CDC’s BioSense system provides near-real time situational awareness for public health monitoring through analysis of electronic health data. Determination of anomalous spatial and temporal disease clusters is a crucial part of the daily disease monitoring task. Spatial approaches depend strongly on having reliable estimated values for counts among the geographic sub-regions. If estimates are poor, algorithms will find irrelevant clusters, and clusters of importance may be missed. While many studies have focused on improved computation time and more general cluster shapes, our effort focused on finding anomalies that are correct according to available BioSense data history.
Objective
We applied spatial scan statistics to data from CDC’s BioSense system and examined the effect of the spatial prediction method on determination of anomalous disease clusters. The objectives were to decide on a reliable spatial estimation method for one BioSense data source and to establish criteria for making this decision using other sources.