Abstract
OBJECTIVES: This study aimed to investigate the impact of meteorological factors on the incidence and multi-route transmission dynamics of hepatitis E virus (HEV) in Jiangsu Province, China, during the pre-COVID-19 era (2005-2018), and to develop predictive models for informing public health interventions. STUDY DESIGN: A dual-model study integrating the Multi-Host and Multi-Route Transmission Dynamic Model (MHMRTDM) and Generalized Additive Model (GAM) was employed to quantify meteorological impacts on multi-route HEV transmission. METHODS: HEV incidence data (2005-2018) and meteorological variables from provincial and national agencies were analyzed. The MHMRTDM quantified transmission rate coefficients (β, β (w) and β (p)'). GAMs linked the transmission coefficients and incidence to meteorological factors, validated using 2017-2018 data. RESULTS: The optimal GAM integrated with the MHMRTDM was established (lowest GCV = 1.705 × 10(-21), R (2) = 0.980, lowest RMSE = 3.682 × 10(-11), lowest MAE = 2.987 × 10(-11)). Analysis of four dependent variables (incidence, β, β (w) and β (p)') revealed distinct climate-driven patterns: (1) Incidence exhibited dual seasonal peaks linked to atmospheric pressure, sunshine duration, and humidity; (2) Host-to-person transmission (β (p)') was most sensitive to climatic conditions, peaking at 1013 hPa and declining sharply above 75% humidity, while susceptible person-to-infected person (β) and environment-to-person (β (w)) transmission were primarily modulated by humidity and wind speed; (3) The GAM validation confirmed robust performance for transmission coefficients (p < 0.001). Predictions for 2019-2021 highlighted persistent seasonal bimodality, reinforcing the model's utility for outbreak forecasting. CONCLUSION: Meteorological factors drive HEV transmission through distinct pathways, with host-to-person interactions being particularly climate-sensitive. While the GAM provided valuable insights, future research incorporating behavioral and land-use factors, as well as causal inference models, will be critical for improving the understanding and predictive accuracy of HEV transmission dynamics.