Answers are grounded in this paper and cite their sources.
Background Hepatitis A virus (HAV) and hepatitis E virus (HEV) are both acutely transmitted via the fecal-oral route and pose a significant global burden of infection. Existing disease predictions often do not account for environmental chemical pollutants, and there is insufficient epidemiological evidence linking mercury and persistent organic pollutants (POPs) to hepatitis incidence and mortality. Methods This study utilized hepatitis surveillance data from China between 2005 and 2020 along with pollutant emission data from the EDGAR database. A SARIMA model was employed for univariate time series forecasting, combined with a BP neural network for multivariate predictions. Key toxic factors were identified using LASSO regression, and WQS regression was applied to analyze mixed exposure effects stratified by age. Furthermore, a random forest model was used to explore sensitive factors and interactions during different dominant periods of hepatitis epidemics. Results The SARIMA monthly model demonstrated superior predictive performance for HAV and HEV incidence, while the quarterly model was found to be more suitable for mortality rate predictions; overall, the predictive performance for HAV was slightly better than that for HEV. Gas-phase oxidized mercury (GOM), particulate-bound mercury (PBM), polycyclic aromatic hydrocarbons (PAHs), and polychlorinated biphenyls (PCBs) were identified as core influencing factors. There was significant heterogeneity in key toxicants and risk effects across different age groups, with those aged 60 and above being more sensitive to particulate-bound mercury exposure. The shift in dominant hepatitis epidemic types significantly altered environmental driving factors: the HAV-dominant period was characterized mainly by GOM, PCB, and HCB, while the HEV-dominant period was characterized by GEM, PCDD/F, and HCB. The dual-dominant period exhibited the most pronounced features of co-exposure to multiple toxicants and yielded the best model fitting results. Conclusions Environmental mercury and POPs are significantly associated with the risk of HAV and HEV infections, with notable age differences and heterogeneous effects during different epidemic periods. Integrating environmental indicators into machine learning and time series models can effectively enhance the predictive accuracy of hepatitis incidence and mortality rates. The findings provide a scientific basis for environmental health risk prevention and targeted interventions for hepatitis across different populations and stages.
No takes yet. Share an insight, caveat, or question.
Weiming Hou (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: