Climate-based modelling and forecasting of dengue in three endemic departments of Peru

Amid profound climate change, incidence of dengue continues to rise and expand in distribution across the world. Here, we analysed dengue in three coastal departments of Peru which have recently experienced public health emergencies during the worst dengue crises in Latin American history. We developed a climate-based spatiotemporal modelling framework to model monthly incidence of new dengue cases in Piura, Tumbes, and Lambayeque over 140 months from 2010 to 2021. The framework enabled accurate description of in-sample and out-of-sample dengue incidence trends across the departments, as well as the characterisation of the timing, structure, and intensity of climatic relationships with human dengue incidence. In terms of dengue incidence rate (DIR) risk factors, we inferred non-linear and delayed effects of greater monthly mean maximum temperatures, extreme precipitation, sustained drought conditions, and extremes of a Peruvian-specific indicator of the El Niño Southern Oscillation. Building on our model-based understanding of climatic influences, we performed climate-model-based forecasting of dengue incidence across 2018 to 2021 with a forecast horizon of one month. Our framework enabled representative, reliable forecasts of future dengue outbreaks, including correct classification of 100% of all future outbreaks with DIR ≥ 50 (or 150) per 100,000, whilst retaining relatively low probability of 0.12 (0.05) for false alarms. Therefore, our model framework and analysis may be used by public health authorities to i) understand climatic drivers of dengue incidence, and ii) alongside our forecasts, to mitigate impacts of dengue outbreaks and potential public health emergencies by informing early warning systems and deployment of vector control resources.Author summary: Dengue fever is a mosquito-borne infectious disease. It is a growing, substantial threat to global public health as rising climatic variability drives the expansion and increase in incidence of the disease. In the northern coastal regions of Peru, dengue is a substantial public health burden with major crises instigated by severe, recent dengue outbreaks. Here, focusing on three Peruvian departments of Piura, Tumbes, and Lambayeque, we describe the extent to which current and past recent climatic conditions, such as temperature, precipitation, and sea-surface temperatures, shape the patterns in reported dengue cases. Furthermore, we display how understanding such relationships with climatic conditions can be used to model historical trends in dengue cases, which can be helpful for addressing incomplete dengue surveillance. Finally, the study demonstrates how climate-model-based forecasting of dengue cases can be employed for reliable prediction of potential future outbreaks up to one month ahead of time. The insights can be used by public health authorities before and during dengue outbreaks to potentially reduce the morbidity, mortality, and strains on health services by informing public health policies such as information campaigns, early warning systems, and targeted mosquito control interventions.