ABSTRACT
Traditional public health surveillance was originally designed as a passive system whereby cases of disease were reported individually as they arose, usually by primary care physicians. In the case of infectious diseases, it is usually important to consider mechanisms of transmission in the formulation. An attempt to use such methods has been made by C. Vidal-Rodeiro and Andrew B. Lawson in a spatio-temporal disease mapping context and considerable time savings were apparent compared to refitting using Markov chain Monte Carlo. While temporal components are more controversial it is also the case that special consideration of the background spatial components may be needed. Bayesian hidden Markov models have also been proposed for temporal surveillance data by Y. L. Strat and F. Carrat, but they only apply the approach to retrospective analysis. A review stresses the range of definitions available but also stresses the usefulness of syndromic surveillance within the greater public health community.
