from COVID-19. That several prediction models are used to model baseline excess deaths underscores the lack of a definitive choice thereby signposting its subjective nature. A general lack of assessment of the assumptions governing such models was another drawback in relying on estimates of
excess mortality derived from them.
Conclusions and Relevance: In assessing the impact of COVID-19 (or any public health
emergency), reported death counts and other mortality statistics, when combined with relevant auxiliary information, can offer a better view of the pandemic impact rather than reliance on a subjective metric such as excess death which can be misleading. More importantly, mathematical modeling though useful in an unfolding pandemic, once data become available, this should supersede forecasted estimates in decision-making or impact assessment.