Twitter® is a popular microblogging site that allows users to disseminate information in 140 characters of text or less. A review of literature indicated that, to date, there has been little inquiry into the health based discussions conceptualized and enacted within and among Twitter® users. Methods for this qualitative study included a directed content analysis, guided by the Public Health Agency of Canada’s Determinant of Health (DOH) framework was completed to explore health based discussions on Twitter®. A 24-hour cross-section of tweets (N=2400) containing the word or hashtag ‘health’ were collected for analysis. Findings revealed predominant themes of health services, personal health practices, and education. Many of the tweeted messages reflected existing political and social issues publicized within the global mass media. This study also considered the evolving dynamic behind the conceptualization of health and how it is co-constructed through news media, advertising, and social network technologies. Discussion of the emerging themes and implications for practice are presented.
Keywords: Twitter®, social media, health, health promotion, tweet, determinants of health, web 2.0, microblogging, social networking, internet, technology, informatics, content analysis, qualitative research
Searching for health information is one of the more popular web-related activities, preceded only by email and general browsing of search engines. More than ever before, use of the Internet has fundamentally altered how people share personal experiences, search for information, and make decisions of daily living, particularly health care decisions (Estabrook, Lee Rainie, & Witt, 2007; Horrigan & Lee Rainie, 2006). To date, 70% of Canadians have used the Internet to search for health-related information (Statistics Canada, 2010). In the United States, 77% of the adult population regularly accesses the Internet and approximately 83% of those online have used the Internet to search for health information (Pew Internet & American Life Project, 2011). Searching for health information is one of the more popular web-related activities, preceded only by email and general browsing of search engines (Pew Internet & American Life Project). Those most likely to search for online health information have been characterized as: highly active Internet users (involved in multiple online activities); women; those with higher levels of attained education; married individuals; those with high-speed Internet access; and individuals living in urban areas (Underhill & McKeown, 2008). In addition, researchers from The Pew Internet & American Life Project reported that 78% of wireless Internet users (e.g., laptops, cell phones) have looked for online health information (Fox, 2010).
The use of social media technology is increasing. “Social media,” largely synonymous with the terms “Web 2.0” and “participatory media,” refers to online services where the content is generated and manipulated by Internet users (Gray, 2011; “Social Media,” 2012). Social media tools, including technologies like blogs, podcasts, wikis, information aggregators, and social networking sites, allow users to engage in the exchange of user-generated material and information (Gray, 2011). A social network site is commonly understood as an online location where individuals can create personal profiles and build networks of contacts among other online participants. Participation in social networking sites currently engages tens of millions of Internet users on a given day (Lenhart & Madden, 2007). This increased prevalence and use of social media technology has facilitated the development of active networks of individuals based on shared interests, activities, or beliefs. Evolving communication behaviours have subsequently led to the embedding of some social media outlets within everyday activities (Lenhart & Madden, 2007; Lenhart, Madden, Smith, & Macgill, 2007).
Recently, microblogging has become a popular addition to the repertoire of social media possibilities available to Internet users. Recently, microblogging has become a popular addition to the repertoire of social media possibilities available to Internet users. Microblogging is a form of social media technology that allows users to share and exchange small amounts of textual content and media ("Microblogging," 2011). Twitter®, a popular microblogging service, limits online journaling to 140 or fewer text characters. To participate, a ‘Twitter®’ account must be activated and details of the account set according to the user’s preferences (e.g., password generation, mobile phone registration, and notification of communication tweets). Messages called "tweets" can be generated from the online Twitter® website (or linked with other social networking applications), a personal mobile phone, or via a third-party Twitter®/instant message client. Currently it is estimated that Twitter® has at least 200 million registered accounts and generates roughly 110 million tweets per day (Chiang, 2011).
During the H1N1 pandemic, Twitter® was used as a communication modality to update the public on wait times at flu vaccination clinics and for the distribution of government alerts. Despite Twitter’s® ubiquitous presence online, there has been limited research to examine this modality of communication as a means of conveying and sharing personal health experiences and information. Existing examples of Twitter® use for health promotion were recently highlighted during the reporting of the H1N1 flu pandemic (Centers for Disease Control and Prevention [CDC], 2009; The Canadian Press, 2009). During the H1N1 pandemic, Twitter® was used as a communication modality to update the public on wait times at flu vaccination clinics and for the distribution of government alerts (CDC, 2009; The Canadian Press, 2009). Similarly, Twitter® was used to stimulate fundraising and the direction of medical services for areas affected by natural disasters or political unrest. During the 2009 Haitian earthquake, a number of entities utilized Twitter® as a communication tool to assist in raising money and awareness of relief efforts (Sample Ward, 2011).
Researchers assessing physician use of Twitter® for professional practice found that clinicians effectively shared medical information with the public, but there were instances of ethical breaches and unprofessional content (e.g., patient privacy violations and conflicts of interest) (Chretien, Azar, & Kind, 2011). Winston et al. (2012) explored online discussions within a Twitter® community of individuals managing type I diabetes. More often than clinical information (e.g., footcare), participants within this network typically posted experiential health information, reporting on daily life events, practical guidelines (e.g., checking blood sugar), and problems in maintaining sports activities.
As a relatively new technological innovation, the use and functionality of Twitter® as a modality for health communication has not been thoroughly explored in the literature. Specifically, there has been a lack of research examining the role microblogging may play in the promotion of health for individuals using this social media service. To begin to explore the use of microblogging in health promotion, the current study sought to capture “everyday conversations” about health broadcast on Twitter®.Therefore, the purpose of this study was to gain an initial understanding of “health” as discussed among a global network of individuals using the microblogging platform Twitter®. This study contributes one of the first examinations of how Twitter® users express and conceptualize issues that influence “health” in alignment with a preexisting determinant of health framework (Public Health Agency of Canada, 2003).
Review of Literature
Health Information Seekers
Evidence indicates that online health information seekers are interested in information on specific diseases, symptoms, surgical procedures, lifestyle factors (e.g., diet, nutrition and exercise), medications, alternative therapies, and the health system. Evidence indicates that online health information seekers are interested in information on specific diseases, symptoms, surgical procedures, lifestyle factors (e.g., diet, nutrition and exercise), medications, alternative therapies, and the health system (Underhill & McKeown, 2008). The predominance of “disease and lifestyle” related online searching supports a lay conceptualization of “health” that aligns with illness, behaviour change, and self care (Underhill & McKeown). Yet evidence suggests that health is influenced by an intersection of social, economic, and policy issues as well as physical factors and conditions referred to as the Determinants of Health (DOH) (Public Health Agency of Canada, 2001; World Health Organization, 1998). The DOH framework acknowledges the intersecting influence of lifestyle practices; social support; health services; education; employment and income; gender; culture; and physical and social environments on health (Public Health Agency of Canada, 2003).
