Fear on the networks: analyzing the 2014 Ebola outbreak/El temor en las redes: Analisis del brote de ebola del 2014/Temor nas redes: uma analise do surto de Ebola de 2014.
The level of penetration, use, and popularity of information and communication technologies among the world's population creates agile bridges of information on a diverse range of events that are of global concern or of particular interest. Disasters, movies, economic debacles, sporting events, political meetings, and even music videos come from and go to all possible corners. Once there is an event of interest, networks are flooded with comments, memes (a viral digital content that acts as cultural and social symbol and idea), photos, blogs, and videos. These posts can be "liked" and shared millions of times, sometimes changing their use, and therefore, their meaning.
All this engagement does not imply that the content to which people react is of good quality. Indeed, separating what is true from what is false is an arduous task, one that the general public may not always be able to perform accurately. Within public health, there are topics of interest whose social media activity is generally stable over time. There is another set of topics that has a more defined time effect; some arise and stabilize quickly, such as those associated with natural disasters; while others develop slowly, and seem to grow exponentially, such as those associated with major epidemics. Polgreen and colleagues have asserted that the Internet has "dramatically changed how people search for medical information ... especially about infectious diseases ... Thus, the frequency of Internet searches may provide information regarding infectious disease activity (5)."
Although there are and always will be, situations that cause fear in society, this report shows that fearful reactions to fears or in individuals normally occur when the situation might affect them personally. In public health emergencies, such as the 2014 Ebola outbreak, it is an enormous challenge to communicate uncertainty without igniting fear and undermining public trust in health authorities (6). Figure 1 compares the term "Ebola" (a term associated with fear) to "soccer" and "futbol" (associated with pleasure), and shows that only in two situations did fear outnumber pleasure. On 2 August 2014, a United States missionary physician infected with Ebola in Liberia was flown to Atlanta, Georgia (United States) for treatment. On 8 October 2014, Duncan, the first person diagnosed with Ebola in the United States, died in a hospital in Dallas, Texas. The Government of the United States ordered that passengers arriving from West Africa at five major airports be screened for fever.
This study aims to support governments, mass media, and health-related institutions in developing strategies that explicitly seek to generate effective communication campaigns with a special focus on online tools, such as social networks and virtual libraries. It also seeks to reduce the know-do gap on the use of information and communication technologies for the prediction of social behavior (7, 8).
EBOLA AND SOCIAL MEDIA
Chronology of key events
A chronology of 63 key events from March-November 2014 associated with the outbreak of Ebola was developed based on data published by Reuters (9). Among these key events were the first reported cases, the infected health workers returning to their home countries, first treatments and potential vaccines, news from known leaders, and so on.
The sources of information chosen to gather data for this study were: Google (Mountain View, California, United States); Wikipedia (San Francisco, California, United States); Facebook (Menlo Park, California, United States) and Twitter (San Francisco, California, United States); and such news, blogs, and analytical tools as Symplur (Riviera Beach, Florida, United States), Topsy (San Francisco, California, United States), and Hashtags.org (Chicago, Illinois, United States). Public tweets and re-tweets (RTs) and Facebook posts that included the hashtag "#ebola" were collected; their distribution was determined for the study period and integrated with the chronology. The most mentioned Twitter profiles were determined, those using #ebola the most, and those with the highest number of impressions. Major peaks were determined, and transcriptions of the tweets around these peaks were analyzed to find the most common topics, as well as the most common words. RTs and content with non-related words (identified as noise) were not considered.
A total of 6 422 170 tweets and 142 Facebook posts were retrieved and organized in several Microsoft Excel[TM] (Microsoft Corp., Redmond, Washington, United States) spreadsheets. The chronological lists of events was integrated with the chronology of tweets and Facebook posts.
