Evaluation of influenza surveillance system in Dongguan, Guangdong province

Objective To evaluate the performance of the influenza surveillance system in Dongguan. Methods The analysis was conducted on the surveillance data and specific survey Results to evaluate the representativeness, timeliness, acceptability, facticity, sensitivity and specificity of the surveillance system. Results The influenza surveillance system in Dongguang is composed of the sub-systems of influenza like illness (ILI) surveillance, etiological surveillance and outbreak surveillance, which covers the city general hospitals, township hospitals and private hospitals and performs well. The correlation coefficient was 0.683 between ILI proportion and reported outbreak (P=0.003) and 0.707 between ILI proportion and detection rate of influenza virus (P=0.000), which were consistent in indicating the incidence trend and seasonality of influenza in Dongguan. The definition of ILI was not fully understood by all physicians surveyed, and about half of them thought the surveillance would increase their work load, 45.45% of them failed to report ILI or report false ILI. The specificity, positive prediction value and negative prediction value were highest (92.42%,58.33% and 85.92% respectively), but the sensitivity was only 41.18% when evaluated by using the indicator of more than 10% increase of ILI proportion compared with previous week. Conclusion The representativeness, timeliness, facticity and usefulness of the surveillance system in Dongguan were good, but acceptability of the system should be improved and the sensitivity and specificity of the system should be further evaluated.