Criticism of medical practitioners, government, and producers of mass media has been directed at their focused attention on physical determinants of health (e.g., physical, lifestyle factors) in neglect of social health determinants (e.g., income equity, gender, culture) (Raphael, 2008, 2009; Raphael, Curry-Stevens, & Bryant, 2008). Where an empirical understanding of the social determinants on health exists in scholarly circles, lay understanding of the link between the social determinants and health tends to be implicitly understood in the form of “experiential knowledge” or through one’s lived experience.
Connectedness and Health
Christakis and Fowler (2010) demonstrated that individuals who are highly socially-networked are at higher risk levels of infection during outbreaks of infectious disease. They proposed that tracking the online behaviour of individuals who are connected via social networks in conjunction with data mined from other information repositories (i.e., Google) could enable public health officials to create real-time information streams regarding the disease epidemic.
...researchers proposed that the information found within social media data has the potential to serve as a proxy for gauging public opinion on a given situation like flu outbreaks. Kostkova, de Quincey, and Jawaheer (2010) summarized their analysis of over one million tweets collected during the recent H1N1 outbreak (May and August 2009) and found numerous notations of specific words and phrases related to the H1N1 outbreak, including 2888 instances of “I have swine flu” and 1530 occurrences of “I have flu.” Ritterman, Osborne, and Klein (2009) also collected tweets during 2009 H1N1 outbreak and used 48 million tweets collected over a three month period to predict locations of flu outbreaks. They purport that further work is needed to ascertain whether microblogging can be leveraged to better understand public knowledge of infectious outbreaks and to create early warning systems for public health. Furthermore, researchers proposed that the information found within social media data has the potential to serve as a proxy for gauging public opinion on a given situation like flu outbreaks (O'Connor, Balasubramanyan, Routledge, & Smith, 2010; Ritterman et al., 2009). McNab (2009) elaborated on the importance of this communication modality by stating that “[T]witter and other social media tools can help to bring accurate health information to many more people than ever before. After all, one fact sheet or an emergency message about an outbreak can be spread through Twitter faster than any influenza virus” (p. 566).
Hughes and Palen (2009) provided a descriptive account of Twitter® use during emergency and large convergence events. They concluded that Twitter® is a potential communication modality for fast and efficient dissemination of information, in part due to a reduction of person-specific replies and greater frequency of embedded web links to emergency information sources.
Topical Content Analysis of Twitter® Messages
In the past few years, researchers have attempted to classify different types of Twitter® messages (Honeycutt & Herring, 2009; Java, Song, Finin, & Tseng, 2007; Naaman, Boase, & Lai, 2010). Java et al. (2007) were one of the first to examine the message and geographic disbursement of Twitter® content in their highly cited study. Based on their findings, they contend that the main intent of Twitter® use includes daily (personalized) chatter, conversations, information sharing, and news reporting. Naaman et al. (2010) extended the work of Java et al. by demonstrating that Twitter® users largely fit two “content camps.” They claim that 20% of the users in their analysis shared information with others (i.e., the ‘Informers’), while 80% of their sample they labeled as “Meformers,” a term outlining the self-focused and/or self-indulgent nature of many users’ tweets. Honeycutt and Herring (2009) also examined a number of features related to Twitter® messages including the use of the ‘@’ symbol. They found that tweets which utilized the @ feature of Twitter® were more likely to be sharing information, wishing to exhort other users, and/or demonstrate the interactivity of the content.
Chew and Eysenbach (2010) collected over two million Twitter® posts during the H1N1 pandemic (data collection May 1 to December 31, 2009) containing the key words ‘swine flu’, ‘swineflu’, and/or ‘H1N1’ and used content analysis to examine a subset of 5395 tweets. Tweet ‘themes’ that emerged from the data included sharing resources, statements of personal experience, opinion/interest, humour, frustration, concern, relief, misinformation, and questioning. Only a small subset of coded tweets (4.5%) was classified as possible misinformation or speculation. Chew and Eysenback (2010) also discovered that over 90% of these tweets contained references (e.g., web links to news or information sites) to information in the message.
Finally, physiotherapist researchers Sullivan et al. (2012) collected 3488 tweets using eight search terms related to the traumatic brain injury, ‘concussions’. Using 1000 randomly selected tweets from the overall database, the researchers developed a coding scheme to determine major content themes contained within each of the 1000 tweets. Overall, Sullivan et al. (2012) discovered that the majority of coded tweets on concussion were discussions related to news events (33%), including stories of professional athletes who had suffered concussions. Sharing of personal information/situations related to concussions (27%), and inferred management of concussions (13%) were also popular Tweet themes.
A qualitative directed content analysis (Hsieh & Shannon, 2005) was used to investigate ‘tweeted’ information that aligned with issues of health. A directed content analysis is an approach “to validate or extend conceptually a theoretical framework or theory” (Hsieh & Shannon, 2005, p. 1281). Using an existing framework, researchers identify key concepts or variables and map these passages to predetermined theoretical codes. Any text not categorizable within the initial coding scheme is provided a new code (Hsieh & Shannon, 2005).
This research was based on the assumption that Twitter® users contributed personally meaningful information to the social networking site and that Twitter® was considered a communication medium for contributing and/or communicating about ‘everyday’ issues (Java, Song, Finin, & Tseng, 2007). Although data published using Twitter® is publically available information, we obtained confirmation from the Office of Research Ethics that ethical research standards were maintained without requiring consent of individual Twitter® users’ whose data was captured. Twitter® usernames and other personally identifiable hashtags (excepting the current President of the United States of America) were stripped from results presented in this article.
Using The Archivist (MIX Online, 2011) data collection software program, a database of 36,042 tweets containing the word ‘health’ was collected over four consecutive days. Since the purpose of the study was to obtain an initial ‘snapshot’ of health conceptualization on Twitter®, the timing of the data collection was not intended to correspond to any specific event or seasonal time period. The June 2009 data collection time frame was a convenience sampling time period.
