An index-based system for early alerts of potential zoonotic disease outbreaks

Abstract

The Information System on Wildlife Health (SISS-Geo) is a free platform designed for the real-time collection of georeferenced data on wildlife health and environmental conditions via mobile devices. It serves as a collaborative tool, enabling health professionals, researchers, environmental managers, and the general public to report information about wildlife health occurrences directly within the system. Since its launch in 2014, SISS-Geo has been successfully applied in supporting decision-making during significant wildlife health events in Brazil. In this paper, we introduce a dynamic alert system based on a Multi-Attribute Zoonotic Alert Index (Z-Alert) to enhance disease outbreak detection and response in wildlife. The proposed approach improves the current alert system of the SISS-Geo platform by integrating clustering techniques with multi-objective optimization. The clustering stage allows to group SISS-Geo records in a way that reflects both spatial and temporal variations, capturing relevant patterns for epidemiological surveillance. Each identified cluster is assigned a numerical alert index, the Z-Alert index, reflecting its criticality level, based on optimally weighted attributes such as the percentage of dead animals, temporal interval, and geographical spread. The system offers customization alert thresholds, allowing health managers to adapt the alert system to their specific needs. To validate the model, we used historical yellow fever case data from Brazils Ministry of Health, achieving over 94% accuracy in identifying critical clusters. By enabling the prioritization of prevention and investigation actions, the system strengthens the responsiveness of health managers, optimizing resource allocation and promoting smart economic strategies for outbreak containment.