Municipality-level ecological analysis of COVID-19 mortality and infections in Austria, 2020 to 2023

@misc{riedmann2026municipalitylevel,
  title = {{Municipality-level ecological analysis of COVID-19 mortality and infections in Austria, 2020 to 2023}},
  author = {U. Riedmann and A. C. Stevenson and C. Dibben and W. B. Grant and M. Sprenger and L. Zgaga and J. Woltsche and P. Wolf and W. Maerz and J. Ioannidis and S. Pilz},
  howpublished = {medRxiv},
  year = {2026},
  month = oct,
  doi = {10.64898/2026.10.15.26365614},
  url = {https://www.medrxiv.org/content/10.64898/2026.10.15.26365614v1},
  note = {CC BY},
}

TY  - UNPB
TI  - Municipality-level ecological analysis of COVID-19 mortality and infections in Austria, 2020 to 2023
AU  - U. Riedmann
AU  - A. C. Stevenson
AU  - C. Dibben
AU  - W. B. Grant
AU  - M. Sprenger
AU  - L. Zgaga
AU  - J. Woltsche
AU  - P. Wolf
AU  - W. Maerz
AU  - J. Ioannidis
AU  - S. Pilz
PY  - 2026
DA  - 2026/10/15
PB  - medRxiv
DO  - 10.64898/2026.10.15.26365614
UR  - https://www.medrxiv.org/content/10.64898/2026.10.15.26365614v1
AB  - BackgroundPrevious studies indicated that regional fluctuations in COVID-19 impact were associated with socioeconomic, demographic and environmental factors early in the pandemic. It is unclear whether factors associated with regional COVID-19 mortality and SARS-CoV-2 infections changed over the course of the pandemic.  MethodsWe evaluated the association of ecological factors with municipality-specific COVID-19 mortality in the national population of Austria between January 2020 and June 2023. Analyses were conducted separately per calendar year and pooled across years via random-effects meta-analysis. Outcomes were analysed using univariable and multivariable zero-inflated negative binomial (ZINB) regression, including annually averaged environmental exposome exposure (including air pollution and UV radiation), socioeconomic, demographic and infection dissemination predictors with significance threshold set to p<0.005. Secondary analysis modelled SARS-CoV-2 infections using of negative-binomial model (NB). Two additional approaches for smaller predictor set selection were applied as sensitivity analyses.  ResultsWithin the entire observation period 18,430 COVID-19 deaths were reported. In univariable analyses (pooled RR (99.5% CI)), COVID-19 mortality was significantly associated with the proportion of commuters (0.79 (0.73-0.86)), residents aged 75-84 (1.18 (1.01-1.37)) and [≥]85 years (1.29 (1.04-1.60)), residents with at most compulsory education (1.15 (1.02-1.29)), employed residents aged 15-64 (0.89 (0.85-0.93)), and homeowners (0.91 (0.83-0.99)), as well as with precipitation (0.94 (0.89-0.99)) and the average time with a mean temperature of 5-15 {degrees}C (0.95 (0.90-0.99)). In multivariable analyses the share of population aged 85 or older (1.30 (1.09-1.54)) and home ownership (0.92 (0.87-0.98)) were independently significant. Mortality predictor estimates were relatively stable across pandemic years. Across 5,963,882 documented SARS-CoV-2 infections share of employed in 15- to 64-year-olds (1.01 (1.01-1.02)), part-time employment rate (1.01 (1.00-1.02)) and district level nursing care allowance recipiency (0.97 (0.95-0.98)) were associated with detected infections. Overnight stays, population 85 or older, income, compulsory education or less and air pollution all had multiple significant estimates with opposite directionality, leading to non-significant overall associations. Sensitivity analyses outcomes mirrored main analyses outcomes.  ConclusionsThroughout the pandemic in Austria, demographic and socioeconomic factors were associated with COVID-19 mortality and infections, but many associations, especially for infections, were unstable over time. These findings underscore the importance of structural vulnerabilities in epidemic impact.
ER  -