Real-time modelling for forecasts during infectious disease outbreaks

Local and international public health bodies often face difficult decisions during infectious disease outbreaks. Limits on human resources or logistics, for example in the speed and scale of vaccine production and delivery, require prioritisation of areas or population groups when rolling out an intervention. It is now possible to collect a wealth of epidemiological, genetic, spatial and behavioural data during an outbreak and make it available immediately for analysis. It remains an open question how these data are best combined for forecasts that can inform decision making. My fellowship aims to systematically investigate and improve the predictive capabilities of mathematical models during outbreaks. In particular, I will develop methods to combine different data sources in order to analyse an outbreak in real time. By testing these methods on a range of recent epidemic datasets, I will assess the predictive capabilities of models using different amounts and types of data. A particular focus will be on translating forecasts into recommendations for decisions, and on how the accuracy of forecasts affects the quality of these decisions. All the work will be conducted with the aim of generating a computational platform that is readily deployable in an outbreak situation for meaningful decision support.