Mishori R, Levy B, Donvan B. Twitter Use at a Family Medicine Conference: Analyzing #STFM13. Fam Med 2014;46(8):608-614.
Twitter Use at a Family Medicine Conference: Analyzing #STFM13
Ranit Mishori, MD, MHS; Brendan Levy, MD; Benjamin Donvan
Background: The use of social media is expanding in medicine. A few articles sought to describe participant behavior using Twitter at scientific conferences. Family physicians are known as active participants in social media, but their behavior and practices at conferences have not been methodically described.
Methods: We recorded all public tweets at the 2013 Society of Teachers of Family Medicine (STFM) Annual Spring Conference bearing the hashtag #STFM13, using commercially available services. We created a transcript of all tweets for the 5 days of the conference and 3 days before and after. We looked at the total number of tweets, number of original tweets and re-tweets, active users, most prolific users, and impressions. We categorized the content based on (1) Session related, (2) Social, (3) Logistics, (4) Ads, and (5) Other. We compared major metrics (but not content) to the 2012 STFM Annual Spring Conference.
Results: There were a total of 1,818 tweets from 181 user accounts: 13% of the conference registrants. The top tweeter accounted for over 15% of the total tweets, and the top 10 accounted for over 50% of the total volume. Most original tweets (69.7%) were related to session content. Social content came in second (14.2%), followed by other, logistics, and advertisement (7.6%, 6.9%, 1.6%).
Conclusions: This preliminary analysis provides an initial snapshot of twitter activity at a family medicine conference. It may suggest avenues for further inquiry: trend identification, “influencer” identification, and qualitative analysis. Interdisciplinary research should focus on evaluation methods that can assess the quality, value, and impact of tweeting.
(Fam Med 2014;46(8):608-14.)
The Society of Teachers of Family Medicine (STFM) Annual Spring Conference is held annually and is a central event in academic family medicine. It offers participants the opportunity to attend educational sessions and engage in formal dialogue, as well as in informal networking and conversation.
Physicians have been urged to become active participants in social media as a form of engagement with the larger health community, patients, and peers1 and as a way to “start an online dialogue” with policy makers and stakeholders.2 Twitter—the microblogging medium—has been hailed as “an essential tool for every physician leader,”3 one that is “crucial to the development of medicine today,”4 and “just what the doctor ordered.”5 This rapidly growing social network has approximately 500 million users worldwide, 140 million of them in the United States.6
The use of Twitter has been encouraged at various conferences, including at STFM meetings, by organizers and physician leaders,7 as a means of increasing participant connectivity and engagement while also enhancing message dissemination.
A few articles have previously sought to describe participant behavior using Twitter at various scientific and medical conferences.8-16 For example, researchers conducted a content analysis at a large kidney conference, looking at tweets they designated as “informative” (educating the reader about kidney disease) versus “uninformative” and further applying sentiment analysis to the content.17 Other investigators looked at trends of use in scientific meetings. One study looked at meetings held by anesthesia professional associations in the United Kingdom, and noted a 530% increase in the number of tweets posted in 1 year compared to the previous year’s conference.16 A “large growth” in the number of twitter users was also noted at a 2012 hospice and palliative medicine conference,18 but the author noted several barriers to physician use including lack of training and concerns about public online communication. Finally, a large virtual participation—“social” participants as opposed to “physical” participants—was noted at the International Conference on Emergency Medicine (ICEM) 2012.19
As the ICEM 2012 paper noted, increasingly, conference-related dialogue and engagement are taking place not only within the convention hall, but also online, among stakeholders who are not physically present at the conference. Internet tools facilitate in-person interactions while extending the reach of events to those unable to attend and even allowing valuable conversations to persist and grow organically long after the posters are packed. Recent analysis has shown that this is becoming a trend and that at many medical conferences there are now more people participating digitally than are even present on the conference floor.20 In this way, social media can simultaneously offer a reflection of the zeitgeist as well as shape it.
Family physicians are known as active participants in social media,21,22 but their cyber behavior and engagement at family medicine conferences have not yet been documented.
We decided to observe and analyze public tweets coming from the 2013 STFM Annual Spring Conference to see who is talking about the conference and in what ways. This preliminary content analysis may provide an initial snapshot of one aspect of the social media presence of this important family medicine conference for future comparison, trend identification, “influencer” identification, and hypothesis generation.
The STFM Annual Spring Conference took place between May 1 and May 5, 2013 in Baltimore, MD. The hashtag #STFM13 was assigned to this conference by organizers. Prior to the conference, we set up both the HootSuite (www.HootSuite.com) and Hashtracking23 (www.HashTracking.com) services to record every tweet bearing the hashtag #STFM13. We recorded and created a complete transcript of all tweets occurring for the 5 days of the conference and 3 days both before and afterwards (https://www.hashtracking.com/reports/blevy/STFM13/P2h51259).
