|
Abstract:
|
The performance of surveillance systems depends on factors including true prevalence in the population at risk, sensitivity/specificity, or sensitivity/predictive value positive. In systems with a long history of operations, changes in both the system and in the underlying population at risk can occur over time, and these changes may affect estimates, performance, and interpretation/inference. Generally, there are many changes in clinical care that can affect system performance. For example, the emergence of bedside echocardiography in newborns or non-invasive chromosomal and other types of screening may affect system prevalence estimates of cardiac septal defects or trisomies. System performance may change when data sources change, or when the mix of active and passive case acquisition changes. Also, changes over time to the catchment area, specifically geographic boundary or make-up of the population at risk, may affect a system’s prevalence estimates. Both hypothetical examples and those from birth defects surveillance systems will be presented along with implications and coping strategies.
|