Twitter. All tweets were grouped into weekly time ranges, from Sunday to Sunday. Through the same platform, the following information was determined: total number of impressions, total number of tweets, total number of participants, average tweets per hour, and average tweets per participant. The number of tweets and their timeline were charted in order to identify areas for further analysis, as follows: (a) first area, 424 tweets from 16 March-24 July 2014; (b) second area, 1 863 764 tweets from 24 July-29 September 2014; and (c) third area, 4 555 982 tweets from 29 September -2 November 2014. The accounts with the highest influence during those periods were determined by the number of mentions. The chronology is integrated with the timeline of tweets in one chart and for each area. Every number represents one specific event in the chronology (Figure 2).
Facebook. Using the Facebook search function, posts that included the hashtag #ebola were identified. The search capabilities did not allow a time range to be specified; so, the results of the search conducted on 10 November 2014 at 12:57 p.m. were scrolled through to collect all posts from 16 September-10 November 2014.
Search trends. Google Trends was used to determine trends in relation to other topics, such as dengue and chikungunya (Figure 3). Similarly, the number of results in other sources, such as SlideShare (LinkedIn Corp., Mountain View, California, United States), and LinkedIn were compiled. The Ebola trend was also referenced with the chronology.
By preliminary analysis, the events that prompted an increase in the number of tweets using #ebola were those related to new cases of infection or the entry of the disease into a new geographic area. Most tweets were RTs of information provided by news agencies and official health organizations. News agencies significantly exceeded health organizations in terms of RTs, and thereby, influence. The most common language was English, as expected since most Twitter users predominantly reside in English-speaking areas. The only time a language (Spanish) surpassed English in Twitter usage was when the first case of Ebola occurred in Europe (in Madrid, Spain).
The number of tweets from official health organizations containing facts and recommendations decreased over time, but there was an increase in information about official reports on the number of cases and deaths. Finally, it was evident that events related to new infections and deaths seem to correlate with an increase in the use of words that express fear and worry. The use of words that users associated with symptoms of Ebola increased with time, but do not seem related to specific events (Table 1).
The trend on Facebook behaved somewhat differently from the trend on Twitter. Moderate peaks were found in October, with a surge of high peaks in November; however, no events were registered for November, at least not according to the constructed chronology. This incongruence must have been related to Facebook's search functionality, which displays a greater number of posts closer to the time when the search is performed. By eliminating the November posts from the timeline, the Facebook and Twitter trends and associated events were more closely aligned.
The results of Google Trends, expressed in relative measures from 0-100, were similar to those of Twitter and Facebook. The same events correlate to similar peaks. In comparison with other infectious diseases, Ebola notably surpassed chikungunya and dengue.
Table 2 summarizes the Ebola-resulted results for various other online services, such as YouTube (Alphabet Inc., San Bruno, California, United States), SlideShare, LinkedIn, other Google search applications, including Google web search, Google news search, Google search videos, Google search books, Twitter, and Hashtags.org.
Beyond simply accessing information, the population is now active in the creation of content: not only does news travel almost instantaneously from one side of the world to the other, but thanks to social networks, reactions also come in from its recipients.
The way media communicates and the way each person reacts to an event is different and conditioned by unique factors and the local circumstances of the place in which the event occurred. This complex interaction creates a challenge for health organizations that are striving to effectively and efficiently communicate with an affected population, whether it is to mobilize, prevent, reduce fear, or encourage. The need to listen in detail to the population before deciding how to communicate, what to communicate, through what media, and to what audience is imperative.
The 2014 West Africa Ebola epidemic was an event that initially developed at a slow pace, without great impact among people--only among health agencies. Eventually, it became a topic of massive interest; Ebola news was distributed through multiple channels and sources, simultaneously causing a social media response by the population, some seeking information to help them make certain decisions, some to make statements, and some to express dissent or concern (10). Some authors have asserted that,
... at the early stages of an outbreak, informal sources can be indicative, not just that an outbreak is occurring, but can highlight disease dynamics through estimation of a key epidemic parameter, the reproductive number. Social and news media, such as from HealthMap and Twitter are a cost-effective data source (11).