The Archivist is a Windows based application that captures and records in real time a longitudinal Twitter® feed of a specified search term. The software also allows for exportation of captured tweets to Excel and XML formats for data-mining purposes. The Archivist program was run on the researcher’s desktop computer from 16 June 2009 (19:32 GMT) until 20 June 2009 (12:02 GMT). The word health was captured either as a single word, part of a word (i.e., healthcare), or in a URL or hashtag (i.e., #health). Tweets that were incomprehensible, not in English, or used health in an inapt context (e.g., word ‘health’ used in reference to ‘health pack,’ an element of first-person shooter video game) were not included in the study. To provide context to Twitter® conversations, a review of major world events preceding the data collection date was explored via review of electronic news media.
Given the enormity of collected data (n = 36,042), researchers selected for analysis the first one hundred tweets from the end of each hour of June 19, 2009 (starting at 05:00 GMT) for a 24-hour period. Friday, June 19, 2009 was purposefully selected largely based on researchers’ assumption that tweets posted at the end of the week, relative to other days of the week, would encompass the week’s events. The final data set used in this study contained 2400 tweets containing the word health. The analysis was guided by the Public Health Agency of Canada’s Determinants of Health (DOH) framework (Public Health Agency of Canada, 2001). The DOH are twelve interrelated domains that influence health. These twelve determinants are briefly described below (Table 1).
Table 1. Determinants of Health (Public Health Agency of Canada, 2001)
Determinant of Health
Influence on Health
Income and Social Status
Examines aspects of both income and social status, including how higher levels of income are associated with agency to live and work in places that are conducive to overall health.
Social Support Networks
Social networks of friends, family, and associates are an important. It has been found that individuals enrolled in meaningful social networks have protective properties in terms of overall health and wellbeing.
Education and Literacy
Education / literacy is positively associated with health. Higher levels of education generally provide greater access to jobs, resources, and opportunities.
Underemployment, unemployment, or unsafe working employment arrangements can significantly impact both emotional and physical health.
Overall health is positively impacted by the provision of social stability, safe communities, and positive interactions with others.
The health of individuals is influenced through natural (e.g., air, water, soil), and physical surroundings of housing/dwellings, transportation systems, and design of communities.
Personal Health Practices (PHP) and Coping Skills
This factor is associated with various social, economic, and environmental attributes that influence people’s lifestyles and the choices they make that have effects on their health. Many times, people’s choices in terms of PHP and coping skills are context dependent and socially constructed, therefore it would be inappropriate to assume that all individuals have agency over their PHP and coping skills.
Healthy Child Development
The early development of infants and children is influential in the growth and development continuum. Development is influenced by parenting, family income, food security, education, and access to health and dental care.
Biology and Genetic Endowment
Genetic endowment or predisposition to certain conditions can severely impact an individual’s long term wellness and health.
Access to both population health and formal health services are an important in the maintenance of health.
Gender is a socially constructed factor that determines norms and health expectancies. Gendered health concerns and conditions revolve around pre-existing notions of gender and can give primacy to one gender’s issues over another.
Various types of marginalization, stigmatization, and lack of cultural sensitivity can impact the health of individuals /families.
Two researchers independently evaluated Twitter® messages captured verbatim using The Archivist program. A coding form structured from the DOH categories guided analysis. The data was initially reviewed using predefined coding categories (Hsieh & Shannon, 2005). Additional coding categories were created inductively from analysis from data not fitting a priori DOH coding categories.
Researchers maintained an audit of their reflections of the analytic process (i.e., captured ideas and thoughts regarding category and theme development) which provided an accounting of analysis process used to assist in resolution of analytical discrepancies (Lincoln, Lynham, & Guba, 2011; Krefting, 1991). Researcher inconsistencies were thoroughly discussed until consensus was achieved. Trustworthiness and validation of findings was established through systematic data analysis, analyst triangulation, and verbatim data collection. Basic descriptive statistics were completed on the health tweets.
Media Context: June 19, 2009
To provide context to Twitter® conversations, we conducted and summarized a review of major world events (i.e., news headlines) over the month of June 2009 preceding the data collection date. The news headlines provided a list of ‘current event’ issues that were reproduced and disseminated by multiple mass media communication channels (e.g., televised news, news magazines, print newspapers, news based television programming). These current events served to stimulate and direct conversation; therefore, we collected Canadian, U.S., and world news headlines from online news feeds. Significant U.S. health-related news was noted in the weeks preceding the data collection date. Two weeks prior to the data collection point, a proposed healthcare reform bill in the United States dominated the American news. Beginning June 9, 2009 the U.S. Democrats unveiled a healthcare reform bill aimed at reshaping the healthcare system in the United States, and became a reoccurring headline in the news sources. Other major news headlines from this time period included the Air France crash over the Atlantic Ocean; Swine Flu (H1N1) cases/deaths; U.S. economic stimulus expenditures; a fatal shooting at a Holocaust Museum; and details regarding the June 2009 Iranian election results and ensuing public demonstrations.
Determinants of Health
The 2,400 tweets were coded to the DOH framework in terms of message content. Tweets that did not fit the DOH framework were handled in three ways. Two additional codes were created inductively from the data. The “other” code was created in response to tweets that did not fit the DOH framework, but still contained significant, but unspecified health content. Similarly, an “advertising” code was added to record tweets focused on marketing health related products or personnel recruitment. Of the 2400 tweets in the data set, 1661 (69%) were coded to the DOH, “other” or “advertising” themes (see Figure 1). Due to use of the word health in an inapt context, 739 tweets (31%) were not able to be coded to any category. The majority of tweets were coded to “health services” (n = 669, 28%), followed numerically in terms of magnitude by “personal health and coping skills” (n = 252, 11%), “other” (n = 237, 10%), “advertising” (n = 176, 7%), and “education / literacy” (n = 131, 5.5%). Overall, these five codes accounted for roughly 87% of coded tweets in the database.
Figure 1. Demographics of Coded Tweets
The DOH themes of income and social status, social support networks, environment/working conditions, social environments, physician environments, healthy child development, biological and genetic endowment, gender, and culture collectively amounted to 196 (8%) tweets. The most prevalent in this collection included tweets examining “social support networks” (n = 41, 2%); “physical environments” (n = 64, 3%); and “gender” (n = 43, 2%). Examples of these lesser tweeted DOH can be found below (see Table 2).
Table 2. DOH Categories and Select Message Examples
1. Income and Social Status
"Know of businesses that offer health coverage to part-time employees. Let me know."
