Abstract / Summary
Successful pandemic preparedness and sustainable healthcare security increasingly rely on One Health—linking humans, animals, and the environment. Novel real-time geolocated data sources from social media, the web, digital traces, and Internet of Things (IoT) sensors, combined with traditional surveillance data, are increasingly used to power AI models and enhance environmental sensing that is essential for preparedness and response in the One Health context. Despite the abundance of data streams and sensing data, the potential of AI and IoT environmental monitoring for One Health remains unfulfilled. The OHSEW (One Health Surveillance and Early Warning) framework architecture fills this gap by defining a human-centric architecture for One Health surveillance, early warning, and response, leveraging AI modelling and IoT environmental sensing, designed in particular for monitoring vector-borne diseases, zoonoses, and other environmental reservoirs of infectious diseases where the risk of spillover to human populations is high. In the OHSEW framework, we identify a rich spectrum of multiple heterogeneous data sources and a pipeline for improved surveillance, data visualisation, and training predictive AI models. These include open-source population and socioeconomic data, land use, satellite sensing, and weather forecasts, enriched by IoT environmental sensing (providing microclimate data streams) and traditional routine surveillance data reported by public health and environmental health experts. These models generate real-time data and actionable evidence for public health managers and policymakers, visualised through customisable dashboards, and provide early warning signals enabling rapid response and effective risk communication with the public. In this article, we outline the framework architecture, individual components, data pipeline, and configurable and localisable setup. To demonstrate the feasibility of the architecture, components of the OHSEW framework were piloted in real-world deployments of mosquito surveillance systems at two sites, Madeira, Portugal, and Cabedelo, Brazil. The architectures and results of these real-world deployments are presented to illustrate the feasibility of the OHSEW framework, with directions for further research outlined.