ABSTRACT
This chapter discusses analytical tools for prospective disease surveillance. It deals with the filtering of the data streams to capture as much of a presumed outbreak signature and as little of the background and systematic noise as possible and analytic methods to recognize and characterize an outbreak signal sooner and more clearly than possible without the analytics. The chapter presents basic concepts and some recent trends in statistical alerting methods. It focuses on monitoring single time series for statistical anomalies that might be signals of health events of interest. The chapter describes the most popular monitoring application of distributed surveillance data, the search for significant clusters of cases in space and time. It addresses concepts and issues derived from the use of SaTScan, many of them apply to other cluster detection methods as well. The chapter explores primary objectives for alerting methods and then addresses the component issues of calculating expectations, deriving test statistics, and setting thresholds for each objective.
