The Application Of Syndromic Data In Influenza Surveillance Of JiangSu Province

Objective:(1) To understand the epidemiologic characters of influenza in Jiangsu Province and to evaluate the influenza surveillance measures available by analyzing the data collected during 2004 to 2007. (2) To make sense of the application of ARIMA algorithm in predicting the number of influenza-like illness(ILI) in sentinel hospitals of Jiangsu Province. (3) To make clear of practice applyment of the variance syndromic data in influenza surveillance . (4) To confirm the possibly existent of human influenza A(H5N1) cases that present no or slight influenza symptom in the high risk population.Methods:(1)To analyze data collected from epidemiological, pathogenic and outbreak surveillance from January,2004 to December,2007. (2) Using the ARIMA algorithm of SPSS to analyze the weekly ILI data aged from 5 to 14. (3) To summarize the result of influenza/human avian influenza synthetical surveillance system surrounding Gaoyou Lake;cross-correlation analysis of time series was done to explore the relationship between different data sources.Results:(1) According to the surveillance of ILI in sentinel hospitals, it showed two peaks distributed in winter-spring and summer in Jiangsu Province. The time of ILI season peak are different between ILI aged below 5 and above 15. The ILI rose above historical level may suggest influenza outbreak, but a high sensitive and specity thredhold value has not been found yet. Most of outbreaks occurred in primary and middle schools, and frequently in winter and spring. The monthly percentage of outpatient visits due to ILI aged below 5 could indicate the outbreak much sensitively than The monthly percentage of outpatient visits due to ILI. The construction of subtype strains usually changed around September, and take turns to be the annually outstanding subtype. (2) The mold of ARIMA(0,1,1) could largely predict the weekly ILI number aged from 5 to 14. (3) School absenteeism surveillance could early detect disease outbreak, the relative relationship between antibiotics sales curves and ILI curves among 7 weeks duration has no statistical significance; curve of common cold pharmaceutical sales fits best with the lag 1 week's ILI curve. Human influenza A(H5N1) case is not found in the high risk population . Few influenza A virus were isolated from the healthy poultry in the area of surrounding Gaoyou Lake,and the isolated virus occurred frequently in winter-spring season.Conclusion: Although sentinel hospital-based surveillance system has provide scientific evidence for making the preventive and control strategies of influenza, its ILI surveillance might not be effective to detect outbreak. Given that,it is necessary to explore more sensitive and synthetical surveillance measures in practice. The performance of influenza/human avian influenza synthetical surveillance system surrounding Gaoyou Lake has provided practical evidences for the application of syndromic surveillance in outbreak early detection, and it still should be perfected.