One of the challenges facing developers and users of automated disease surveillance systems is being able to accurately evaluate the performance of their systems for the wide variety of public health threats that are possible. A variety of methods have been used in the past to create data sets for use in testing algorithm performance. Synthetic data has been created using agent-based simulations where data is created based on the hypothesized activity of individuals with contagious diseases. This data is only as accurate as the social models and variety of assumptions which must be made permit. Real data containing elevated levels of respiratory and gastrointestinal activity have been used to evaluate the ability of algorithms to detect the elevated levels. Routine unvalidated outbreaks are typically not public health emergencies and may not represent signals of interest. Another approach is to use real background data and inject a variety of different types of synthetic cases representing various types of outbreaks on top of that background.
With the introduction of the American Health Information Community (AHIC) Minimum Data Set (MDS), the public health surveillance community should have the potential to obtain greater specificity for alerts generated in automated systems. The introduction of these additional data elements increases the complexity of algorithms using linked data elements. Creating synthetic data sets that accurately estimate relationships among chief complaint, pharmacy, laboratory and radiology is an added complexity in creating synthetic outbreaks for performance evaluation.
Objective
The objectives of this presentation are to describe the need for synthetic data containing the elements of the AHIC MDS. Approaches for creating synthetic data with MDS data elements will be presented and methods for insuring maintenance of confidentiality will be discussed.