This study evaluated long‑term mortality rate changes and patterns of excess deaths across pandemic and post‑pandemic years. The authors aimed to assess cumulative impacts for 2020–2025 while accounting for age structure and long‑term mortality trajectories, and to compare multiple baseline modelling approaches to quantify how methodological choices affect excess death estimates.
Analyses used 38 populations with reliable death registration. Processed, age‑stratified annual deaths and mid‑year populations were prepared in twenty five‑year age bands. Raw inputs derive from the Human Mortality Database, Eurostat, and national statistical offices; processed data and the code to reproduce figures and tables are provided in an openly available repository referenced in the paper.
Four baseline approaches were used to estimate excess mortality for 2020–2025:
The authors emphasize that choice of baseline and handling of long‑term trends materially affect excess death estimates.
Analyses were performed with granular age stratification rather than a single aggregate rate. Two summary metrics are noted for age‑stratified mortality rates across populations:
The paper reports substantial heterogeneity in both slopes and slopes‑of‑slopes across the 38 populations, which motivated the multi‑model approach and the use of shrinkage to stabilize trend estimates.
Results varied by model but show consistent high‑level patterns:
The authors underscore that estimates are model‑dependent and that inclusion of post‑pandemic years materially alters cumulative excess death totals.
Population‑level outcomes were heterogeneous:
Differences among populations were more pronounced during 2020–2023. By 2024–2025, an estimated 33 of the 38 countries exhibited death deficits, indicating a widespread post‑pandemic reduction relative to baselines in those years.
Key methodological and substantive implications drawn by the authors:
Lack of age stratification and relying on unchanged linear trends for baselines can substantially upward‑bias excess death estimates.
Socioeconomically more vulnerable populations experienced higher pandemic mortality but also showed more pronounced post‑pandemic death deficits. The authors suggest this pattern may partly reflect mortality displacement: deaths of people with relatively modest remaining life expectancy during the pandemic resulting in fewer deaths in subsequent years.
Consideration of changing population age structure and long‑term mortality trajectories is essential when estimating excess deaths, and inclusion of post‑pandemic years can change the interpretation of cumulative mortality impact.
All processed age‑stratified data (annual deaths and mid‑year populations by age bands) and analysis code reproducing the paper’s figures and tables are openly available in the authors’ public repository. Raw mortality source data are not redistributed but are available from the Human Mortality Database, Eurostat, and national statistical offices as documented in the repository and supplementary methods.
The authors declared no competing interests and reported adherence to ethical and reporting guidelines. Funding sources noted include NIH grants and the Office of Naval Research. The study is available as a medRxiv preprint with the DOI provided in the source.
Specific limitations beyond model dependence and heterogeneity are not detailed further in the summary; the paper emphasizes transparency of code and data processing to allow replication.
Any additional methodological subtleties, sensitivity analyses, or country‑specific caveats are contained in the full paper and repository; those precise details were not reported in the provided source excerpt.