Differences between notifiable and administrative health information in the spatial–temporal surveillance of enteric infections

PURPOSE: The purpose of this study is to compare the spatial and temporal information generated from two distinct health data sources available for the surveillance of intestinal infections associated with Escherichia coli O157:H7. METHODS: Our study area is the province of Alberta, Canada. Data are from two sources: a fee-for-service administrative health data system and a notifiable disease data reporting system. The study period is between 1999 and 2005. We compare the systems by observing correlations in the infections over time, the variability in the overall distribution of cases (as measured by a geographic dissimilarity index), and the relative locations of spatial-temporal clusters of infection. RESULTS: Our results indicate considerable variability in information generated from these two systems. The geographic distribution of cases varies considerably, with annual indices of dissimilarity suggesting considerable variation in the geographic distribution of cases throughout the study period (D=0.445). The temporal patterns identified by these two sources of information are negatively correlated (-0.40, p<0.001). Notifiable disease clusters occur in the summer in southern regions of the province, whereas cases identified from administrative health data system cluster in the winter season, and further to the north. CONCLUSIONS: Notifiable disease data may suffer from selection bias; administrative health data may be insufficiently precise without laboratory confirmation. Our results illustrate differences in the spatial and temporal information generated from these two systems of case identification. Future surveillance of gastrointestinal illness of infectious origin may benefit from case ascertainment algorithms based on both sources of data.