Study on early warning based on influenza surveillance data in Guangdong province

ObjectiveTo build the earlywarning systemof Guangdong province based on influenza surveillance data.MethodsAccording to theweekly influenza-like illness consultation rate(ILI%)of fourteen sentinel hospitals in Guangdong province from 2004 to 2006, the surveillance data of 2007 were fitted into models for early warning of inflecenza with threemethods including control chartmethod, historical linit method and exponentially weighted moving average(EWMA) method.ResultsThe control chartmethod showed that the influenza activity began to increase atweek 14 in 2007 and lasted twenty weeks The result of historical linit method showed that the influenza activity had no significant increase in 2007 compared with the data of sinilarperiods in recent years.The EWMA method showed that significant increase of influenza incidence occurred at week 24 compared with forecast values based on the recent trend.ConclusionThe sinulative early warning signal in 2007 is approxinately consistent with the actual influenza incidence, which means that the method is scientific and reliable.