Data from social media can be used to draw hypotheses about how people and institutions behave in relation to public health events. In the case of epidemics, panic-related behavior in affected communities might be closely related to the type and frequency of words used; therefore, communication by health organizations should adapt dynamically and rapidly to the type of panic and the increased volume of data. Data can be used as a proxy for a major panic outbreak. An in-depth analysis of the impact of their social media efforts on the population should be an integral part of every health organization's communication strategies.
This analysis also showed that in many cases, news agencies have more engagement with the public than do health organizations; given this, some level of content surveillance by health organizations is needed. At the same time, social media data--unstructured data--requires sizable computational capacity to conduct proper analysis and to establish sound conclusions. This study faced a considerable methodological challenge, which was confronted by selecting small samples, but which led to difficulties with generalizing the conclusions.
Health institutions must continue developing communication and communication-based surveillance strategies that explicitly seek to generate more impact as influencers of information networks--especially in relation to news agencies--or, to at least insure that news agencies are using health institutions as important sources. Strategies must be applied to both ongoing and sudden public health issues, using shared information as a rich source for evaluating behavior expressed on social media. Communication and community outreach should be seen as vital components of an integrated public health response plan that is based on established science. When health communications initiatives are developed independently from other parts of the response plan, they may become marginalized and ineffective (12).
Notwithstanding, the unstructured nature of these data proves a challenge for fast and accurate analysis. For qualitative assessments, in particular, data needs to be in a readable and decipherable format. However, even in raw form, social media data can provide key insights into people's attitudes, which can in turn allow for fine-tuning of communication strategies and improved support for socio-epidemiological assessments.
This study considered only tweets and Facebook posts that included the hashtag "#ebola," which may have reduced the total possible number of tweets and posts of interest. Since the quantitative approach includes all tweets and re-tweets, it is likely that some repeat or the use of unrelated words may also have biased the conclusions. A strategy to avoid noise in content should be included in further analysis, particularly for the quantitative analysis. This noise avoidance is carried out only in the qualitative analysis. On the other hand, it is not possible to obtain all Facebook posts of interest through its search engine; and the results may have been filtered and/or customized by Facebook, which limited the real study scope for this social network.
Clearly, communication campaigns in the current context of the "Information Society" should have a special focus on online tools, social networks, and strategic content development. Health communication strategies should include creating and providing high quality content and also verifying content on websites such as Wikipedia, monitoring discussions and questions on LinkedIn, sharing all presentations related to diseases and health issues, such as Ebola on SlideShare, and improving web resources, including open libraries, mailing lists, among others.
Further research is needed to determine if unstructured data taken from social networks can be trusted to predict epidemics and how these nontraditional sources and tools can best be used for monitoring future epidemics and reducing fear among the population.
Facts and recommendations on how to manage disease, prevent new infections, and identify symptoms should be continuously delivered in order to raise awareness and keep users, other health organizations, and news agencies informed. The content of tweets should be written in a way that aims to control fear and paranoia among people. Tweets with valid and complete information and recommendations, rather than links to further information, are preferred. The most influential news and health information sources should be continuously reviewed for accuracy. Finally, health organizations must be in front of news agencies, developing specific and continuous Internet communication and strategies for knowledge dissemination in order to become the most influential source of public health information.
Conflict of interests: None declared.
Disclaimer. Authors hold sole responsibility for the views expressed in the manuscript, which may not necessarily reflect the opinion or policy of the RPSP/ PAJPH and/or PAHO.
(1.) Wang Y, Qingfei M, Shengnan H. Understanding the effects of trust and risk on individual behavior toward social media platforms: a meta-analysis of the empirical evidence. Comput Human Behav. Z2016;56:34-44.
(2.) Carlson J, Suzanne Z, Ranida B, Harris K, Harris J, Carlson DS. Social media use in the workplace. J Organ End User Comput. 2016;28(1):15-31.
(3.) Yousefi-Nooraie R, Dobbins M, Marin A, Hanneman R, Lohfeld L. The evolution of social networks through the implementation of evidence-informed decision-making interventions: a longitudinal analysis of three public health units in Canada. Implement Sci. 2015;10(1):1.