2. Social Support Networks
"totally broken hearted hearing the news about my brother's health.... please pray for him and me. ill find out soon..."
"I'm not in the best health right now, but my spirits are high because God has blessed me with such loving, caring people :-)"
3. Education / Literacy
"Please read this article on health care http://bit.ly/wea8v"
"Went to an inspiring talk in Seattle about Global Health given by Dr. Paul Farmer. He is the focus of a book 'Mountains beyond Mountains.'"
4. Employment/Working Conditions
"Managing Chronic Illness at Work (NYT): http://bit.ly/aM99N#health"
"If you're an employer changing group health plans midyear, be sure to ask if your new carrier will credit deductibles from your prior plan!"
5. Social Environments
"Oh no, there are no health disparities in Australia. None at all. (We're just talking about white people, yeah?)"
"Army’s 1st study of the mental health of troops who fought in Iraq found that 1 in 8 reported symptoms of post-traumatic stress disorder."
6. Physical Environments
"Oh, my lungs are burning. Gridlock is bad for your health"
7. Personal Health Practices and Coping Skills
"I'm more than a little in love with personal health days!!"
"Men's Health unveils iPhone workout app http://tinyurl.com/mnsrrb"
8. Healthy Child Development
"BBC: Half of teachers believe the health/safety culture in schools is damaging children's learning and development http://tinyurl.com/lcjjq9"
"USA: RI #H1N1 | HEALTH does not automatically recommend closing schools [Jun 19 Providence] http://bit.ly/uhuIi"
9. Biology and Genetic Endowment
"Another BPA Study Shows Reproductive Health Effects http://bit.ly/Ex0pz"
10. Health Services
"Why are health care costs perpetually increasing? Did the cost of an x-ray go up? Price sure did."
"Holy shit! Having health insurance is awesome! Back when I had my sinus infection i went to urgent care, I only have to pay $25 out of $100!"
"Male Anorexia is an often unseen, undiagnosed health problem on the rise today http://tinyurl.com/knzbek"
"Domestic violence, the most common cause of injury to women ages 15 to 44, can lead to long-term health problems an http://bit.ly/14xads"
"Understanding health disparities: Race, class, both? AAs are sicker & dying at faster rate http://tinyurl.com/oo3v4q (via @user)"
"@user Yeah, and I hated health class..."
"6 hours of community dental health....yay! how exciting. cant wait til this day is over!"
"Cheap health insurance is AVAILABLE. message me."
"Max Muscle has solutions for everyday life, from the casual gym user to bodybuilders and athletes to general health, stop in today!"
The five predominant themes of “health services” (n = 669, 28%); “personal health and coping skills” (n = 252, 11%); “other” (n = 237, 10%); “advertising” (n = 176, 7%); and “education/literacy” (n = 131, 5.5%) are discussed in greater detail in the following section.
Health services. “ Health services” was the most tweeted theme (n = 669, 28%) in the data set. The high frequency of tweets related to “health services” appears to align with the highly publicized and debated U.S. health care reform that frequented the mass media news reports. A sub-theme of “dividedness” resonated from the polarized tweets about socialized healthcare in the United States. Within this sub-theme there were tweets opposing proposed U.S. health care reform:
“Not a single thing sounds good or reasonable about the government health care ideas”
“@BarackObama I would sign her cast, but I'm afraid it might have to be reused under your health care plan...mmm rationing”
There were also tweets endorsing the proposed legislation:
“hoping for affordable health care for all....Come'on Pres Obama..make it happen......you promised!”
“just donated to the Obama health care movement. join me....”.
Although personal opinion statements to proposed health changes were frequent, tweets that forwarded news media headlines were also common. A number of tweets contained the URL to the source of information or media which supported their opinion as indicated by the following:
“Democrats push for new government health plan: WASHINGTON (Reuters) - Democrat lawmakers on Friday proposed guar.. http://ping.fm/LOsNW”
“Socialized Health Care Horror Stories: Cancer: With Obamacare increasingly becoming the f.. http://bit.ly/uXun3”
There were also a number of tweets reflecting U.S. health reform indifference, resignation, or frustration with the process of political reform. The frustration some experienced when observing the political-healthcare process is reflected below:
“Debating whether I should write to congress about health care. Last time I tried, they talked down to me, so I doubt it'll do anything.”
“I got a feeling the only way I'm ever going to know what public health care is to move to Europe....”
Personal health practices & coping skills. “Personal health practices and coping skills” (n = 252, 11%) was the second most tweeted DOH theme. Given conversational and declarative tendencies of Twitter® users (Kelly, 2009), it was anticipated that this determinant might have been foremost in terms of frequency. Further research is needed to determine whether this presumption holds true during a time period that is not subject to the influencing forces of high-level political changes and related news media coverage in regards to health care.
Views of health are largely constructed by individuals and people within a particular social context; tweeters in this analysis conceptualized health practices from a behavioural perspective. For instance, there seemed to be a proclamation of lifestyle practices within some tweeter’s postings, as indicated by the following:
“Improving my health, gone vegan”
“Leaving work, taking a half day. For my own mental health and well-being”
Tweets appeared to describe a health practice that the individual felt was meaningful enough to broadcast to followers, and ultimately, the ‘twittersphere’. Although lack of context precludes any generalization, it does seem that some individuals tweeting within this theme appeared to be reporting on actions taken to improve their health. This notion is captured in the following tweets:
“Oh okay i cant drive due my health but i can walk around the entire river walk”
“I'm here to tell you... heart disease doesn't have to happen to you if you decide once-and-for-all to take charge of your own health!”
Tweeters also seemed eager to share health advice with followers and suggested treatments and/or therapies they perceived as healthful as indicated below:
“cranberry juice is awful. but it promotes urinary health, folks!”
“I recommend sprouted grain breads and sourdoughs to my patients for better health and blood sugar control.”
Given the 140 character message limitation of Twitter®, the context of individual tweets was not discernible. Because of this, it is unclear as to whether tweeters recommending certain health practices actually understood or appreciated the safety or efficacy of their recommendations:
“Eating Cayenne peppers is good for your health. Lowers cholesterol”
“I have been eating stuff for digestive health and drinking this tea which promotes more urine production, yep cleaning myself out again!”
Other. The Other theme captured numerous health related tweets (n = 237, 10%) that could not be classified into the 12 determinants of health. A variety of topics found in the Other category included tweets regarding conversational or declarative messages with humorous backgrounds; health used in a sarcastic context; or prompted readers to individual websites for more information than could be conveyed in the tweet. Although there were no specific health issues that dominated this theme, there were a number of interesting health-related tweets captured.
