In the event of outbreaks of infectious diseases, mathematical models can be used to inform decision makers regarding the likely spread of disease and the impact of control strategies. However, in the early stages of novel outbreaks, there can often be significant uncertainty regarding the spatiotemporal spread of disease and the likely impact of any intervention policy. However, policy makers do not often have the luxury to wait for any uncertainty to resolve before introducing an intervention, so it is crucial that models are developed to take account of the most up to date information available.

In this work, we analyse historical outbreaks of livestock and zoonotic infectious diseases and investigate the predictability of infectious disease models in the early stages of these epidemics to determine how any predictions change as more data are accrued. Our results indicate that the substantial epidemiological uncertainty at epidemic onset can lead to misleading forecasts of the impact of any intervention policy. However, robust predictions can be obtained after the first two to three weeks owing to a resolution of uncertainty during this period. We conclude that real time information is vital to ensure that policy makers can select the most appropriate intervention policy to minimise the impact of any ongoing epidemic.