(4.) Boontarig W, Papasratorn B, Chutimaskul W. The unified model for acceptance and use of health information on online social networks: evidence from Thailand. Int J E-Health Med Comm. 2016;7(1) 31-47.
(5.) Polgreen PM, Chen Y, Pennock DM, Nelson FD, Weinstein RA. Using internet searches for influenza surveillance. Clin Infec Dis. 2008;47:1443-8. DOI: 10.1086/593098
(6.) Rosenbaum L. Communicating uncertainty--Ebola, public health, and the scientific process. New Engl J Med. 2015;372(1):7-9.
(7.) Pan American Health Organization. Knowledge Management and Communications Strategy. Proceedings of the 28th Pan American Sanitary Conference, Washington, DC, 17-21 September 2012. Available from: http://iris.paho.org/xmlui/ bitstream/handle/123456789/4455/CSP2812-e.pdf Accessed on 22 August 2016.
(8.) World Health Organization. Bridging the "know-do" gap meeting on knowledge translation in global health. 2005;25:2006. https://www.measureevaluation.org/resources/training/materials/high-impact- research-training-curricula/bridgingthe-know-do-gap.pdf Accessed on 22 August 2016..
(9.) Oatis J. US health Ebola chronology. Reuters. 2 October 2014. Available from: www.reuters.com/article/us-health-ebola-chronology-idUSKCN0HR2F820141002 Accessed on 9 November 2014.
(10.) Towers S, Afzal S, Bernal G, Bliss N, Brown S, Espinoza B, et al. Mass media and the contagion of fear: the case of Ebola in America. PLoS One. 2015;10(6):e0129179.
(11.) Chunara R, Andrews JR, Brownstein JS. Social and news media enable estimation of epidemiological patterns early in the 2010 Haitian cholera outbreak. Am J Trop Med Hyg. 2012;86:39-45. DOI: 10.4269/ ajtmh.2012.11-0597.
(12.) Santibanez S, Siegel V, O'Sullivan M, Lacson R, Jorstad C. Health communications and community mobilization during an Ebola response: partnerships with community and faith-based organizations. Public Health Rep. 2015;130(2):128.
Manuscript received on 25 August 2016. Accepted for publication on 31 January 2017.
Marcelo D'Agostino,  Felipe Mejia,  Ian Brooks,  Myrna Marti,  David Novillo-Ortiz,  and Gerardo de Cosio 
 Health Information and Analysis Unit, Department of Communicable Diseases and Health Analysis, Pan American Health Organization, Regional Office of the World Health Organization, Washington, DC, United States of America. Send correspondence to Marcelo D'Agostino, firstname.lastname@example.org
 International Consultant, Bogota, Colombia.
 School of Information Sciences, University of Illinois, Urbana-Champagne, Illinois, United States.
 Department of Knowledge Management, Bioethics, and Research, Pan American Health Organization, Washington, DC, United States.