“Declining mental health? There's an app for that”
“Will I ever meet my sixth grade health teacher someday so I can tell her that the anti-drug unit only made us especially eager to try drugs?”
“It's a pity water is not harmful to health. It would be a delicious sin.”
“I'm wondering if there would be much interest in Health Camp Wales? http://bit.ly/1yBW7 “
Advertising. Advertising related tweets (n = 176, 7%) were either product based or employment recruitment in nature. Various products and roles were advertised from employment positions, to various pharmaceuticals and health aids.
“Hamilton Jobs Registered nurse - public and community health (Hamilton - Ontario)”
“physical therapist - Providence Health & Services - SEATTLE, WA #tweetmyjobs #jobs”
There were a number of tweets advertising “cheap” or “affordable” health insurance. Finally, evidence of potential phishing tweets (e.g., fraudulent messages) offered non-descript improvements to life and health wellbeing as indicated in these examples:
“Cheap health insurance is AVAILABLE. message me.”
“Health Breakthrough works for 100% of people who take product must see”
Education / Literacy. Individual tweets within the theme of “education / literacy” (n = 131, 5.5%) seemed to possess a higher level of information literacy than some of the tweets captured in other themes. Tweeters commented upon or forwarded articles and/or reviews from scholarly journals or reputable health websites. Many of these tweets seemed to be interwoven into media related websites, outlining results captured in forthcoming medical/health journal publications. The shortened URLs embedded in some of the tweets provided interesting insight into the dissemination of health education information using Twitter®. A few examples are provided below:
“[The Lancet] [Articles] Financing of global health: tracking development assistance for health from .. http://tinyurl.com/mygvw4” [links to an abstract outlining Ravishankar et al.’s (2009) report, entitled: Financing of global health: tracking development assistance for health from 1990 to 2007]
“Health More dads working long hours, study confirms: BlackBerry in hand and baby in tow, dads don't have m.. http://tinyurl.com/lgvknr” [links to an MSNBC article that references a report from Van Echtelt et al. (2009) Post-Fordist Work: A Man's World? Gender & Society, 23(2).]
The majority of tweets in this theme carried a URL that directed others to the primary information source or a secondary source of health information. A wide-cross section of topics from health economics to disease prevention, recovery, and self-help techniques were included within the 131 tweets.
Interestingly, there were a number of school-related tweets from students seeking advice, information, or connecting over health-related academic work:
“PhD student seeking information on maternal mental health, any experts out there? Please @ reply or DM me”
“Waah! Somebody help me in researching what are the rights of health care provider?.. T_T”
The purpose of this investigation was to explore a cross-section of health-related Twitter® messages captured over a 24 hour period, using a DOH framework (Public Health Agency of Canada, 2003). A total of 2,400 tweets were included in the overall analysis which was complemented with a collection of Canadian, American, and world news headlines one month leading to the data collection date. By mapping tweets onto the DOH framework (Public Health Agency of Canada, 2003), the themes of health services; personal health practices and coping skills; and education accounted for the majority of coded tweeted within the data set. The other and advertising domains, inductively generated from the data, also captured a significant number of the overall tweets. Almost one third of tweets including the term “health” were ultimately excluded for using health in an inapt context (n= 739, 31%). The remaining tweets suggested several points for consideration.
Media and Health
The current investigation supports the importance of social context and politics in stimulating the frequency and quantity of health services-related tweets. For instance, the United States health care reform coincided with the data collection time period and likely inflated the number of tweets related to the “health services” theme. The influence of other political and world events is less clear when examining the highly tweeted themes of “personal health practices and coping skills” and “education / literacy.” Perhaps certain DOH themes are more sensitive and reactive to contextual and political events, whereas others, such as “personal health practices and coping skills,” retain a baseline of popularity due to relevance for a large number of individuals. A larger data set is required to examine these proposed inferences.
Similarly, it is difficult to speculate why domains like “social environments” and “income/social status” were not as popular as other themes found within the DOH framework. It may be that tweets within these themes were less generalizable to other users in terms of practical knowledge and information utility. In popularly tweeted themes, users seemed to value sharing conversational and informational knowledge that they believed was valuable both to them and others. Given the specificity of some of the lesser tweeted domains, it could be theorized that individuals who microblog tend toward topics that they believe their followers (or the twittersphere in general) might find interesting or useful. Zhao and Rosson (2009) in their exploratory study of Twitter® in the workplace found that tweeters generally shared information that they themselves find useful or interesting.
Mostly general information sharing between and among Tweeters and their followers has been reported in the current literature (Ebner, Lienhardt, Rohs, & Meyer, 2010; Huberman, Romero, & Wu, 2009; Zhao & Rosson, 2009). This study did not include an analysis of health information seeking behaviours using Twitter® (e.g., where health information was generated through a keyword search, or, pushed to a user as a subscribed follower). As Huberman et al. (2009) discovered, the immediate number of ‘followers’ an individual possesses does not automatically imply interaction between them. Instead, a ‘hidden social network’ of “active users” operates within the wider and denser network. It is these active users who are prolific participants, disseminate messages, or create trending topics (Huberman et al., 2009). This notion of active users is also collaborated by McCandless (2009) who found that roughly 5% of the Twitter® users create around 75% of all tweets. Further research is needed to determine if and how many participants function as key broadcasters in disseminating various types of health information.
Media and Politics
The use of Twitter® by politicians/government and media agencies is gaining in popularity (Golbeck, Grimes, & Rogers, 2010). Our analysis found a thematic link between tweets and current events/politics, particularly in terms of the reformation of health care systems. Although this finding is not unprecedented given the recent use of microblogging during politically charged events (e.g., United States presidential election 2008; Iranian presidential election demonstrations 2009; G20 Toronto demonstrations 2010; Arab Spring 2011; Canadian Federal election 2011), it does draw attention and gives a “voice” to the public who use social media technology to actively engage in co-constructing ideas regarding health and healthcare in the 21st century (Hampton, Goulet, Rainie, & Purcell, 2011). Recognized benefits of social media technology include the ability to accelerate the pace of discovery, broaden social networks, and refine questions within personal health care encounters (Fox & Jones, 2009). Our findings suggest that these same processes (e.g., expanded social networks, enhanced knowledge) directed at individual health care may also benefit community and social health issues using social media technology. ...the interaction of news media and politics appears to be a far more complex and had significant influence on the conversational focus of health issues reported by tweeters... Social media technology is moving individuals from passive recipients of information to active participants in shaping messages (Benkler, 2006). Scanfield et al. (2010) similarly demonstrated that expanded social networks and increased access to information about antibiotic use/misuse exists among individuals using Twitter®.They tracked dissemination of ‘antibiotic’ information through networks of followers and ‘‘retweeting’’ activity and in this way demonstrated the use of microblogging among informal social networks for dissemination of both valid and invalid information. While the current investigation did not track interactivity of tweeted messages, our findings also suggest that social media technology is moving individuals from passive recipients of information to active participants (Benkler, 2006).
