A SENTIMENT ANALYSIS OF MERS-CoV OUTBREAK THROUGH TWITTER SOCIAL MEDIA MONITORING

Background: The latest outbreak of Middle East respiratory syndrome coronavirus (MERS-CoV) in South Korea in 2015 has drawn attention from across the world. How MERS impacts the general public’s sentiment remains unknown. This is the first study utilizing social media data to monitor MERS. Methods: A total of n = 6,722,700 tweets with search terms “MERS” and 32 other popular health topics from June 8 to July 25, 2015 were analyzed. We assessed trends in MERS sentiment scores using a piecewise regression model. The sentiment scores from MERS and those from other popular health topics were compared by t-tests. Results: We identified a significant increasing trend in MERS sentiment scores. The results from the change point modeling suggest that MERS sentiment scores had stabilized and the panic arising from the MERS outbreak had stopped after July 14, 2015 (P = .002). The average MERS sentiment score was significantly lower than those of all 32 other popular health topics (P < .0001) in June 2015, though the gap narrowed in July 2015. Conclusion: Our study suggests that Social Media Monitoring (SMM) could serve as a supplementary tool for traditional infectious disease surveillance and that sentiment scores could reflect real-time public opinions and emotions related to MERS. Received: May 28, 2018 Accepted: July 16, 2018