Integrated Surveillance for Human and Animal Brucellosis in Kenya: A Predictive Analysis

Abstract

Brucellosis is a bacterial zoonotic disease which poses a significant public health challenge globally. In Kenya, it is a priority zoonosis, causing morbidity and losses in humans and animals. Here, we used monthly surveillance data from 2014 to 2022 from the official human and animal health surveillance databases. We conducted spatiotemporal analysis, tested associations between human and animal brucellosis using Time Series Linear Models, and forecasted the incidence of human brucellosis for twelve months using Seasonal Autoregressive Integrated Moving Average (SARIMA) models. Our analysis revealed a significant disparity in brucellosis cases, with a much higher cumulative number of human cases (4,688,787) compared to animal cases (1214). Human incidence depicted a relatively stable trend, with occasional fluctuations. However, cattle and camel incidences displayed sporadic peaks and troughs. Only cattle brucellosis was significantly associated (estimate: 0.355; 95% CI: 0.004 to 0.707) with human brucellosis. SARIMA models demonstrated reasonable predictive accuracy for human incidence, but incorporating animal data did not significantly improve model performance. Our study highlights the weaknesses in the existing surveillance systems and the need for comprehensive evaluation of the systems and implementation of integrated surveillance to address gaps in surveillance, improve the accuracy of predictive analysis, and enhance early detection for zoonotic diseases.

Affiliated Institutions

• University of Tennessee at Knoxville US
• Ministry of Health KE
• International Livestock Research Institute KE
• Ministry of Agriculture Livestock and Fisheries KE
• University of Nairobi KE
• Washington State University US
• University of Edinburgh GB

Related Publications