As demonstrated from this data set, the interaction of news media and politics appears to be a far more complex and had significant influence on the conversational focus of health issues reported by tweeters in the current study.
The current analysis suggests the use of a collaborative model of shared health information made possible through the use of microblogging. Fox and Jones (2009) examined how people used the Internet for health care, and found that many people go online to connect with health professionals, friends, and families. In addition, more people read blogs, listen to podcasts, and post comments on their own and others’ social network sites.A number of coded tweets in the Personal Health Practices & Coping Skills theme demonstrated this notion of apomediation – individuals bypassing traditional health information gatekeepers and using apomediaries to guide their information search and collection. The 1,661 (69%) of the 2,400 tweets in this study reflected a “determinant of health” issue supports an evolving health communication framework and demonstrates an emerging avenue for consumers to provide and obtain health information. The “crowd-sourcing” nature of Twitter® allows health consumers, researchers, or students to tap into a global source of advice, support, and/or information (Eysenbach, 2008; Gray, 2011).
Eysenbach's (2008) concept of ‘information apomediation’ seems to be supported from the results obtained in this study. Information apomediation describes a process whereby individuals circumvent traditional information gatekeepers (i.e., healthcare providers) and are able to access information directly. Apomediaries are people and tools that ‘assist’ the individual in searching for information, but do not act as a gatekeeper. For example, Fox (2010) has reported on the growing presence of a unique online participant group, ‘patient opinion leaders’ (POLs), who specialize in issues of cancer, diabetes, HIV, mental health, and other chronic diseases. A number of coded tweets in the Personal Health Practices & Coping Skills theme demonstrated this notion of apomediation – individuals bypassing traditional health information gatekeepers and using apomediaries to guide their information search and collection.
As with all research there are limitations. The current research only examined a one-time ‘snapshot’ of tweets published in English, and constrained to the first one hundred tweets from the end of each hour on 19 June 2009 until 20 June 2009 (24 hour period). Tweets published in other languages (e.g., French, Spanish, languages of First Nations groups) may have emphasized different DOH. Similarly, the date of data collection and the current events occurring across the world likely influenced the predominance of select DOH themes over others. While this initial snapshot of tweets provided some insight into use of microblogging about health, a longitudinal data collection period and/or data collected across several points of time would be beneficial in demonstrating patterns within conversations regarding health and whether some DOH themes retain a baseline popularity/frequency that is unchanged by emerging global events.
Another limitation was the use of only one search term (i.e., health) for data mining. Recognizing that many health-related discussions can occur without use of the word “health” future research examining Twitter® users’ “health-related” discussions will require broader search algorithms to enhance comprehensiveness of data capture. Finally, the methods of this study did not enable the researchers to collect data on interactions between users, nor evaluate accuracy of tweeted health information. These two attributes were not sought as part of this study, but do present as limitations during interpretation of results. Without exploring the network of interactions between users, it is difficult to ascertain patterns of information and knowledge exchange within this communication modality. Equally, not knowing the accuracy of the tweets also hampers overall interpretation of results since it is unclear if misinformation related to health was perpetuated throughout the data set.
Implications for Practice
...issues that determine “health” are broadly discussed by users of social media technologies such as Twitter®. This study has presented a snapshot of messages posted to Twitter® (n = 2400) over a 24 hour period in June 2009. Tweets were mapped to the Public Health Agency of Canada's (2001) DOH framework and further analyzed for thematic content. Study results demonstrated that issues that determine “health” are broadly discussed by users of social media technologies such as Twitter®. Although results of this study are not generalizable, the findings do reinforce the emerging potential and functionality of social media technologies like Twitter® in contributing to health based conversations directed at individual, community, and societal levels. For public health educators, Twitter® and other social media networks can play a substantial role in effective information dissemination but only if strategic communication processes (e.g., a dense network made up of followers and followees, consistent and ‘real-time’ contact) are used (Winston et al., 2012). Thackery et al. (2012) warn against use of social media technologies as virtual ‘pamphlet walls.’ Recognizing that public health departments are ‘early adopters’ of social media technologies, they emphasize the need to exploit the interactive nature of Health 2.0 modalities (e.g., Twitter®) and engage with their audience(s) to meet their information needs rather than ‘push’ information out broadly.
Little is known about best practices for assisting clients in their acquisition, critique, and implementation of health information obtained from Internet and social media sources. The limited published literature on social media technologies and health care creates some consternation that we are “late to the social media game.” Little is known about best practices for assisting clients in their acquisition, critique, and implementation of health information obtained from Internet and social media sources. That approximately 50% of North Americans function below the expected minimum health literacy level and are challenged by increasingly complex material (CCL, 2007; Human Resources Development Canada, 2003; National Center for Education Statistics, 2006) highlights the role of clinician as health information navigators. Researchers assessing physicians’ use and misuse (e.g., ethical breaches, privacy violations) of Twitter® for professional practice underscore the need for additional education and research about this evolving practice role (Chretien et al., 2011) and others, such as “health information navigator.”
The social media-politics dynamic and its influence on public discussions of health is an area that deserves further exploration... A large number of tweets analyzed in this study (especially in the “Health Services” theme) appear influenced and shaped by both social and political contexts. This finding demonstrates the fluid and dynamic linkages between mass media, microblogging, and the concept of health that require additional research inquiry. For future work, a longitudinal survey of data generated on Twitter® should be studied to capture the subtle communication nuances and patterns present in large scale networks like those found between and among Twitter® users. The social media-politics dynamic and its influence on public discussions of health is an area that deserves further exploration specifically to ascertain how representation of health and use of social media technology impact public conceptualization and enactment of health and healthcare.
