Forecasting Ebola

Efforts to build highly detailed and interactive computer simulations of world populations appear to be bearing fruit in the fight against the spread of Ebola. The Network Dynamics & Simulation Science Laboratory (NDSSL), part of the Virginia Bioinformatics Institute, located on the campus of Virginia Tech in Blacksburg, Va., has been working with the Department of Defense to forecast the extent of the Ebola epidemic in Africa by using computational epidemiology.

“The models combine data on the growth of the outbreak with Census records, clinical data about the disease, and other contextual information to predict the number of new cases in the coming months,” wrote Caitlin Rivers, a Ph.D. student with the NDSSL, in a recent article on Nextgov.com. “We also use them to learn how best to respond to outbreaks.”

The computer simulations require huge amounts of computing power and data because they are highly detailed replications of human populations and the interactions of those populations. Researchers created hundreds of thousands of autonomous “agents” (which could be a single human being) in the model and program them with very simple interaction instructions. Researchers then insert agents infected with Ebola into these “synthetic populations” and observe the resulting large-scale system dynamics.

These detailed, synthetic, interactive models can help researchers understand the characteristics of an Ebola outbreak, particularly its spread, as well as identify vulnerable populations. Most importantly, researchers can use the models to forecast the extent of the epidemic, taking into account the potential impact of various interventions.

Earlier this year, as part of a research project sponsored by the National Institutes of Health, NDSSL scientists used earlier versions of the models to show that pharmaceutical interventions were less effective than supportive care and personal protective equipment for healthcare workers in slowing the spread of Ebola.

Unlike traditional epidemiology, which uses data from field investigations, surveys, and hospital records, computational epidemiology uses near-real time data from numerous online sources to help decision makers plan interventions. But getting the quality data needed to build reliable models quickly can be a challenge.

“The more information that can be included in the models about the epidemic to date, the better the predictions are about what is coming next, which in turn makes the results more useful to decision makers,” said Rivers. “The tighter the loop between data collection, model building, and public health response, the better.”

Cartographers without borders

Aid organizations working in Ebola-stricken West Africa need more than doctors and medical equipment. They need maps that are more accurate than the Google Maps of the area, and they need them quickly. OpenStreetMap (OSM), a combination of crowdsourcing and open source mapping software, is helping humanitarian teams create detailed, interactive digital maps within days.

“Google’s business model is selling advertising, so it’s simply the business case,” said Andrew Buck, a volunteer coordinator with the Humanitarian OpenStreetMap Team (HOT) in the online magazine Fast Company. “Starbucks isn’t paying for Google to advertise over there so there’s very little incentive for Google to improve its maps.”

The HOT team quickly brings together OSM volunteers on the ground with GPS devices and remote editors to create detailed and freely available maps for organizations such as Doctors Without Borders and the Red Cross.

“People feel like they are having a direct impact when they see that a group like Doctors Without Borders needs their work,” said Buck. —M.C.