ARIMA-based prediction of global trends in Trichostrongylus infection among humans, ovines, and bovines

Abstract Trichostrongylus spp. are globally distributed gastrointestinal nematodes that affect ruminants and humans, posing significant veterinary and public health challenges. Despite their zoonotic potential, the temporal dynamics of Trichostrongylus infection remain poorly understood globally. This study aimed to estimate long-term trends in Trichostrongylus prevalence in humans, ovines, and bovines using time series modelling. A systematic review identified 240 eligible studies with annual prevalence data across 60 countries. Following Kalman smoothing, annual prevalence time series were constructed for each host species covering 1947–2024 for humans, 1966–2024 for ovines, and 1962–2024 for bovines. ARIMA models were fitted separately: ARIMA(0,1,1) for humans, ARIMA(3,0,0) for ovines, and ARIMA(0,1,1) for bovines. Model selection was based on Stationary R 2 , RMSE, MAPE, and the Ljung-Box Q test for residual independence. Forecast 95% confidence intervals were reported to convey uncertainty in the projected trends. All three models demonstrated good in-sample fit and adequate residual diagnostics. Infection rates in humans and bovines are projected to decline, from 4.64% to 3.73% in humans and from 20.11% to 11.76% in bovines by 2034. In contrast, the ovine model forecasts an increase in infection rates, from 6.50% to 15.56%. This increase in ovines may reflect greater pasture exposure and environmental persistence of infective larvae, while improvements in hygiene and livestock management likely contribute to the declining trends observed in humans and bovines. The rising infection rate in ovines, coupled with sustained zoonotic risk, underscores the need for integrated One Health surveillance and control efforts.