Lorie Donelle, PhD, RN
Lorie Donelle is an assistant professor with Western University’s Faculty of Health Sciences jointly with the School of Nursing and School of Health Studies in London, Ontario, Canada. Under the umbrella of health promotion her program of research focuses on the concept of health literacy particularly among marginalized populations. As well, an intersecting research tract acknowledges the importance of the “Information Age of Healthcare” and includes the investigation of information technology use (e.g., Internet, electronic/personal health records, social media) in promoting health.
Richard G. Booth, MScN, RN
Richard Booth is a doctoral candidate and lecturer at Western University (London, Ontario, Canada) studying clinician learning and adoption of health technology. He works clinically as a psychiatric-mental health nurse in the adult psychosis program at St. Joseph’s Health Care London. Currently, he teaches in the undergraduate program at the Arthur Labatt Family School of Nursing at Western University, and holds an adjunct professor position at the Institute of Health Policy, Management & Evaluation, University of Toronto. His current research interests include nursing education, social media, health informatics, and socio-technical perspectives.
© 2012 OJIN: The Online Journal of Issues in Nursing
Article published September 30, 2012
Benkler, Y. (2006). The wealth of networks: How social production transforms markets and freedom. New Haven: Yale University Press.
Canadian Council on Learning (CCL), (2007). Health literacy in Canada: Initial results from the international adult literacy and skills survey 2007, Ottawa, ON: Author.
Centers for Disease Control and Prevention. (2009). Social media at CDC. Retrieved from www.cdc.gov/socialmedia/campaigns/h1n1/#Microblogs
Chew, C., & Eysenbach, G. (2010). Pandemics in the age of Twitter: Content analysis of tweets during the 2009 H1N1 outbreak. PLOS ONE, 5(11). doi:10.1371/journal.pone.0014118
Chiang, O. (2011, January 9). Twitter hits nearly 200M accounts, 110M tweets per day, focuses on global expansion. Forbes. [Web Log] Retrieved from http://blogs.forbes.com/oliverchiang/2011/01/19/twitter-hits-nearly-200m-users-110m-tweets-per-day-focuses-on-global-expansion/
Chou, W., Hunt, Y., Beckjord, E. B., Moser, R. P., & Hesse, B. W. (2009). Social media use in the United States: Implications for health communication. Journal of Medical Internet Research, 11(4). doi:10.2196/jmir.1249
Christakis, N. A., & Fowler, J. H. (2010). Social network sensors for early detection of contagious outbreaks. PLOS ONE, 5(9). doi:10.1371/journal.pone.0012948
Chretien, K. C., Azar, J., & Kind, T. (2011). Physicians on twitter. The Journal of the American Medical Association, 305(6), 566-8. doi:10.1001/jama.2011.68
Ebner, M., Lienhardt, C., Rohs, M., & Meyer, I. (2010). Microblogs in higher education – A chance to facilitate informal and process-oriented learning? Computers & Education, 55(1), 92-100. doi:10.1016/j.compedu.2009.12.006
Estabrook, L., Rainie, Lee, & Witt, E. (2007). Information searches that solve problems. Retrieved from www.pewinternet.org/Reports/2007/Information-Searches-That-Solve-Problems.aspx
Eysenbach, G. (2008). Medicine 2.0: Social networking, collaboration, participation, apomediation, and openness. Journal of Medical Internet Research, 10(3), e22. doi:10.2196/jmir.1030
Fox, S. (2010). Mobile health 2010. Retrieved from www.pewinternet.org/Reports/2010/Mobile-Health-2010.aspx
Fox, S., & Jones, S. (2009). The social life of health information. Retrieved from www.pewinternet.org/Reports/2009/8-The-Social-Life-of-Health-Information.aspx
Golbeck, J., Grimes, J. M., & Rogers, A. (2010). Twitter use by the U.S. Congress. Journal of the American Society for Information Science and Technology, 61(8), 1612-1621. doi:10.1002/asi.21344
Gray, C. (2011). Social media: A guide for researchers. Research Information Network. Retrieved from www.rin.ac.uk/our-work/communicating-and-disseminating-research/social-media-guide-researchers
Hampton, K., Goulet, L., Rainie, L, & Purcell, K. (2011). Social networking sites and our lives. Retrieved from http://pewinternet.org/Reports/2011/Technology-and-social-networks.aspx
Honeycutt, C., & Herring, S. C. (2009). Beyond microblogging: Conversation and collaboration via Twitter. 2009 42nd Hawaii International Conference on System Sciences (pp. 1-10). IEEE. doi:10.1109/HICSS.2009.89
Horrigan, J., & Rainie, Lee. (2006). The internet’s growing role in life’s major moments. Retrieved from www.pewinternet.org/Reports/2006/The-Internets-Growing-Role-in-Lifes-Major-Moments.aspx
Hsieh, H.-fang, & Shannon, S. E. (2005). Three approaches to qualitative content analysis. Qualitative Health Research, 15(9), 1277-88. doi:10.1177/1049732305276687
Human Resources Development Canada. (2003). Reading the future: A portrait of literacy in Canada – highlights. Ottawa, ON: Author.
Huberman, B. A., Romero, D. M., & Wu, F. (2009). Social networks that matter: Twitter under the microscope. First Monday, 14(1). Retrieved from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/viewArticle/2317/2063
Hughes, A. L., & Palen, L. (2009). Twitter adoption and use in mass convergence and emergency events. International Journal of Emergency Management, 6(3/4), 248. doi:10.1504/IJEM.2009.031564
Java, A., Song, X., Finin, T., & Tseng, B. (2007). Why we twitter: Understanding microblogging usage and communities. Proceedings of the 9th WebKDD and 1st SNA-KDD 2007 workshop on Web mining and social network analysis - WebKDD/SNA-KDD ’07 (pp. 56-65). New York, NY: ACM Press. doi:10.1145/1348549.1348556
Kelly, R. (2009). Twitter study. San Antonio, Texas. Retrieved from www.scribd.com/doc/18548460/Pear-Analytics-Twitter-Study-August-2009
Kostkova, P., de Quincey, E., & Jawaheer, G. (2010). The potential of social networks for early warning and outbreak detection systems: The swine flu Twitter study. International Journal of Infectious Diseases, 14, e384-e385. doi:10.1016/j.ijid.2010.02.475
Krefting, L. (1991). Rigor in qualitative research: the assessment of trustworthiness. The American Journal of Occupational Therapy, 45(3), 214-22.
