A data-driven dynamical system model for assessing targeted interventions against brucellosis in Tanzania

Abstract

Brucellosis remains a critical public health and economic concern in Tanzania, particularly in regions with high livestock-human interaction. This paper developed a data-driven dynamical system model to evaluate targeted interventions for controlling brucellosis transmission. The model integrated human, animal, and environmental components, emphasizing public health education and domestic animal management practices. Data from ten regions namely Dodoma, Morogoro, Manyara, Arusha, Mara, Kagera, Shinyanga, Mwanza, Tanga, and Singida, collected between January 2021 and May 2024 were used for parameter estimation. The model was fitted to observed data using a Markov Chain Monte Carlo (MCMC) method with the Adaptive Metropolis algorithm, while a graph-theoretic approach derived the effective reproduction number (R_e) and the force of infection to quantify intervention impacts. Results showed a marked decline in brucellosis transmission. The force of infection dropped from 0.99262 in 2021 to 0.13265 in May 2024, and annual cases fell from 173 in 2021 to 93 in 2024. Public health education improved safe animal handling practices, while enhanced animal management strategies such as fencing, health checks, isolation of infected animals, and controlled breeding curbed disease spread and reduced environmental contamination. Sustained interventions kept R_e < 1, achieving long-term disease control. These findings underscore the importance of continuous monitoring, expanded diagnostics, and region-specific interventions to sustain progress and address remaining challenges.