Enabling SaTScan On the Cloud: Large Scale Spatial Analysis of Epidemics On-Demand

SaTScan is a compute intensive application that is commonly used to model the progression of epidemics. In practical applications ensemble runs of SaTScan are desirable to provide public health analysts with better insight of the epidemics’ progression. These ensemble runs require on-demand computational resources beyond those available in a typical health department. Cloud computing is a valuable approach to provide these resources. This presentation reports our successful implementation of a SaTScan Cloud using the Nimbus TP2.X software. To demonstrate its use we present the analysis of epidemic data from a high-fidelity, agent-based model of a pertussis infection simulated for a population of 2.2 million individuals in Utah. We established that SaTScan ensembles can be run in parallel on the cloud greatly decreasing the turnaround time. This presentation shows that using cloud computing it is possible to provision on demand resources for epidemic analysis using a compute intensive application like SaTScan. Therefore, implementing a cloud across the existing internal infrastructure of a health department may be a viable approach for large-scale epidemiology surveillance on demand.