Myanmar Malaria Early Warning System (MMEWS): Multi-sensor satellite data fusion system for monitoring socio-environmental predictors of malaria in Myanmar

Myanmar carries a high burden of malaria within Southeast Asia and also has documented artemisinin resistant Plasmodium falciparum parasites. The country's malaria elimination program has been successful in the past decade and has achieved 82% reduction in malaria cases from 2012 to 2018 . However, small hotspots of malaria continue to exist, especially, in remote rural areas which, are difficult to monitor and eliminate. We present an early warning system (MMEWS) to support malaria elimination efforts in Myanmar. The conceptual framework of MMEWS relies on identifying the malaria risk as a combination of environmental factors that control vector populations and environmental, socio-demographic, and public health factors that control population exposure and vulnerability. Spatially explicit malaria risk is modeled from three components: hazard (the prevalence of malarial parasites / vectors), exposure (human interactions with the landscape that expose them to malarial parasites), and vulnerability (factors that make some people more likely to get the disease). We model the 'hazard' as a combination of environmental conditions supporting vector populations obtained from satellite-derived data for vegetative stress, surface water and land surface temperature and prevalence of parasites obtained from annual parasite incidence reports of the Myanmar Department of Public Health. 'Exposure' is modeled as a combination of where people are (population distribution) and their interaction with the landscape (occupation related exposure). We use our mapped rural settlements for Myanmar to create a gridded population dataset providing a finer resolution population distribution for Myanmar and combine this with our land cover land use basemap modified to identify land covers and uses that might increase the likelihood of being exposed to malaria. Finally, 'vulnerability' is modeled as a combination of access to care (determined as a function of ease /difficulty to reach hospitals) and socio-economic vulnerability (obtained from Myanmar's vulnerability assessment). Our MMEWS model thus uses data fusion to provide spatially explicit MBP. MMEWS identifies developing malaria hotspots across Myanmar at an 8-day reporting frequency and is designed to help with malaria elimination efforts in the country.