Abstract
In Japan, as part of surveillance for seasonal influenza, the number of patients per influenza sentinel site is counted on a weekly basis. Currently, reference values are set for the weekly reported number of influenza cases per sentinel, and pre-epidemic and epidemic warnings are issued based on these values. In this study, we examined the association between these reference values and the effective reproduction number (R(t)) using surveillance data for Miyazaki Prefecture collected from 2010 to 2011. There are nine public health centre jurisdictions in this prefecture, and R(t) exceeded 1.0 at the time when pre-epidemic warnings were issued in almost all the jurisdictions. Thus, it was indicated that the validity of the reference value was also high for influenza transmission. However, our results indicated the presence of secondary epidemic caused by infections originating both from other jurisdictions and inner jurisdictions, and it is occasionally not possible to evaluate the end of an epidemic in a jurisdiction using only the reference value of termination. It is necessary to establish new methods after considering the situation in the surrounding jurisdictions for more detailed epidemic predictions.