[Spatial-statistical analysis of infectious disease notification data in lower saxony].

BACKGROUND

In Lower Saxony the analysis of notification data according to the infection protection law is based on two pillars: one pillar is the weekly analysis to identify clusters and locally non-detectable outbreaks by an early warning system. The second pillar is an annual reporting system for obtaining an overall picture and for identifying long-term trends. In both fields methods of spatial statistics are applied.

METHODS

For the 16 most frequent notifiable diseases, the notification data of the districts of Lower Saxony are analysed and presented on the Internet. The presentation starts with an overview page, on which differences between observed and expected case counts for these diseases are presented tabulated and graphically. On a separate page for each disease the temporal and spatial distribution is shown by means of the following graphs: Time series presentation (weekly case counts and moving averages), chloropleth map of regional incidence rates and dot map of regional case counts. For identification of outbreaks, Kulldorff's "scan statistic" is used to search the data for clusters and to assess the significance of these clusters. The methods are executed mainly automatically using predominantly public domain software.

RESULTS AND DISCUSSION

The approach attempts to meet the surveillance principle of providing the participants with immediate feedback. The scan statistics is a universal objective instrument for detecting and assessing spatial clusters. A special advantage is the identification of clusters which include several districts. The properties of the identified clusters (spatial extension, number of cases) offer some insight into the characteristic epidemiology of the diseases.