A prerequisite for most studies of the informational value of surveillance data for outbreak detection is an adequate sample of surveillance data. This prerequisite can be difficult to satisfy. The surveillance data must be from an outbreak.The surveillance data should be a nearly complete sample from the outbreak region, otherwise a negative result from the study will be difficult to interpret. It will either indicate that the data are of no value or that the sampling may not have been sufficient.
To obtain an adequate sample, an evaluator may have to obtain data from multiple hospitals, physicians' offices, veterinarians, laboratories, or retail stores. Occasionally, it is possible to find a data aggregator, such as Information
Resources, Inc. or ACNielsen, or a health department that can provide a suitable data set. Note that, unlike evaluations of detection algorithms, there is no role for synthetic data in evaluation of surveillance data; the properties of real data are of primary interest.
Was this article helpful?