The spatial aspect of Swedish seasonal influenza data was investigated and modelled with the main aim of finding patterns that could be useful for outbreak detection, i.e. for detecting an increase in incidence as soon as possible.
Quality problems with data on laboratory-diagnosed cases (LDI) collected by a number of laboratories and other data were studied. Parametric and nonparametric regression methods were used for estimation of the excepted incidence. Multivariate analysis was used to determine the impact of different spatial components.
Quality problems were found for all types of data. LDI was found useful for the present aim. No evidence for a geographical pattern was found. It was found that the influenza outbreak started at about the same time in the metropolitan areas and about 1 week later in the rest of the country. Both parametric and nonparametric regression models are suggested.
There was a time difference between the outbreaks in the metropolitan areas and the rest of the country. This can be utilised to improve outbreak detection.