Using multi-source web data for epidemic surveillance: A case study of the 2009 Influenza A (H1N1) pandemic in Beijing

Abstract

Timely and effective surveillance is critical for the prevention and control of epidemics. However, due to technical challenges and shortage of human resources, comprehensive and timely data collection required for effective surveillance, especially collection of data about sudden epidemic outbreaks, is still very difficult. In this paper, we propose the use of multi-source web data for epidemic surveillance. We use the 2009 Influenza A (HtNt) pandemic in Beijing as a case study to demonstrate the utility of our proposed approach. Experiments using data from the Beijing Center for Disease Control and Prevention (CDq and several search engines show encouraging results. This case study also has direct practical values in the real setting.

Early detection of an epidemic erythromelalgia outbreak using Baidu search data

“…Despite the significance of this, there are no principles or standards for guidance 18 26 27 . Previous studies generally chose the names or clinical symptoms of target diseases as their core keywords 22 23 25 26 . As EM is a little-known disease within the lay Chinese community, insufficient search volume of this word leads to Baidu’s failure in calculating its search information.…”

“…At the same time, an increasing number of people, including patients and their family members, are inclined to search online for health information before seeking medical service 16 17 , making it possible to monitor the health status of the population by tracking changes in frequencies of specific search keywords. Internet search engines are now the most common tool to obtain information for Internet users 18 19 , and data from different search engines have been successfully utilized for early detection of diseases such as influenza and dengue 17 18 20 21 22 23 24 25 26 . Such studies suggest that Internet search data-base surveillance might be a novel way to monitor epidemic EM outbreaks in near real-time.…”

A Data-driven inference algorithm for epidemic pathways using surveillance reports in 2009 outbreak of influenza A (H1N1)

“…), which have been shown closely related to the spreading of epidemics [30], [7], [9], [31], [32], [33], [34], [35], [36], [37]. One can readily incorporate these population-mobility-related factors into the extended agent-based simulations to modify the inference results of the invasion trees and epidemic pathways.…”

Discovering seasonal patterns of smoking behavior using online search information

“…Web search queries have been used to monitor infectious and noninfectious diseases such as influenza epidemics [10][11][12] and other diseases [13,14]. Tobacco control researchers have used Internet search queries to explore issues such as tracking the popularity of electronic cigarettes and monitoring tax avoidance and smoking cession after cigarette tax increases [15,16].…”

“…Searching for the term "smoking" and " 吸 烟 " means that the user has the need for smoking information. Based on the previous studies [10][11][12][13][14][15], we assume that the trend of smoking behavior in real life can be reflected by the trend of the corresponding searching behavior on the Internet. Then we applied the periodogram method to detect the seasonality, and calculated the pairwise cross-correlations among these countries.…”