Lenhart, A., & Madden, M. (2007). Social networking websites and teens. Retrieved from http://www.pewinternet.org/Reports/2007/Social-Networking-Websites-and-Teens.aspx
Lenhart, A., Madden, M., Smith, A, & Macgill, A. (2007). Teens and social media. Retrieved from http://www.pewinternet.org/Reports/2007/Teens-and-Social-Media.aspx
Lincoln, Y.S., Lynham,S. A., & Guba, E.G. (2011) Paradigmatic controversies, contradictions, and emerging confluences, revisited. In Lincoln, Y. S, & Guba, E. A. (eds.) The sage handbook of qualitative research (pp. 97-128) (4th ed.). Thousand Oaks, CA: Sage Publications, Inc.
MIX Online. (2011). The archivist: Save and analyze tweets. Retrieved from http://visitmix.com/work/archivist-desktop/
Madden, M. (2010). Older adults and social media. Retrieved from http://pewinternet.org/Reports/2010/Older-Adults-and-Social-Media.aspx
McCandless, D. (2009). Information is beautiful: If twitter was 1000 people... Retrieved from www.informationisbeautiful.net/2009/more-truth-about-twitter/
McNab, C. (2009). What social media offers to health professionals and citizens. Bulletin of the World Health Organization, 87(8), 566-566. doi:10.2471/BLT.09.066712
Microblogging. (2011). Microblogging. Retrieved from http://en.wikipedia.org/wiki/Microblogging
Naaman, M., Boase, J., & Lai, C. (2010). Is it really about me?: Message content in social awareness streams. Proceedings of the 2010 ACM conference on Computer supported cooperative work - CSCW ’10 (p. 189). New York, NY: ACM Press. doi:10.1145/1718918.1718953
National Center for Education Statistics. (2006). A first look at the literacy of America's adults in the 21st century. Retrieved from: http://nces.ed.gov/naal/pdf/2006470.pdf
Institute of Medicine (2004). Health literacy: A prescription to end confusion. Retrieved from, www.nap.edu/catalog/10883.html
O'Connor, B., Balasubramanyan, R., Routledge, B. R., & Smith, N. A. (2010). From tweets to polls: Linking text sentiment to public opinion time series. Proceedings of the International AAAI Conference on Weblogs and Social Media (pp. 1-8). Washington D.C. Retrieved from www.cs.cmu.edu/~rbalasub/publications/oconnor_balasubramanyan_routledge_smith.icwsm2010.tweets_to_polls.pdf
Pew Internet & American Life Project. (2011). What internet users do online. Retrieved http://pewinternet.org/Trend-Data/Online-Activites-Total.aspx
Public Health Agency of Canada. (2001). Towards a common understanding: Clarifying the core concepts of population health. Retrieved from www.phac-aspc.gc.ca/ph-sp/docs/common-commune/vision-eng.php
Public Health Agency of Canada. (2003). What makes Canadians healthy or unhealthy. Retrieved from www.phac-aspc.gc.ca/ph-sp/determinants/determinants-eng.php#income
Raphael, D. (2008). Grasping at straws: A recent history of health promotion in Canada. Critical Public Health, 18(4), 483-495. doi:10.1080/09581590802443604
Raphael, D. (2009). Escaping from the phantom zone: Social determinants of health, public health units and public policy in Canada. Health Promotion International, 24(2), 193-8. doi:10.1093/heapro/dap005
Raphael, D., Curry-Stevens, A., & Bryant, T. (2008). Barriers to addressing the social determinants of health: Insights from the Canadian experience. Health Policy, 88(2-3), 222-35. doi:10.1016/j.healthpol.2008.03.015
Ritterman, J., Osborne, M., & Klein, E. (2009). Using prediction markets and Twitter to predict a swine flu pandemic. 1st International Workshop on Mining Social Media Programme (pp. 1-7). Sevilla, Span. Retrieved from www.socialgamingplatform.com/msm09/proceedings/paper2.pdf
Sample Ward, A. (2011). Citizen tech: Social media in disaster response. Retrieved from http://amysampleward.org/2011/02/19/citizen-tech-social-media-in-disaster-response/
Scanfeld, D., Scanfeld, V., & Larson, E. L. (2010). Dissemination of health information through social networks: Twitter and antibiotics. American Journal of Infection Control, 38(3), 182-8. doi:10.1016/j.ajic.2009.11.004
Smith, Aaron, Schlozman, K., Verba, S., & Brady, H. (2009). The internet and civic engagement. Retrieved from www.pewinternet.org/Reports/2009/15--The-Internet-and-Civic-Engagement.aspx
Social Media. (2012). Social media. Retrieved from http://en.wikipedia.org/wiki/Social_media
Statistics Canada. (2010). Canadian internet use. Retrieved from www.statcan.gc.ca/daily-quotidien/100510/dq100510a-eng.htm
Sullivan, S. J., Schneiders, A. G., Cheang, C.-W., Kitto, E., Lee, H., Redhead, J., Ward, S., et al. (2012). “What’s happening?’ A content analysis of concussion-related traffic on Twitter. British Journal of Sports Medicine, 46(4), 258-63. doi:10.1136/bjsm.2010.080341
Thackeray, R., Neiger, B. L., Smith, A. K., & Van Wagenen, S. B. (2012). Adoption and use of social media among public health departments. BMC Public Health, 12, 242. doi:10.1186/1471-2458-12-242
The Canadian Press. (2009, November 20). Ont. health units seek ways to shorten flu shot lines. CBC News. Retrieved from www.cbc.ca/news/health/story/2009/11/20/h1n1-ontario.html?ref=rss
Underhill, C., & McKeown, L. (2008). Getting a second opinion: Health information and the Internet. Retrieved from www.statcan.gc.ca/pub/82-003-x/2008001/article/10515-eng.htm
Winston, E. R., Dawn Medlin, B., & Adriana Romaniello, B. (2012). An e-patient’s end-user community (EUCY): The value added of social network applications. Computers in Human Behavior, 28(3), 951-957. doi:10.1016/j.chb.2011.12.016
World Health Organization. (1998). Health promotion glossary. Retrieved from www.who.int/hpr/NPH/docs/hp_glossary_en.pdf
Zhao, D., & Rosson, M. B. (2009). How and why people Twitter. Proceedings of the ACM 2009 international conference on Supporting group work - GROUP ’09 (p. 243). New York, NY: ACM Press. doi:10.1145/1531674.1531710