Transmission acceleration outperforms the endemic-channel threshold for dengue outbreak detection

Outbreak detection determines whether surveillance can support timely public-health action, yet outbreaks are still identified operationally through heuristics derived from historical incidence. In dengue surveillance, this has made endemic-channel thresholds the routine basis for outbreak alerts without independent validation of whether they detect epidemics early enough to guide response. Here we show that transmission acceleration provides a substantially stronger basis for seasonal dengue outbreak detection than the conventional rolling-baseline outbreak threshold. We evaluated detector performance across a densely populated hyperendemic city with sustained transmission, 17 tropical regions spanning diverse climatic conditions, population densities and surveillance capacities, and eight dengue-endemic countries. Across these settings, acceleration-based detectors consistently outperformed the threshold in pre-peak burden capture, lead time, warning persistence and sensitivity. In the city-scale evaluation, acceleration-based detection captured 4,257 cases per season before the epidemic peak, compared with 1,119 for the outbreak threshold, and achieved 100% sensitivity versus 30% while triggering within the four-to-eight-week actionable window. The same ordering is held at regional and country scales. These findings establish transmission acceleration as a more valid and operationally useful basis for dengue outbreak detection from routine surveillance data.