The complete transcript included the message itself, the sender, and any additional hashtags used, beyond #STFM2013. The hashtracking.com service also calculated several key metrics (number of tweets, number of re-tweets, number of individuals or accounts generating tweets) for us based on the data, including providing a histogram of the users sending Tweets most frequently and those that were most widely disseminated.
We then excluded all re-tweets and manually categorized only the original tweets based on criteria developed in previous studies19 as follows: (1) Session related—all tweets whose content was related to a conference session (quotes, impressions, statements, knowledge gained, for example: “#STFM13 Dr John Saultz—in one year—you can’t change the world—but you can begin conversations, you can change a person or two”), (2) Social—tweets not related to session content; tweets related to social gatherings, meet-ups, new followers, joining twitter, humor, banter, musings, and jokes, for example: “Thanks #STFM13 for a great conference. Shout out to all my new twitter followers from mtg. Time to head home!”, (3) Logistics—tweets informing participants of official STFM events, sessions, and meetings, including room numbers, speakers, and activities, for example: “The 46th STFM Annual Spring Conference starts Wednesday! Follow the #STFM13 hashtag to catch live-tweets and conference highlights,” (4) Advertising—tweets by commercial entities and companies promoting their services and products and job openings. For example: “Need ride to and from the airport? #STFM13 attendees get @SuperShuttle discount when they use the code A626D,” (5) Other—all tweets that do not fall into the above categories (including advocacy, resources, tweets in other languages, links and conversations not related to the conference directly). The third author categorized all the tweets, each of which was later reviewed by the lead author in consultation with the second author in order to reach a consensus on the correct category.
As a form of comparison, we also looked at similar analytics generated by The HealthCare Hashtag Project, Symplur, which generates some data for free and posts it on their web site (Symplur.com).24 Additionally, Symplur had, on their public website, data available about the 2012 STFM Annual Spring Conference, and we compared the Symplur-generated 2012 to the 2013 Annual Spring Conference on key common metrics.
Additionally, to supplement our quantitative data with some qualitative materials, we asked the top tweet generators why they tweet and why tweeting from a conference is important, via direct email inquiries.
The study was deemed “non-reviewable” by our institution’s IRB due to the public availability of the data and the fact that it does not involve human subjects.
In total there were 1,818 tweeted messages from 181 different user accounts. This is compared to the total conference attendance, which is reported by STFM to be approximately 1,370 (personal communication). It is equal to 13% of the total in-person conference registrants.
The majority of tweets were original messages (61%) (Table 1), and over half of all tweets (56%) originated from a group of only 10 users (Table 2).
The top tweeter with the handle @SingingPenDrJen accounted for over 15% of the total tweets, and the top 10 tweeters accounted for over 50% of the total volume.
Most original tweets (69.7%) were related to session content (Figure 1). Social content came in second (14.2%), following by other, logistics, and advertisement, respectively (7.6%, 6.9%, and 1.6%).
The Symplur data (Table 3, Figure 2) shows that there was actually a slight decrease in the number of tweets recorded compared to the previous year (2012) but a slight increase in the number of participants. Over the duration of the event the tweets appear to follow approximately the same pattern with slightly more tweets at the beginning and ending of the conference in 2013.
Impressions measure the maximum possible dissemination of a message. They are calculated by taking the number of tweets sent by each participant and multiplying it by the number of followers that a participant currently has and adding them all together. While it is possible to calculate the number of impressions, calculating their “value” is challenging. There is data from internet marketers that attempt to objectively measure their “value,” showing that the average “value” of an “impression” to company revenue is comparable to paid internet advertising.25 Thus, while there is a way to establish the monetary value of a tweet, it remains to be seen what this means for the propagation of scientific or educational information.
Interestingly, in this case, despite the increase in participants the actual number of impressions went down almost 41% between 2012 and 2013. This may be largely because a small number of highly active and followed users were less active in 2013 or absent from the conference.
Reasons for Using Twitter
Based on the responses of the top tweet generators (n=9—we excluded STFM because they have an obvious reason to tweet from their own meeting; 100% response rate), we identified four main themes: (1) Sharing information with others, particularly those not in attendance, (2) Facilitate networking, (3) Family medicine and medical education advocacy, (4) Note taking and note keeping (Table 4).
