Title: Detection of outbreaks in notifiable disease data
Authors: Matias Salibian-Barrera - The University of British Columbia (Canada) [presenting]
Tae Yoon Lee - The University of British Columbia (Canada)
Abstract: In order to detect and control possible disease outbreaks, the British Columbia Centre for Disease Control (BC CDC) monitors approximately 60 reportable disease counts from 8 branch offices in 16 health service delivery areas. Instead of relying on the judgement of staff on whether the reported numbers are higher than expected, in the early 2000s the BC CDC commissioned an automated statistical method to detect outbreaks in notifiable disease counts. To accommodate the seasonality of some diseases like meningococcal infections, longer term trends such as the periodicity of pertussis cases, and the decline in diseases like acute hepatitis B, this method is based on a generalized partially linear additive model. However, the current approach relies on certain ad-hoc criteria, and the BC CDC is interested in considering other alternatives. We discuss an outbreak detection method based on robust estimators of the distribution of the number of cases of each disease. The proposal builds on recently proposed robust estimators for additive, generalized additive, and generalized linear models. Using real and simulated data we compare our method with the approach currently used by the BC CDC and other natural competitors.