Dysentery in Mainland China, 2004–2020: spatiotemporal dynamics and changes during the COVID-19 pandemic period

Dysentery remains a significant enteric disease globally, causing substantial morbidity and economic burden. Incidence in China has declined over two decades, yet considerable temporal and spatial variation in disease burden and transmission persists across provinces. It also remains unclear how reported dysentery incidence in 2020 deviated from the pre-pandemic temporal pattern. We analysed CISDCP surveillance data from January 2004 to December 2020 across 31 provinces. Joinpoint regression identified long-term trends; time-series decomposition quantified seasonality; Global Moran’s I and LISA assessed spatial autocorrelation and local clusters. Six forecasting models(neural network, ETS, SARIMA, hybrid, BSTS, Prophet) were evaluated per province via a composite standardised index. Province-specific models trained on 2004–2019 data generated pre-pandemic 2020 baselines, compared with observed values. A total of 3,971,083 cases and 661 deaths were reported. Incidence declined throughout, accelerating after 2017 (APC: −11.47 for 2004–2017; −18.85 for 2017–2020). Seasonal patterns were stable year to year, though peak amplitude decreased. Spatial clustering was significant during 2004–2019 and weakened to its lowest in 2020, though 2020 significance depended on spatial weight specification. Beijing and Tianjin had the highest bacillary dysentery incidence; southwest provinces showed higher amoebic burden. Infants aged 0–1 years accounted for the largest single-year age-band share(13.1%). Observed 2020 national incidence was 18.1% below the pre-pandemic baseline projection (95% PI: [6.9%–28.3%]), with wide provincial variation (below the lower 95% bound in six provinces). Dysentery burden in mainland China fell steadily from 2004 to 2020, but seasonality, spatial clustering, and provincial disparities persisted. Observed 2020 incidence fell below the pre-pandemic baseline projection, temporally coinciding with COVID-19 control measures. These findings support hotspot-targeted control, priority actions for infants and childcare settings, and province-level early warning.