Seasonal influenza vaccination coverage and population health burden in Europe during the 2019/2020 and 2021/2022 influenza seasons: an ecological panel analysis using principal-component-adjusted multivariable regression

BACKGROUND: Few studies have examined whether country-level seasonal influenza vaccination coverage is associated with population health burden. We assessed the associations of influenza vaccination coverage with all-cause mortality and disability-adjusted life years (DALYs) across countries participating in the Survey of Health, Ageing and Retirement in Europe (SHARE). METHODS: We conducted an ecological panel analysis using SHARE Waves 8 (2019-2020) and 9 (2021-2022) across participating EU/EEA countries, Switzerland, and Israel. Survey-weighted vaccination coverage and structural indicators were derived among adults aged ≥55 years. All-cause mortality, all-cause DALYs, lower respiratory tract infection (LRTI) DALYs, and chronic disease-specific DALYs were obtained from the Global Burden of Disease database for the same age group. Principal component analysis (PCA) was performed on the pooled country-season dataset to summarize correlated structural covariates. Associations were estimated using linear mixed-effects models with country-specific random intercepts. RESULTS: Mean vaccination coverage increased from 34.5% to 47.6% but remained below the EU target of 75% in many countries. After adjustment for period and PCA-derived covariates, a 1-percentage-point higher vaccination coverage was associated with 14.9 fewer all-cause deaths per 100,000 adults aged ≥55 years (95% CI: -22.8 to -7.1), 328.1 fewer all-cause DALYs per 100,000 (95% CI: -480.4 to -175.8), and 6.3 fewer LRTI DALYs per 100,000 (95% CI: -10.6 to -2.0). Subgroup-specific vaccination coverage was also inversely associated with diabetes DALYs (-12.1; 95% CI: -20.3 to -3.8). CONCLUSIONS: In this two-period ecological analysis, higher country-level influenza vaccination coverage was inversely associated with mortality and DALY burden among adults aged ≥55 years. These findings should be interpreted as hypothesis-generating population-level associations rather than causal effects.