Spatial disparities in the mortality burden of the covid-19 pandemic across 569 European regions (2020–2021)

The COVID-19 pandemic, emerging in December 2019, caused a significant global increase in deaths. While the World Health Organization declared an end to the global health emergency in May 2023, assessing the pandemic's overall burden, particularly during its peak in 2020 and 2021, remains crucial. Excess mortality, the difference between observed deaths and those expected in the absence of the pandemic, is considered the gold standard for estimating this impact. Previous studies often focused on national-level analyses, hindering comparisons between regions across different countries due to varying methodologies and indicators. This study addresses these limitations by providing a detailed analysis of the spatial disparities in excess mortality across 569 comparable-sized regions in 25 Central and Western European countries during 2020 and 2021. The researchers used a robust forecasting approach based on CP-splines to account for regional diversity and provide confidence intervals for excess mortality measures. This methodology is crucial given the inherent variability associated with fine-grained analyses often involving small populations. This work is essential because it helps to quantify the pandemic's true impact in different regions of Europe and to understand the regional patterns and the evolution of these patterns across the two pandemic years. The spatial approach provides insights into the spread of the pandemic and the impact of geographical factors on excess mortality rates. This study, therefore, provides a comprehensive and nuanced understanding of the pandemic's heterogeneous impact across Europe.

Numerous studies have quantified the impact of the COVID-19 pandemic using excess mortality, but most have focused on national levels. A few studies examined regional variations within single countries, but methodological inconsistencies (e.g., baseline mortality estimation using pre-pandemic levels or forecasting techniques, different indicators for assessing excess mortality) prevented meaningful cross-country comparisons. Only two previous studies enabled a simultaneous comparison of regional excess mortality across multiple European countries for 2020, and another covered 2020-2022 but excluded data for several countries. This current research builds upon this limited existing literature by expanding the analysis to include a larger number of regions across a more extensive array of European countries and by considering both 2020 and 2021.

This study utilized subnational death and population counts from Eurostat, the Human Mortality Database, and national statistical offices for 25 European countries. To ensure comparability, the researchers primarily used NUTS-3 level data, with adjustments made for data availability and territorial changes. The data was harmonized into single-year age intervals up to 95+. A total of 569 harmonized spatial units were analyzed, with population sizes ranging from 40,000 to 6,750,000. To estimate baseline mortality, the researchers employed a CP-spline approach, combining two-dimensional P-splines with historical demographic patterns specific to each region. This non-parametric approach is flexible, generates smooth age profiles and time trends, and enhances robustness for smaller populations. The method dynamically optimized region-specific timeframes to forecast 2020 and 2021 mortality, ensuring accurate trend validation. The researchers selected the starting year (up to 2010) that minimized the deviance between observed and forecasted 2019 mortality. Age-Standardized Years of Life Lost (ASYLL), calculated by summing the years of life lost for all individuals in each age group and applying age-standardization using the 2013 European Standard Population, was used as the primary indicator for assessing excess mortality, as it facilitates comparisons between populations with varying age structures and can be added across time to assess the full burden of the pandemic in 2020 and 2021. All calculations were performed using R version 4.3.1 and ArcGIS 10.8.1 for mapping. The study provides extensive details on data sources and the analytical procedure.

The study revealed significant spatial disparities in ASYLL across the 569 European regions during 2020 and 2021. In 2020, high ASYLL values were concentrated in northern Italy, southern Switzerland, central Spain, and Poland, coinciding with early outbreak areas. Regions in Western Europe generally experienced smaller losses or even gains. However, by 2021, a clear East-West gradient emerged, with Eastern European countries (particularly Slovakia, Hungary, and Latvia) exhibiting the highest losses, especially among men. Western European regions showed more moderate losses or gains. Across both years, only seven regions showed gains in 2020, and four in 2021, while 362 regions experienced losses in 2020 and 440 in 2021. Significantly, only eight regions had ASYLL exceeding 20 years per 1000 population in 2020, compared to 75 in 2021. Analysis of the combined impact of 2020 and 2021 showed that Central and Eastern European (CEE) countries generally experienced higher excess mortality, particularly among men, although there was significant variation within countries. The study further examined the relationship between pre-pandemic (2015-2019) life expectancy and excess mortality in 2020 and 2021. While a negative association was apparent overall, the relationship was unclear when stratifying by region (CEE vs. West) and year, with varying positive and negative associations depending on the year and the geographic location. The study also highlighted the most and least affected regions, showing that Eastern European regions (particularly in Poland and Slovakia) were among the most affected, while Scandinavian countries, Germany, and France had regions among the least affected or even with gains. The disparity between men and women in years of life lost was more pronounced, with significantly higher losses observed among men, especially in CEE countries.

This study's findings at a fine geographical level underscore the importance of regional analysis for understanding the COVID-19 pandemic's impact. National-level aggregates mask considerable regional variations, particularly in 2020, highlighting the need for tailored health policy responses. The shift from initial outbreak areas to an East-West gradient in 2021 suggests evolving spatial dynamics influenced by factors such as virus strains and variations in vaccination uptake and compliance with public health interventions. The lack of a clear relationship between pre-pandemic mortality levels and excess mortality in 2021, when analyzing Western and Eastern European countries separately, suggests the influence of other factors like socio-economic conditions and public health policies on pandemic outcomes. These disparities should be explored in future studies to fully understand pandemic’s impact. The study's regional analysis contributes to a more complete understanding of the pandemic's complex interplay with geographic and societal factors.

This study offers a comprehensive regional-level analysis of COVID-19 excess mortality across 569 European regions during 2020 and 2021, demonstrating substantial spatial disparities and a shift in geographical patterns over time. The findings highlight the need for regional-specific public health strategies. Future research should investigate the underlying causes of regional variations, considering factors such as socio-economic conditions, health system characteristics, and policy interventions. A comparative analysis with other historical mortality crises would further contextualize the COVID-19 pandemic's impact.

The study's limitations include the lack of spatial autocorrelation consideration in the mortality modeling, which might affect the accuracy of excess mortality estimates. Furthermore, the analysis does not adjust for potential harvesting effects, and the focus on 2020 and 2021 due to data availability may limit the study's comprehensive scope. Future research incorporating spatial dependence in modeling and expanding the timeframe of analysis could provide a more robust and comprehensive understanding of the COVID-19 pandemic’s mortality burden.