Statistical disease cluster surveillance of medically treated self-inflicted injuries in Alberta, Canada.

Routine surveillance of cases of disease can highlight geographic regions that need further study and intervention. Statistical disease cluster detection methods are one way to statistically assess the number of cases in administrative areas. Traditionally, disease cluster detection methods are used to monitor the incident cases of disease. We review a statistical cluster detection method that is applicable for regions with diverse administrative area population sizes. We apply the method to assess clustering of self-inflicted injury presentations to emergency departments in Alberta, Canada. Analyses focus on the pediatric population and are adjusted by the age and gender distributions of subregional health authorities. Fifteen clusters of self-inflicted injuries are identified and, based on age and gender distributions, the clusters are not likely chance occurrences. We believe that these clusters represent areas of excessive self-inflicted injury and that special intervention programs should be considered.