Model for Malaria Tracks Impact of Simulated Interventions

Fellows Spring 2011

By David Taylor

October 10, 2011

Also published in Microbe magazine

Malaria Tools, a model for tracking this parasitic disease, is robust enough to predict outcomes of interventions now being contemplated, according to its developers, Azra Ghani and colleagues at Imperial College in London, England. It simulates various means for controlling malaria against the dynamics of transmitting the disease, while taking into account important factors such as climate and public health interventions. Users can evaluate, for example, the extent of mosquito net use, access to standard treatments, and how much pesticide is being sprayed in a particular geographic sector. The model simulates how those strategies fare against the population dynamics of the malarial parasite, Plasmodium falciparum, and the female Anopheles mosquitoes that deliver it to humans.

Malaria Tools can also help in evaluating the effectiveness of a malaria vaccine, such as one now in Phase 3 trials, according to John Marshall of Imperial College. When the model runs with the (wildly optimistic) assumption of 90% vaccination coverage across Africa, it shows the area of malaria sufferers steadily decreasing over 20 years. Thus, it predicts that the vaccine virtually vanquishes the disease.

To produce such scenarios, the Malaria Tools program relies on data summarizing the population dynamics of P. falciparum at six locations in Africa, across a range of seasonal patterns and infection intensities. Vectors are accounted for by species, according to Jamie Griffin, also of Imperial College, since the behavior of any one mosquito species affects the impact of control measures at each of those sites. For example, some Anopheles species bite indoors and rest on interior walls after feeding. The model accounts for people as individuals to reflect, for instance, differing immune responses to infections. Tracking such details bedeviled other modelers, he says.

The complex interactions between parasites and hosts partly account for why malaria is so difficult to control. Decades of simulating this disease with models failed to capture its shifting patterns. "There are individual factors, household factors, and geographic factors, but they're not as predictable as you'd think," says David Smith from the University of Florida.

Researchers belonging to malERA, a loose affiliation of modelers, recently wondered: "How can models and model systems ask key questions?" Their subsequent efforts yielded a blueprint for harmonizing several models, including Malaria Tools, according to malERA scientist Marcel Tanner, with the University of Basel in Switzerland. Under that plan, various teams will employ competing modeling strategies in parallel, rather than follow any single approach, while testing those models with hard data. "There's a new way to use models," says Tanner. "Modelers aren't just lunatics staring at a computer screen."

David A. Taylor, a freelance writer based in Washington, D.C., reported on malaria in West Africa with a fellowship from the International Reporting Project (IRP).