Scaling Malaria Early Warning to the National Level in Ethiopia

Abstract:

Early warning of the timing and locations of malaria epidemics can facilitate targeting of resources for disease prevention, control, and treatment. In Ethiopia, malaria outbreaks are driven by many factors, including climate variation. The Epidemic Prognosis Incorporating Disease and Environmental Monitoring for Integrated Assessment (EPIDEMIA) system uses real time epidemiological data combined with environmental data from satellites to provide an early detection platform and a 12-week forecast of malaria caseload at the woreda (district) level. Over the last 10 years, EPIDEMIA has been developed and piloted from research to the operational phase, and it has supported malaria early warning in epidemic-prone regions of the Ethiopian highlands.

Our objective was to develop a roadmap for scaling-up and implementing malaria early warning at a national level in Ethiopia and other malaria-impacted countries. To accomplish this goal, we evaluated national-level malaria surveillance data and upgraded the EPIDEMIA software to support malaria forecasting for multiple regions in Ethiopia. We conducted virtual engagements with key stakeholder groups to obtain feedback on EPIDEMIA and discuss the current opportunities and barriers associated with scale-up of malaria early warning systems.

Surveillance data collected through Ethiopia’s Public Health Emergency Management system were found to be suitable for weekly modeling of malaria incidence. We used robust time series models to flag suspect data and imputing missing values. Environmental data on land surface temperature, precipitation, greenness, and surface moisture were obtained from Earth-observing satellites. Distributed lag models were combined with a genetic algorithm to identify optimal groups of districts for modeling P. falciparum and P. vivax malaria. EPIDEMIA generated forecasting reports that incorporated risk maps and time-series control charts and validation reports that characterized geographic variation in forecasting skill. The stakeholder engagement process identified key areas for future technical improvements to EPIDEMIA, but also emphasized the importance of capacity building and developing networks of individuals and institutions to support the broader use of malaria forecasts.