Abstract
Twitter.com is a “micro-blogging” website. Although Twitter use is growing rapidly, little is known about health behavior discussions on this site, even though a majority of messages are publicly available. We retrieved publicly available Twitter messages during a 5-week period in early 2012, searching separately for the terms “Pap smear” and “mammogram.” We used content analysis to code each 140-character message, generating a separate coding framework for each cancer screening term and calculating the frequencies of comments. Using the brief account description, we also coded the author as individual, organization, or news media outlet. There were 203 Pap smear and 271 mammogram messages coded, over three fourths of which were from individual accounts. Overall, 22 % of Pap smear messages and 25 % of mammogram messages discussed personal experiences, including attending appointments, negative sentiment about the procedure, and results. Other messages from both individuals and organizations (8 % Pap smear, 18 % mammogram) promoted screening. About one quarter of the messages expressed personal experiences with cancer screening. This demonstrates that Twitter can be a rich source of information and could be used to design new health-related interventions.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
Introduction
Online social networks provide an opportunity for social interaction through the Internet—a space to share ideas, opinions, and information. Their use has rapidly increased in recent years [1], with evidence to suggest that 62 % of adults worldwide now use social media [2]. In addition, these sites have broad population reach [1, 3, 4]. Because of their ease of use and the speed of information dissemination [5], social media channels have transformed the way society shares ideas and beliefs, news, and information about products and services [6]. The extraordinary potential of rapid electronic communication has been clearly demonstrated not only by recent global events and protests [7] but also real-time public health developments such as tracking flu outbreaks [8, 9].
The micro-blogging site Twitter™ (www.twitter.com) is the second most popular US social networking site behind Facebook. Unlike Facebook, in which content shared between users is often private, the majority of Twitter content is estimated to be publicly available [10], as it represents a mix of news, organizational updates, and personal accounts. A recent report estimates that 15 % of US online adults (more than 20 million individuals) are currently using Twitter in 2012 [11, 12], representing significant growth in recent months (24 and 32 % increases in users in 2010 and 2011, respectively [13]). Numerous health care organizations and campaigns use Twitter as a media platform for health promotion and education, including the Centers for Disease Control and Prevention (username, @CDCgov), Agency for Healthcare Research and Quality (@AHRQNews), and the National Institutes of Health (@NIHforHealth). Despite this surge in use, there has been no research to our knowledge conducted on any health behavior content on Twitter. In order to better utilize online social media sites to deliver health information and design targeted interventions (including those that may reach a more diverse audience compared with traditional communication channels or reach groups not accessible via other channels), understanding existing health communication on sites such as Twitter will be critical.
To address this lack of knowledge about health-related dialogue on Twitter, we conducted an exploratory qualitative analysis of Twitter messages in early 2012. We chose to focus on two cancer screenings—Pap smears and mammograms—due to the demonstrated efficacy of these cancer screening procedures [14]. We sought to ascertain (a) if cancer screening content occurs on the Twitter platform, (b) if there is sufficient volume for meaningful analysis, and (c) the nature of the content.
Methods
Study Setting: Twitter
Twitter is an online social networking site that allows individuals to share information in short messages called “tweets” that are 140 characters or less. Twitter is largely a public forum where users follow real-time information and accounts range from personal (from friends and family to celebrities and politicians) to organizational (including news sources, national associations, and local groups). Networking means following the “feed” (message stream) of other accounts, so that news and/or information is customized within each user’s timeline (Fig. 1). Although the text of each tweet is limited, the content shared is often rich, especially through links to longer stories, entire websites, pictures, and videos.
We evaluated the content of a cross-sectional sample of publicly available tweets during a 5-week period in April and early May 2012. Tweets for our study were collected using two search terms: (1) “pap smear” and (2) “mammogram,” filtered to only show top tweets. Twitter determines top tweets from measurable interactions for a tweet (e.g., clicks, re-tweets, number of followers) in combination with a ranking algorithm that defines the interestingness of a tweet based on how it exceeds expected interactions among a follower group in a given time period (personal communication, Twitter). In this first exploratory examination of twitter content, we chose to limit the analysis to the top tweets to focus on messages that generated increased attention on the site, rather than all content. We conducted four separate searches for each term over the study period and varied the day of the week of the search, as the top tweets are automatically refreshed based on recency of the messages.
