Rift Valley fever is a viral disease of animals and humans and a global public health concern due to its ecological plasticity, ability to adapt, and potential for spread to countries with a temperate climate. Outbreaks of Rift Valley fever in Egypt are sporadic. A new paper by John Drake and colleagues John Beier (University of Miami) and Ali Hassan (Ain Shams University) in the Journal of Vector Ecology reports on an attempt to develop a statistical model for forecasting Rift Valley fever outbreaks in Egypt. The effort failed to develop a method for forecasting when outbreaks might start — this is probably due to stochastic factors not well captured by data that has been collected historically — although the method was better than random at predicting the termination of outbreaks. Key predictive variables included the interaction between mosquito abundance and recent history of RVF activity, monthly rainfall, and river discharge, linking environmental and entomological conditions to public health. This study underscores the need for machine learning methods that perform well with small to medium amounts of data.

• Drake, J.M., A.N. Hassan & J.C. Beier. 2013. A statistical model of Rift Valley fever activity in Egypt. Journal of Vector Ecology 38:251-259. [online]

Supported by a grant from the National Institutes of Health