Last Updated: 14/03/2023

Improving Response to Malaria Outbreaks in Amazon-Basin Countries


To improve malaria response in the Amazon by enhancing knowledge on when where, and which targeted interventions will have the greatest impact. 

Principal Investigators / Focal Persons

William Kuang-Yao Pan

Rationale and Abstract

There is a critical need for improved malaria control—since 2011, no region in the world has experienced a larger increase in malaria than the Amazon. Several events contributed to this rise: extreme weather (i.e., El Nino), expanded resource extraction, political unrest in Venezuela, and withdrawal of the Global Fund from South America. The unprecedented malaria resurgence has been particularly high near border regions where migration and poor health care facilitate transmission. The current surveillance system has a 4-week delay in cases reported, which is completely inadequate, resulting in reactive vs. preventive intervention strategies. To respond, the team developed a Malaria Early Warning System (MEWS) with NASA support for Loreto, Peru, where over 90% of malaria cases in Peru occur. The MEWS forecasts outbreaks with >90% sensitivity and >75% specificity 8-12 weeks in advance in sub- regions (EcoRegions using unobserved component models [UCM]) and districts (via spatial Bayesian models), and fits community-based agent based models (ABMs) to evaluate behavioral factors associated with transmission. However, gaps remain: MEWS has unknown performance outside of Peru; it does not incorporate migration; forecasts are not downscaled for hotspot detection; forecasting performance is poor near border regions; and the models are not integrated across scales. 

This project will significantly improve current surveillance efforts by providing both current estimates and forecasts of malaria using state-of-the-art climate, hydrology and land cover models. The MEWS is expanded by obtaining surveillance and population data from Ecuador and Brazil, and merging these with hydro-meteorological data. New EcoRegions that ignore administrative borders are defined and UCMs are applied. Spatial Bayesian models are used to estimate both district- and downscaled sub-district level malaria incidence. This proposal responds to the WHO 2016-2030 Global Technical Strategy for Malaria and the recent initiatives by the Pan American Health Organization calling for improved malaria surveillance as a core intervention to improve response to high malaria burden.


Sep 2021 — Aug 2026

Total Project Funding



Related Resources

Related Projects