If an evaluator does not have access to time series surveillance data collected during an outbreak, there are still methods that he can use to study the informational value of surveillance data for outbreak detection. Linked record analysis requires, as a prerequisite, only a set of patients with the disease of interest. The name "linked record'' refers to the fact that the patients of interest are identified by links between records in the surveillance data and other data that are a gold standard for the presence of disease in the individual.
With this method, an evaluator can measure the fraction of individuals with a disease that appear in the surveillance data. An evaluator can also measure the average time that an individual is encountered in the surveillance data, relative to the onset of disease, or some other reference data.
Examples of linked record analysis can be found in Henry et al. (2004), Chapman et al. (2003a), Chapman et al. (2004a), Chapman et al. (2005), and Gesteland et al. (2005).
Was this article helpful?