A one health regional framework for predictive dengue surveillance and early warning in Southeast Asia

Dengue continues to pose a significant public health challenge across Southeast Asia, where rapid urbanization, climate variability, and highly heterogeneous landscapes hinder timely outbreak detection and response. The SEA-ROADS initiative (South East Asia – Regional One Health Approach for Dengue Surveillance), launched in 2025 across Phnom Penh (Cambodia), Bangkok (Thailand), Vientiane (Laos), and Hanoi and Ho Chi Minh City (Vietnam) for predictive dengue surveillance. Standardized dengue and mosquito surveillance was implemented in all cities using a unified methodological framework. Urban areas were first delineated into spatial clusters based on building-level voronoi polygons and five-dimensional feature; vectors capturing structural and landscape characteristics. These clusters were refined using minimum spanning trees, principal component analysis, and hierarchical clustering to derive representative urban classes per city. Mosquito surveillance conducting using 72 larvitraps per city (three traps per class), and sampled weekly to quantify Ae. aegypti and Ae. albopictus abundance. Complementary epidemiological monitoring consisted of monthly dengue case collected in collaboration with local health authorities. Human mobility patterns were integrated into the surveillance framework using point- of-interest (POI) data from Meta (Facebook). Four to six representative urban classes were identified per city, ranging from highly vegetated open areas (Class 1), low-height semi-open areas (Class 2), and moderately dense zones (Class 3), to very dense small-building areas (Class 4), extremely dense micro-building zones (Class 5), and high-rise clusters (Class 6). POI counts and category diversity were aggregated into 200-m grid cells, interpolated to fill gaps, and summarized using principal component analysis; the first PCA axis was used as a proxy for urban centrality. Preliminary temporal analyses revealed a moderate positive and short-term (0–4 weeks) delayed effect of Ae. aegypti abundance on dengue incidence and Ae.albopictus remained comparatively lower and more stable. By integrating multisectoral datasets into predictive models, SEA-ROADS enhances dengue risk forecasting, identifies potential cross-border transmission pathways, and supports the development of national and regional dashboards for near real-time surveillance. This coordinated One Health framework strengthens regional capacity for early detection, targeted intervention, and proactive dengue control across Southeast Asia.