Abstract for: The Ebola Outbreak in West Africa: Important Lessons about Modeling and Simulating Uncertain Dynamic Issues

During the first half of 2014, the Ebola outbreak in West Africa was severely underestimated. But during the second half of the year, many modelling studies showed catastrophic projections of cumulative Ebola cases and deaths. Recently, these modelling studies have been criticized for severely overestimating the outbreak. As a consequence, the usefulness of simulation models during outbreaks has even been questioned, even in Nature. This study exposes some of the causes for overestimation as well as for underestimation when using simulation models for current uncertain dynamic issues. Addressing some of these causes by calibrating more complex instead of less complex transmission models to more, or more recent, data is shown to reduce the Ebola projections from millions to tens of thousands of cases. This study also shows that the current outbreak was likely to be curbed by the current massive deployment and behavioural changes before accelerated vaccination campaigns can even be rolled out. It is shown that the quality of the model and results can be improved substantially, but also that some uncertainty cannot be reduced, and that communicating results under uncertainty to decision makers, the media, and other scientists remains difficult.