38278 Evaluating Engagement of Cdc's Social Media Messages in Response to Hurricane Matthew

Sabrina Riera, MPH, EHHE, CDC, Atlanta, GA

Background:  On October 6, 2016, CDC activated the EOC in response to Hurricane Matthew. As a part of the response, CDC delivered health information to the public via CDC’s English and Spanish social media channels (i.e. Facebook and Twitter). The messages were delivered using a phase-based approach, providing targeted messages before, during, and after the storm. Messages focused primarily on hurricane safety and preparedness, with an emphasis on flooding and carbon monoxide poisoning. To assess overall engagement, likes and shares were tracked and analyzed for each post.

Program background:  CDC uses a number of social media channels to quickly disseminate health safety and preparedness based information during natural disasters. For Hurricane Matthew, the Joint Information Center (JIC) used the following social media channels:

  • @CDCEnvironment Twitter account with 16,000 followers
  • @CDC Facebook page with over 600,000 followers
  • @CDCEspanol Twitter page with 39,000 followers
  • @CDCEspanol Facebook page with 75,000 followers 
Many of these tweets were shared by CDC’s Emergency Twitter account, @CDCEmergency (1.8M followers).
 

Evaluation Methods and Results:  Average likes and shares were tracked and analyzed for each post using Microsoft Excel.

  • A total of 54 messages were sent out through four social media channels:
    • @CDCEnvironment Twitter: 31 posts
    • @CDC Facebook: 7 posts
    • @CDCEspanol Twitter: 8 posts
    • @CDCEspanol Facebook: 8 posts
  • Messages posted on the @CDC Facebook page received the most attention with an average 133.14 likes and 70.71 shares per post.
  • Messages posted on @CDCEnvironment received, on average, 11.77 likes and 18.74 shares per post.
  •  Messages distributed via CDC’s Spanish social media channels received the least amount of attention, with an average of 5.8 likes and 2.38 shares per Facebook post and an average of 1.5 likes and 2.38 shares per Twitter post.

Conclusions:  While the majority of messages were distributed via CDC’s Environmental Health Twitter page, @CDCEnvironment, few messages were posted via CDC’s Facebook page. Messages posted on the @CDC Facebook page, however, gained the most attention with a much higher average of likes and shares per post than the @CDCEnvironment Twitter page. Despite having a relatively large number of followers, messages posted on the @CDCEspanol Twitter (more than 39,000 followers) and Facebook (more than 75,000 followers) pages had the fewest average likes and retweets per post.

Implications for research and/or practice:  The information gathered supports a continual effort to improve CDC’s social media practices. Given the amount of attention posts on Facebook received via @CDC, it would be beneficial to increase the amount of messages posted on this channel during emergencies. Considering the lack of engagement for posts on the @CDCEspanol pages, it would be helpful to further evaluate whether these are effective channels for disseminating information to Spanish speaking populations during emergencies. In future evaluations, we may want to consider tracking additional measures of engagement, including clicks and location. While clicks were not tracked during this response, they would provide additional engagement information for individuals who may have clicked on a link without liking or sharing the post. Location information would also be helpful in further evaluating how successful messages were at targeting particular regions.