Particulate matters 2.5 and its organic constituent on seasonal influenza transmission: Insights from Bayesian spatiotemporal modelling in mainland China

Seasonal influenza transmission is influenced by environmental and climatic factors, but the role of particulate matter (PM 2.5 ) and its organic constituent remains unclear. This study uses Bayesian spatiotemporal models to explore the effects of PM 2.5 and Organic Matters (OM) on influenza transmission and identify spatiotemporal heterogeneity across socio-economic statuses to enhance early warning systems (EWS) for seasonal influenza. Weekly influenza cases number from 2014–2018 were collected at the county level in Guangzhou, Beijing, and Wuhan. PM 2.5 and OM data were sourced from Tracking Air Pollution, and socio-economic factors, including floating population, population density, and GDP per capita, were obtained from government records. A spatiotemporal model with spatial and temporal autoregression terms, implemented under the Bayesian inference framework, and generalized linear models (GLMs) were used to assess associations and compared with spatiotemporal models. PM 2.5 was positively associated with increased influenza transmission in all cities under the Bayesian spatiotemporal models (Guangzhou: Relative Risk (RR)=1.16, 95% Credible interval (95% CrI): 1.08–1.23; Beijing: RR=1.37, 95% CrI: 1.18–1.60; Wuhan: RR=1.18, 95% CrI: 1.10–1.26). OM showed significant associations in Guangzhou (RR=1.16, 95% CrI: 1.10–1.23) and Wuhan (RR=1.15, 95% CrI: 1.08–1.22). Spatiotemporal models had better performance than regression models and identified high-risk clusters of influenza in suburban areas. PM 2.5 and OM significantly influence seasonal influenza transmission, even after adjusting for socio-economic factors. Bayesian spatiotemporal models outperformed GLMs in capturing environmental effects and identifying high-risk areas, supporting their use in EWS to enhance public health interventions. • Spatiotemporal models provide robust influenza risk estimates across three cities. • PM 2.5 significantly impacts influenza, with highest risk observed in Beijing. • Socio-economic factors influence influenza transmission dynamics. • Spatiotemporal models outperform GLMs in capturing influenza-environment interactions.