How Did Consumers Respond to and Say about the 2009 Peanut Products Recall? Evidence From a Pilot Social Media Research

Theoretical Background and research questions/hypothesis: 

Public health agencies routinely initiate communication during public health crises to disseminate risk information to help the public cope with and adopt appropriate response to the crisis.  According to the CDC’s Program Evaluation Framework, credible and relevant evidence about a program and its effect can help enhance the utility and effectiveness of a communication program.  This presentation will report on a pilot study that uses social media data to examine consumer awareness of and reactions to the 2009 peanut products outbreak and the information that FDA and CDC disseminated during the outbreak.  The study was designed to (1) identify the sources of information mentioned/referenced in online conversations, (2) identify the prevalence and trends of discussion themes, (3) identify understanding of and reactions to risk information, and (4) assess sentiments and comments about the outbreak and parties involved in the outbreak.                

Methods:  A contractor used proprietary software and keyword searches to identify relevant online conversations and messages recorded at social media sites such as blogs, discussion forums, and Twitters and within the study timeframe.  The data were used to (1) calculate the volume and trend of conversations related to the outbreak, (2) identify the evolvement and trend of themes and consumer sentiment, (3) identify concepts and issues that were most closely related to the outbreak and how they changed over time, and (4) identify opinion leaders.

Results:  There was a significant amount of buzz about the outbreak during the study timeframe.  Conversation was more prevalent on forums than on Twitter or blogs.  Top forums driving the buzz include news, health-oriented and women’s online communities.  The outbreak was discussed among a wide range of audiences.  Food safety bloggers expressed a higher degree of negative sentiment.  The number and frequency of references to traditional media suggest the latter was a central source of information.  The FDA Recall Link, which listed products recalled, was widely referenced and the CDC YouTube channel saw a high level of buzz traffic.  The data also suggested consumer uncertainty about which products were being recalled.  Finally, consumer sentiment toward the FDA declined during the period of the outbreak.   

Conclusions:  The study provides the FDA a chronology of social media users’ awareness of and reaction to the outbreak.   The information can be used to evaluate the reach and influence of the risk information that the FDA disseminated via both traditional and social media channels.  It also can be used to help prompt identification of consumer misinformation and negative sentiment and develop strategies to counter these effects.  The FDA plans to replicate the study in the future to establish an evaluation mechanism by harnessing the power of social media to refine its health risk information campaigns during foodborne illness outbreaks or recalls.

Implications for research and/or practice:  The methodology used in this study can be applied to a variety of health risk communication programs.  The speed and coverage of social media feedback will enable risk communicators to keep a real-time and constant pulse of audience’s reactions to messages and to address any issues that may reduce the intended effects of the messages.