Qualitative Coding
After retrieving all of the top tweet content from these searches, we used content analysis [15, 16] to code the discussion categories for each cancer screening type [17]. First, the research team examined 20 % of all tweets and developed an overall coding framework that included four to six major categories for each cancer screening type. Then, two members of the team coded the remaining tweets (AL coded all messages and CL independently coded 40 % of messages for reliability), followed by a joint review of aggregated messages. Specifically, data were compiled in an Excel spreadsheet to directly compare categorization of tweets across coders. Discrepancies were resolved by regular meetings and discussion. If new categories (including sub-categories) emerged, the coding framework was changed and the tweets were reread according to the new structure. In addition, frequencies and representative quotes were determined by category.
Finally, to determine whether the author of the tweet was an individual, a media outlet, or an organization, we used the 140-character description and photograph each account holder provided to classify them. We were able to group all accounts into these categories based on this description, as organizational and news outlet accounts were specific in their descriptions and their use of a logo for their account photograph.
Results
Overall, a total of 203 Pap smear tweets and 271 mammogram tweets were coded. The majority of top tweets were from individual accounts (79 %) compared with organizations (20 %), and few media outlet sources (2 %) (Fig. 2). This varied by screening test, however, as more organizations tweeted about mammography (24 %) than Pap smears (14 %). In addition, the discussion categories were similar across the cancer screening terms, with conversations about personal experiences, promotion of screening behaviors, recommended screening guidelines, and other miscellaneous topics—although the distribution across categories differed by cancer screening type. Table 1 provides detailed information about each discussion category by cancer screening type, including the total number of tweets within the categories (not mutually exclusive) and representative quotes.
Pap Smears
A large proportion of top tweets mentioning Pap smears discussed personal experiences about receipt of this cancer screening test (44 tweets from individuals, 22 % of total). Most often, these were straightforward statements about making or going to a doctor’s appointment (25 tweets). An additional 14 tweets expressed negative comments about the dread or embarrassment around the procedure. For example, comments expressed sarcasm possibly used as a means to deflect apprehension (“Great, now they asked me to come back for a pap smear. Oh the joys of being a woman!”), as well as direct comments about their unease (“lord knows i cant stand a pap smear but that 5 mins of uncomfyness is better than dealing with cancer 4 a lifetime”).
There was another group of top tweets that promoted the use of Pap smear screening (16 tweets, 81 % from individuals). This was most often in the form of calls for women to get screened, such as, “In less time than it takes you to read this, you could’ve done a pap smear. Early detection save[s] lives. Visit your gynae or a clinic today.” A few organizations also sent promotional messages about Pap smear screening (three tweets).
In addition, since we analyzed data following a change in Pap smear screening guidelines from the US Preventive Services Task Force (USPSTF) [18], there were 28 tweets about this topic (14 % of total top tweets, 18 from organizations or media outlet sources). Most of these messages (22 tweets) were links to the longer new stories themselves (e.g., “Women 21 to 65 should get a Pap smear every 3 years, according to #USPSTF recommendation released today”); however, another 2 % (six tweets) were commentary about these new guidelines (e.g., “I don’t believe this ‘you don’t need a pap smear every year’ bs.”).
The single largest remaining category of tweets was Other (103 tweets). Although these were varied in nature, they were most often jokes or insults that used the term “pap smear” in a negative way when describing something else. For example, “I would rather be on a job interview while getting a pap smear than go car shopping.”
Mammograms
The discussion categories of personal experiences, promotion, and screening guidelines were also evident in the mammogram tweets, but with differing contexts. Personal experiences were present in 68 individuals’ tweets or 25 % of all mammogram top tweets. The majority (44 tweets) also referenced making or going to appointments, including seven tweets that had a specific negative sentiment about the procedure and six mentioning dread about the appointment: “That mammogram was brutal! But so necessary.” In addition, a relatively large proportion (15 tweets) specifically mentioned family and friends: “Can everyone pray for my mom? She’s going in for an ultrasound because they found something when she went for a mammogram and I’m terrified.” Finally, these personal experience tweets frequently discussed the results of the mammogram (nine tweets): “Tears of joy—mammogram was clear and perfect! Very grateful and relieved!”
In addition, promotional content was present in the mammogram tweets (48 tweets, 18 % of total). This included personal messages such as, “Ladies, if it’s been more than a year and you’re 40 or older, please schedule a mammogram, because I ♥ you.” Yet in contrast to the Pap smear tweets, 3 of the 25 organization’s promotional tweets specifically referenced a location or a mobile mammogram vans to promote screening (“Mobile #mammography at #[university] [city] clinic today, phone (xxx) xxx–xxxx to schedule your mammogram”).
