In May 2017, hydrologist Antarpreet Jutla and a team of civil scientists used predictive algorithms to forecast an outbreak of cholera in Yemen. Cholera, a waterborne bacterial disease, primarily blooms during hot and dry seasons in coastal, developing countries lacking sophisticated sanitation and water infrastructure. To identify areas where these conditions are prevalent, Jutla’s team used satellite imagery to monitor temperature patterns, water storage, population migration, regional topography, and precipitation throughout Yemen. That data was fed into a processing algorithm that predicted areas most likely to experience an outbreak in the near future—particularly cities in West Yemen along the Red Sea.

Less than a month later, the model’s predictions rang true. Because the algorithms were built and tested using data from other regions, such as the Bengal Delta in South Asia, the team did not anticipate such accurate results in Yemen and chose not to preemptively warn local officials of the model’s predictions. In June, highly populated cities along the country’s West coast (including Al Hudaydah, Hajjah, and Taiz) saw tens of thousands of inhabitants suffer moderate to severe cholera symptoms.

The epidemic confirmed the model’s effectiveness beyond the team’s expectations. The refinement of such a system to a near-certain level of accuracy would offer huge advantages to hospitals and medical professionals, such as the ability to prepare treatment facilities and appropriately allocate supplies and vaccinations.

A similar disease forecasting effort in fall 2017 predicted malaria outbreaks in the Peruvian Amazon. NASA has partnered with university researchers who leverage NASA’s satellite fleet to identify areas where popular breeding grounds for the anopheles darlingi mosquito (the species most responsible for spreading malaria) overlap with concentrated human populations, leading to high infection rates. Using the Land Data Assimilation System (LDAS), NASA can pinpoint warm temperatures and calm waters like ponds or groundwater flooding—ideal conditions for darlingi to lay eggs. Regional models analyze this data and jump forward 12 weeks to predict where malaria is most likely to erupt. Health ministries are then encouraged to administer preventative treatment, bed nets, and other resources to specific health posts throughout Peru.

Disease forecasting remains an imperfect science, but as it is refined to a point of repeated, reliable accuracy, it will play a more significant role in containing and responding to dangerous disease outbreaks. 

Photo Credit: World Health Organization (WHO)

Related


Posted by Andrew Foerch