Forecasting potential malaria epidemics

Scientists led by Madeleine Thomson of Columbia University in New York have used the knowledge that mosquitoes need pools of standing water to lay their eggs to find a better way to give early warning of a malaria epidemic. As a result, they have increased the forecast time from a matter of days or weeks to five months, says the Economist.

The new system should:

Help people in countries such as Botswana, South Africa and Swaziland, which even now have the capacity to detect and respond to a malaria epidemic within two weeks.

Prove invaluable to those living in worse organised places such as Zimbabwe, which cannot respond so quickly.

The new early-warning system integrates many different climate-forecasting computer models that couple together information about how the atmosphere and the oceans combine to create the weather, explains the Economist.

The scientists tested it using old data for Botswana dating from 1982 to 2002 and found a huge improvement.

Its predictions of an impending malaria epidemic would be available at the beginning of November, while other models warn only in early March, just a month before an epidemic typically starts.

The advanced warning comes at a price, however. The earlier the prediction is made, the less confidence the scientists have of its accuracy. Indeed, the system would appear to be slightly less accurate than other models, says the Economist.

The gain in lead-time should nevertheless provide governments and non-governmental organisations with the opportunity to plan for a bad season. Bed nets treated with insecticide cost about $5 a piece and, WHO estimates, can cut death rates by 20 per cent, says the Economist.

Source: Casting Shadows, Economist, February 2, 2006.

For text (subscription required): http://www.economist.com/science/displaystory.cfm?story_id=5466137

For more on Health: http://www.ncpa.org/iss/hea/

FMF Policy Bulletin/ 14 March 2006