There was also some discussion about mammography screening guidelines and tests (24 tweets, 9 % of total)—the majority of which were from organizations or media outlet sources. For example, there was reaction to a study released that found that younger high-risk women should be screened (“High risk women may benefit from mammograms starting at age 40”), as well as a discussion of new technology for screening (“Mammogram Plus MRI or Ultrasound Catches More Cancer”).
Finally, the Other category for mammogram tweets (46 tweets, 27 % of total) also included a variety of messages. These included political comments about health reform (“Thanks to HCR [health care reform] my wife just received free mammogram coverage from our health insurance.”), jokes or off-hand comments (“I did ask if that full body scan came with a complimentary mammogram when I was scanned last fall. TSA employees…”), as well as references to celebrities who were publicly being screened for breast cancer.
Discussion
To our knowledge, this is the first examination of cancer screening discussions on a social media site. In this analysis, a large number of top tweets consisted of messages relating to personal experiences with Pap smear and mammogram procedures (22 and 25 % of all top tweets, respectively). This demonstrates that Twitter can be a rich source of real-life health experiences—not limited to events such as flu outbreaks [8, 9] but also including examining personal behaviors and decision making processes. In addition, there was immediate reaction through Twitter about cancer screening guidelines, including comments about the changes in the USPSTF Pap smear recommendations and the age of initiating regular mammography among high-risk women.
The personal experience tweets often expressed dread or negative sentiment about both cancer screening procedures. However, there was more balance of positivity in the mammogram tweets—particularly as they more frequently mentioned support for friends and family members or reported non-cancerous screening results. The negative comments about Pap smears were expected, based on previous literature about the barriers to receipt of this screening procedure. In particular, there was a strong sentiment about dreading an upcoming Pap test and/or comments about how uncomfortable the test could be, which may be directly related to previous literature that has shown that pain perception can be a significant barrier [19]. There was virtually no discussion of Pap smear results in the top tweets. This is consistent with prior work about the potential shame in receiving Pap test results, especially abnormal test results resulting from human papillomavirus and linked to sexual behaviors [20, 21]. Although a substantial proportion of the mammogram tweets also expressed negative sentiment about the comfort of the procedure itself, there were also direct comments conveying anxiety surrounding the anticipation of results—also in line with previous evidence [22, 23].
In addition, there were more organizations tweeting about mammograms compared to Pap smears, and a few mammogram tweets directly mentioned access to mammograms, such as through mobile screening vans or upcoming events. This promotion of community access to cancer screening could be expanded given CDC’s national efforts through the National Breast and Cervical Cancer Early Detection Program (http://www.cdc.gov/cancer/nbccedp/), which provides population-wide access to both screening services for un- or underinsured women. Continued promotion of such services is critical, as previous research has documented the significantly lower rates of both Pap smears and mammograms among women who do not have regular health care visits and/or a usual source of care [24, 25].
Finally, although the personal experience and promotional tweets were well represented among top tweets in our analysis, the Other category for both screening modalities was also common. These tweets demonstrate how people contextualize cancer screening in their lives, but they also create “noise” in analyzing social media messages. As opposed to an informative data “signal” in isolating specific types of social media conversations about personal experiences with cancer screening, these Other tweets introduce sometimes random content. This could present a challenge if future social media text analysis relies solely upon an automated data-mining approach, such as counting total mentions for a keyword. This highlights the need for more in-depth qualitative content analysis when designing further social media studies.
Our study has several limitations to note. First, as this was an exploratory analysis, we did not collect all tweets in a specific timeframe. However, our approach to analyze top tweets over 5 weeks allowed us to focus on those messages that are likely to have higher impact on Twitter (i.e., reach more individuals). In addition, we cannot rule out that the conversations during this timeframe could be different in some way from other points in time. Future work examining additional keywords and tweets over time will be particularly informative. Finally, the tweets were brief in nature and that may limit our ability to interpret the context and meaning. However, we did find rich and personal information across many messages that provided insight into cancer screening perceptions.
Social media has not only revolutionized communication worldwide, but it also holds promise as a new medium for understanding how individuals communicate about health-related behaviors. In particular, a deeper understanding of health-related communication on social media sites like Twitter could aid in (1) understanding real-time reactions to changes in prevention and screening recommendations and (2) the design of new health behavior interventions that deliver messages through these online social networks directly. Engaging individuals through channels in which they are already actively participating could have substantial research and population health implications.
