A Multi-Criteria Evaluation and a Proposed Integration Framework for Global One Health Epidemic Intelligence Systems

Zoonotic pathogens make up roughly 61% of all recognised human pathogen species, yet early detection remains hindered by fragmented surveillance. Coordinated epidemic response is currently undermined by incompatible modalities, including official reporting delays, low signal precision in event-based systems, and an absence of prospective validation for AI and LLM tools. Following PRISMA-ScR guidelines, we conducted a scoping review (December 2025-January 2026) to evaluate 55 global epidemic intelligence (EI) tools. These tools spanned five categories: official indicator-based systems (IBS), event-based open-source intelligence (EBS), AI/NLP/LLM frameworks, participatory field networks, and genomic/ecological meta-surveillance. Tools were mapped using a heuristic framework of eight ordinal dimensions (scored 1-5), including timeliness, reliability (positive predictive value, PPV), and accessibility. The full evaluation matrix and three sensitivity analyses are provided for transparency. Event-based systems were found to detect a substantial share of outbreak events ahead of official notification (e.g., 43% of avian-influenza events for HealthMap), but with lower predictive reliability. Of the 55 tools evaluated, 39 (71%) are actively maintained and publicly accessible. However, only 20 (36%) possess published quantitative evaluations, and crucially, none of the 12 Category-3 AI/NLP/LLM systems has been prospectively validated in operational settings. To our knowledge, this is the first comprehensive, multi-criteria mapping across all five EI modalities. Our findings indicate that the primary bottleneck in early detection is not a lack of tools, but rather the integration gap between them. To address this, we propose a conceptual multi-source architecture that fuses IBS, EBS, AI, genomic, and field inputs through a unified validation layer, providing a framework for coordinated One Health surveillance.