Main Menu Browse by Day or Program Author Keywords

537 The Feasibility of Using Internet Search Data for Monitoring Infection Trends of Dengue Fever and H1N1 Influenza

Sunday, April 3, 2011
Trinity Ballroom (Hilton Anatole)
Chin-mei Ke, MS , Infection Control Unit, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan
Yusen Eason Lin, PhD, MBA , National Kaohsiung Normal University, Kaohsiung, Taiwan
Yao-shen Chen, MD , Infectious Diseases Section, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan
Background: An early warning system, proposed by the World Health Organization in 2007, is critical to respond to the H1N1 influenza pandemic.  Literatures suggest that keyword search is a gateway of information acquisition, and the volume of the keyword search counts has high correlation to the interest of such keyword. Currently the US Center for Disease Control and Prevention (CDC) has investigated the feasibility of using an internet tool, namely Google Flu Trend, to predict the influenza epidemic in major US cities. 

Objective: To use Google Insight Search (Google IS) for outbreak prediction in dengue fever and H1N1 influenza in Taiwan

Methods: We use Taiwan CDC’s case reporting system as our reference data to test the hypothesis that Google IS data can be a predictive tool for future outbreaks.  Pearson correlation and simple linear regression were used for analysis.

Results: Our results showed that Google IS search volume and search perspective of dengue fever (2006: r＝0.96, r2=0.91; 2007 r＝0.90, r2=0.80; 2008 r＝0.85, r2=0.72) and H1N1 influenza (2009: r＝0.79 , r2=0.63; Taipei City: r＝0.77; Taichung City: r＝0.71 ) have significant correlations between the Google Insight Search perspective and Taiwan CDC reporting data.

Conclusions: The results of this study can be used to monitor Google search perspective in real time for prediction of disease outbreak of dengue fever and H1N1 influenza, and to understand what the most people concerned about in public health and epidemiology. It should be noted that the internet accessibility could affect the correlation of the prediction.  Since Taiwan CDC has emphasized its public health education advocacy network in information age, we recommend this application to be experimented by the government health authorities for infectious diseases perspective as an early warning system for other epidemics.  More correlation analysis is needed to build a robust prediction system.