Objective: To reveal the characteristics and stability of the system through the analyzing the surveillance data of respiratory-feverous syndrome via the syndromic surveillance system which was established during the Shanghai World Expo in Pudong New District and provide references for the development and operation optimization on this Mass Gatherings Surveillance Systems.
Methods: Data used was from the surveillance data of respiratory-feverous syndrome collected from Pudong New District Syndromic Surveillance System, through May 1 to October 31, 2010. On the basis of description of data characteristics, correlation analyses were conducted, when compared to the surveillance data of respiratory-feverous syndrome and Pudong influenza-like illness (ILI) used as reference. Comparison of variances on the surveillance data and the report lag time of the earlier and later surveillance periods were also carried out to evaluate the quality and stability of data.
Results: Reports on the respiratory-feverous syndrome showed a peak in late September with day-of-week effects and holiday effects. Correlation between respiratory-feverous syndrome and ILI was the strongest in the same day (r = 0.596, P < 0.05). In the earlier surveillance period from 2010-05-01 to 2010-07-31, the correlation between respiratory-feverous syndrome and ILI was not obvious (r = -0.058, P > 0.05); however, the two-time series showed consistent trend with the correlation coefficient as 0.798 (P < 0.05), in the later period from 2010-08-01 to 2010-10-31. In addition, variability of the surveillance data on respiratory-feverous syndrome was less in the later period than in the earlier one, with quality of the report on related data better in the later period. Analyses on the correlations of reference sequence, variability and quality of report indicated that the stability of the later surveillance period was better than the earlier one.
Conclusion: Only with the operation of syndromic surveillance system for a certain period of time, could data in the system maintain stability. Surveillance data showed both day-of-week effects and holiday effects, suggesting that there was a need to choose early warning models with short baseline data.