Spatiotemporal Early Warning of Human Brucellosis in Hubei Province Using Coupled Ecological-Time Series Models

This study explored transmission patterns of brucellosis in Hubei Province, a Class II region in China, through analysis of spatiotemporal distribution and environmental factors to provide evidence for precise prevention.

We analysed epidemiological data from Hubei Province (2010-2023), including human cases, demographics, livestock numbers and environmental factors. The MaxEnt model was used for spatial risk prediction, while the seasonal autoregressive integrated moving average (SARIMA) model analysed temporal patterns and warning effectiveness.

Human brucellosis showed a 'rapid rise-effective control-local rebound' pattern, with incidence rates rising from 0.0087/100,000 (2010) to 0.5932/100,000 (2015), declining to 0.1535/100,000 (2019) and rebounding to 0.4508/100,000 (2023). The epidemic displayed seasonal peaks (April-September) and affected primarily males (71.13%) and farmers (76.54%). Spatial distribution showed a 'high-north low-south' pattern, spreading from Suizhou to neighbouring areas. The MaxEnt model identified high-risk areas (probability ≥ 0.713) in livestock-intensive regions, with sheep density (55.1%) and annual mean temperature (9.9%) as key factors. The SARIMA model indicated earliest warnings in high-risk areas (mid-late April), followed by medium-risk (early-mid April) and low-risk areas (late May).

Brucellosis transmission in Hubei is influenced by the livestock industry, control measures and climate. The integrated MaxEnt-SARIMA model provides risk classification and differentiated warning schemes, offering guidance for targeted prevention across different risk areas.