We begin Part IV, "Newer Types of Surveillance Data,'' with this chapter on methods for evaluating surveillance data. The other chapters in Part IV discuss the results of experiments using the methods described in this chapter.
The goal of data evaluation is to understand (1) the ability of data to contribute to the earlier detection of outbreaks (or of individuals with disease); (2) their availability; and (3) the cost to obtain the data. A successful evaluation will characterize these factors in a manner that they can be compared against each other to form an overall understanding of the cost and benefit of the data.
The questions that data evaluation seeks to understand about surveillance data include:
• Can the surveillance data contribute to the detection of disease or outbreaks (and which ones)?
• How much earlier can outbreaks or cases of disease be detected using the new data?
• What level of diagnostic precision of case or outbreak detection can the data support?
• How difficult are the data to obtain? At what cost?
• How available are the data for a region of interest?
The primary application for the methods described in this chapter is the study of new types of surveillance data. Examples of such data include chief complaints of patients, absenteeism rates, telephone calls to appointment lines and advice centers, 911 calls, and orders for laboratory tests. There are many types of data that may be of value in biosurveillance. These techniques form the basis for a research agenda whose goal is to screen them for possible inclusion in biosurveillance systems.
We note that the importance to biosurveillance of traditional surveillance data—reports of confirmed cases, results of diagnostic laboratory tests, and the results of routine environmental testing—is beyond question: these data have proven their value in the field.Thus, an evaluator would only be motivated to study their costs and benefits in order to compare them with newer types of data that might provide the same information, but perhaps be easier to obtain, at lower cost, or have the potential to improve the timeliness, sensitivity, or diagnostic precision of outbreak detection and characterization.
We begin the chapter with a discussion of the characteristics of surveillance data that an evaluation should elucidate-informational value, availability, and cost to obtain. We follow with a review of experimental techniques that evaluators have used in published research to measure informational value. Since evaluators use different experimental methods to measure the informational value of data for case detection and for outbreak detection, we discuss these methods in separate sections. The scope of this chapter is limited to methods appropriate to laboratory evaluations of surveillance data. We discuss methods for field testing of surveillance data in Chapter 37.
Was this article helpful?