Caption: FIGURE 1. Worldwide Google search trends on "Ebola" in English and Spanish, compared to "soccer" and "futbol" 28 February-31 October 2014
Caption: FIGURE 2. Worldwide tweets using #ebola integrated with 63 Ebola-associated chronological events (in red dots) in the time range, 16 March-16 October 2014
Caption: FIGURE 3. Worldwide Google trends for Ebola, dengue, and chikungunya with 63 Ebola-associated chronological events (in red dots) in the time range, 28 February-31 October 2014
TABLE 1. Content analysis of tweets in comparison with ebola-related topics based on the chronology of events. Area Characteristics of Language use the news I Suspicion of symptoms Baseline Launch of worrying --Languages: Arabic, reports French, German, Italian, and Korean --Use of phrase "out of control" II First case in Nigeria Use of "#ebola is real," "protect yourself" Use of words such as "worse," "worst," "worry," and "myths" II First entry into --Use of words like United States of "worry," "worse," America "fear," "risk," and "scare," and symbols such as ":(" --Use of the word "God" --Use of the word "vaccine" II International aid Use of the words initiatives "fear," "worse," "worry," "frightening," "panic" II First case in the Use of words "fear," United States "worse," "worry," "frightening," "panic" III First case in Spain --Use of words First case in France "fear," "worse," "worry," "frightening," "panic" --Tweets in Spanish --Tweets in Frenchc III First case in United Use of words "fear," States-- and spread "worse," "worry," declaration "frightening," "panic" Area Most common topics I --Consequences of the disease in the body. --Spread of the disease to other countries. --Ability of other African states to control the entrance of Ebola. --Mortality rate of Ebola. --Relation of the Ebola outbreak with other, seemingly unrelated events. II --Likelihood of Ebola hitting their countries. --The need to include other health aspects within the Ebola trend of information. --Touching others who might be infected. --Call for keeping panic under control. II --Demand for action. --Using the trend of Ebola to hide other important global topics. --Ebola hitting the Middle East. --Ebola hitting other continents. II --Excess of information related to Ebola. --Ebola mutations. --Airborne contagion. --Locals are "left to die;" are "flown out." --Doubt of the capabilities of the World Health Organization II --Ebola-infected person in United States. --Increase in Ebola facts especially from the White House. III --Symptoms that users associate with Ebola. --Other symptoms. --Airborne disease. --Conspiracy attitudes. --Use of facemasks. III --Symptom that users associate with Ebola. --Airport screenings to detect symptoms. --Contagion by health workers. --Living in areas with cases. --Little focus on African patients. --Virus considered manufactured. TABLE 2. Ebola by the numbers: worldwide content available on 16 March-2 November 2014 Tool Description YouTube YouTube allows billions of people to discover, watch, and share originally created videos. SlideShare SlideShare has grown to become the world's largest community for sharing presentations and other professional content. LinkedIn LinkedIn is a business-oriented social networking service, principally used for professional and career purposes. Google web search Google is one of the five most popular search engines in the world. Google news search Google News is a computer-generated news site that aggregates headlines from news sources worldwide. Google search videos Google Videos is a video search engine from Google. Google search books Google Books is a service that searches the full text of books and magazines that Google has scanned. Twitter Twitter is an online social networking service that enables users to send and read short 140-character messages called "tweets." #ebola hashtag On social media sites such as Twitter, a word or phrase preceded by a hash or pound sign (#) and used to identify messages on a specific topic. Tool Ebola content YouTube 806 000 videos SlideShare 5 345 presentations LinkedIn 4 338 posts Google web search 72 500 000 results Google news search 8 910 000 results Google search videos 25 200 000 results Google search books 1 250 results Twitter 7 605 318 tweets #ebola hashtag 1 281 731 tweets using #ebola hashtag Tool Additional information YouTube Search term: Ebola outbreak 2014 SlideShare --3 352 in English --8 79 in Spanish --5 44 in Portuguese --161 in French LinkedIn --70 professional groups discussing Ebola. --70 people self-identified as Centers for Disease Control and Prevention (United States). Google web search Search term: Ebola outbreak 2014 Google news search Search term: Ebola outbreak 2014 Google search videos Search term: Ebola outbreak 2014 Google search books Search term: Ebola outbreak 2014 Twitter Tweets on Ebola. Last 30 days. #ebola hashtag #ebola. Last 30 days.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Special report texto en ingles|
|Author:||D'Agostino, Marcelo; Mejia, Felipe; Brooks, Ian; Marti, Myrna; Novillo-Ortiz, David; de Cosio, Gerar|
|Publication:||Revista Panamericana de Salud Publica|
|Date:||Dec 1, 2017|
|Previous Article:||Factors associated with timely treatment of malaria in the Brazilian Amazon: a 10-year population-based study/Factores asociados con el tratamiento...|
|Next Article:||First detection of dengue and chikungunya viruses in natural populations of Aedes aegypti in Martinique during the 2013-2015 concomitant...|