References
Chou WY, Hunt YM, Beckjord EB, Moser RP, Hesse BW (2009) Social media use in the United States: implications for health communication. J Med Internet Res 11(4):e48
The social skinny (2012) 99 New Social Media Stats for 2012, Available at http://thesocialskinny.com/99-new-social-media-stats-for-2012/ Accessed 29 May 2012
Gibbons MC, Casale CR (2010) Reducing disparities in health care quality: the role of health IT in underresourced settings. Med Care Res Rev 67(5 Suppl):155S–162S
Pew Internet & American Life Project (2010) Mobile health 2010. Available from http://pewinternet.org/Reports/2010/Mobile-Health-2010.aspx. Accessed 23 April 2012
Rajani R, Berman DS, Rozanski A (2011) Social networks-are they good for your health? The era of Facebook and Twitter. Qjm Int J Med 104(9):819–820
Hawn C (2009) Take two aspirin and tweet me in the morning: how Twitter, Facebook, and other social media are reshaping health care. Heal Aff 28(2):361–368
Harlow S, Johnson TJ (2011) Overthrowing the protest paradigm? How The New York Times, Global Voices and Twitter covered the Egyptian Revolution. Int J Comm 5:1359–1374
St Louis C, Zorlu G (2012) Can Twitter predict disease outbreaks? BMJ 344:e2353
Signorini A, Segre AM, Polgreen PM (2011) The use of twitter to track levels of disease activity and public concern in the US during the Influenza A H1N1 pandemic. Plos One 6(5):e19467
Moore RJ (2009) Twitter data analysis: an investor’s perspective. Available from http://techcrunch.com/2009/10/05/twitter-data-analysis-an-investors-perspective-2/. Accessed 29 April 2012
eMarketer (2011) U.S. Digital media usage: a snapshot of 2012
Pew Internet & American Life Project (2012) Twitter use 2012: Pew Research Center, Available at http://pewinternet.org/Reports/2012/Twitter-Use-2012.aspx. Accessed 9 July 2012
eMarketer (2012) Facebook’s US user growth slows but Twitter sees double-digit gains
United States Preventive Services Task Force. Recommendations for adults, Available at http://www.uspreventiveservicestaskforce.org/adultrec.htm. Accessed 9 July 2012
Pope C, Mays N (1995) Reaching the parts other methods cannot reach—an introduction to qualitative methods in health and health-services research. Br Med J 311(6996):42–45
Sandelowski M (2000) Whatever happened to qualitative description? Res Nurs Heal 23(4):334–340
Graneheim UH, Lundman B (2004) Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse Educ Today 24(2):105–112
Moyer VA (2012) Screening for cervical cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 156(12):880–891
Hoyo C, Yarnall KS, Skinner CS, Moorman PG, Sellers D, Reid L (2005) Pain predicts non-adherence to Pap smear screening among middle-aged African American women. Prev Med 41(2):439–445
Waller J, Marlow LA, Wardle J (2009) Anticipated shame and worry following an abnormal Pap test result: the impact of information about HPV. Prev Med 48(5):415–419
Kahn JA, Slap GB, Bernstein DI et al (2007) Personal meaning of human papillomavirus and Pap test results in adolescent and young adult women. Health Psychol 26(2):192–200
Hislop TG, Harris SR, Jackson J et al (2002) Satisfaction and anxiety for women during investigation of an abnormal screening mammogram. Breast Canc Res Treat 76(3):245–254
Montgomery M, McCrone SH (2010) Psychological distress associated with the diagnostic phase for suspected breast cancer: systematic review. J Adv Nurs 66(11):2372–2390
Taylor VM, Yasui Y, Nguyen TT et al (2009) Pap smear receipt among Vietnamese immigrants: the importance of health care factors. Ethn Heal 14(6):575–589
Blewett LA, Johnson PJ, Lee B, Scal PB (2008) When a usual source of care and usual provider matter: adult prevention and screening services. J Gen Intern Med 23(9):1354–1360
Acknowledgments
All authors have fulfilled the criteria for authorship established by the International Committee of Medical Journal Editors and approved submission of the manuscript.
Conflict of interest
None of the authors had conflicts of interest. Dr. Sarkar was supported by a AHRQ career development award K08 HS017594. None of the funders had any role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
R. Lyles, C., López, A., Pasick, R. et al. “5 Mins of Uncomfyness Is Better than Dealing with Cancer 4 a Lifetime”: an Exploratory Qualitative Analysis of Cervical and Breast Cancer Screening Dialogue on Twitter. J Canc Educ 28, 127–133 (2013). https://doi.org/10.1007/s13187-012-0432-2
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13187-012-0432-2