Evaluation of outbreak detection performance using multi-stream syndromic surveillance for influenza-like illness in rural Hubei Province, China: a temporal simulation model based on healthcare-seeking behaviors.

Background: Syndromic surveillance promotes the early detection of diseases outbreaks. Although syndromic surveillance has increased in developing countries, performance on outbreak detection, particularly in cases of multi-stream surveillance, has scarcely been evaluated in rural areas.

Objective: This study introduces a temporal simulation model based on healthcare-seeking behaviors to evaluate the performance of multi-stream syndromic surveillance for influenza-like illness.

Methods: Data were obtained in six towns of rural Hubei Province, China, from April 2012 to June 2013. A Susceptible-Exposed-Infectious-Recovered model generated 27 scenarios of simulated influenza A (H1N1) outbreaks, which were converted into corresponding simulated syndromic datasets through the healthcare-behaviors model. We then superimposed converted syndromic datasets onto the baselines obtained to create the testing datasets. Outbreak performance of single-stream surveillance of clinic visit, frequency of over the counter drug purchases, school absenteeism, and multi-stream surveillance of their combinations were evaluated using receiver operating characteristic curves and activity monitoring operation curves.

Results: In the six towns examined, clinic visit surveillance and school absenteeism surveillance exhibited superior performances of outbreak detection than over the counter drug purchase frequency surveillance; the performance of multi-stream surveillance was preferable to signal-stream surveillance, particularly at low specificity (Sp <90%).

Conclusions: The temporal simulation model based on healthcare-seeking behaviors offers an accessible method for evaluating the performance of multi-stream surveillance.