Sensitivity, specificity and predictive values of health service based indicators for the surveillance of influenza A epidemics

Abstract

Background: The Regional Influenza Surveillance Group (GROG) is a French surveillance network set up in 1984. It collects virological specimens and health service based indicators on a weekly basis. Our aim was to assess the predictive value of the health service based indicators for the detection of influenza A epidemics.

Methods: Virological data were used as a gold standard for defining the epidemics. For each health service based indicator, a statistical threshold was used as a test for the identification of epidemic weeks. Finally, an epidemiological criterion was defined in order to improve the specificity and the speed of detection of outbreaks.

Results: Health service based indicators have a positive predictive value of around 0.80. They also advance the detection of outbreaks by between 1 and 4 weeks.

Conclusions: These indicators are easy to collect and are useful for the surveillance of influenza epidemics. Such a system is the prerequisite for the rational use of preventive tools.