This study sought to characterize and describe Twitter activity at one of the major family medicine-
specialty annual meetings. Social media research is a field still in its infancy. Descriptive analyses are the first steps in trying to understand this medium, by looking quantitatively at the messages themselves, their dissemination, the messengers, or the recipients. Such basic data can benefit users and organizers of future meetings and suggest avenues for inquiry and, perhaps, offer a better understanding of participants’ actions and behaviors related to conference activities. This type of information may assist organizations to adjust their social media strategic planning, identify “top influencers” and design advocacy or information dissemination activities. Our study reveals that most participants used the medium to relay content from the scientific sessions. Content delivery included direct quotes, paraphrasing the speaker, photos of the speaker(s), personal commentary related to the content, and speaker encouragement.
Social use was notable as a means of having an informal dialogue between users, making new introductions, welcoming new users, and setting “Tweetups”—face-to-face meetings of twitter users who may have known each other only virtually but now have a chance of translating their social media connection to an in-person engagement. It offers a good example of how Twitter can be used as a “conversation starter” among like-minded individuals and can help create social connections and networking opportunities among people who might otherwise never get to talk to each other. Advertising content, primarily from conference sponsors, vendors, job offers, and transportation services, comprised the smallest portion of the tweets. It would be interesting to follow this trend, as commercial entities and industry sponsors are likely to enhance their efforts to reach a wide audience via Twitter in the coming years.
Around 30% of the total number of generated tweets were re-tweets—original messages penned by somebody other than the user, that the user decided to relay to his or her followers. In previous research, these tweets tend to propagate in a power law distribution, with most tweets not being re-tweeted at all and a small number being re-tweeted thousands of times, with some messages persisting and spreading for as many as eight re-tweeting generations.26 The retweets allow participants to engage in the dialogue albeit in a somewhat more passive way. It may be a suitable, preferred, or even common way for newcomers to partake in the conversation until they feel more confident and adept at this medium to generate their own content. The re-tweeters have the power to propagate the message and they are, in fact, the key to dissemination efforts. When comparing the basic metrics from 2012 to 2013, a few things are noted: (1) The volume of tweets is nearly similar, varying by less than 10% between 2012 and 2013, (2) The number of participants has grown from 166 to 180. Although seemingly a small number, this change has the potential to be very significant depending on the number of followers of each new user. The more followers a user has, the larger the dissemination potential of the message, (3) Impressions—that measure the maximum possible dissemination of a message—decreased between 2012 and 2013. This could be due to the absence of several key participants or cumulative changes in the number of followers across the population. It is important to note that despite the increased number of participants or content generators, the volume of tweets has not followed suit. In fact, the number of tweets per hour and tweets generated per person were larger in 2012: fewer participants generated more content than in 2013. Analysis of individual users and “top generators” or “top influencers” and their sphere of potential influence (eg, number of followers and re-tweets) can shed more light on this matter in the future and help identify key individuals who may serve as the “social authority” when it comes to family medicine or conference-related content.
Our study has several limitations. (1) As far as content is concerned, we only looked at the 2013 STFM Annual Spring Conference and therefore can provide only a brief snapshot of Twitter activity for this one meeting. In order to document trends, and even impact, comparison with future conferences and follow-up studies will need to be carried out, (2) It is extremely hard, with the current analytic methods that are available, to gauge the dissemination of any messages sent and the impact of the tweets. Tweet volume cannot provide much insight about content and impact. While we can analyze what was said and how widely that message was broadcast, it is impossible to tell who is reading the tweets and what action the recipients are taking as a result, (3) We captured only the tweets that bore the official hashtag (#STFM13) of the conference. It is possible that tweet generators sent messages related to the conference without the official hashtag or that some generators used the wrong hashtag for some, or all, their tweets. As a result, we may have missed tweets, which would have affected the accuracy of our content analysis and other statistics.
What is clear, however, is that the use of twitter during medical conferences (in family medicine and outside our specialty) is expanding overall.
Participants vary from year to year, and their likelihood of tweeting is not a constant and may be influenced by multiple issues (time, other activities, Wi-Fi access, technological platform and gadgets, and mood, among others). Several “social participants” were not physically at the conference, which demonstrates, even in this small sample, how traditional hierarchical information dissemination from authority can coexist and engage in a way that allows outside voices, even those physically outside the convention, to participate in dialogue. The importance of technologies that allow degrees of virtual participation is clear, and their role is expanding. Organizations and events that encourage and actively engage in electronic discussions may be able to broaden their reach and attract new followers. It will be interesting and beneficial to examine virtual participation and its effects on the dissemination of content-related materials, the expansion of conference-generated dialogue, and on conference registration and participation.
Interdisciplinary research, both quantitative and qualitative, from fields such as health communication, linguistics, social and computer sciences, tele-health, and others should be developed to design evaluation methods that can begin to assess its quality, value, impact, and outcomes based on tweeting behavior.
Corresponding Author: Address correspondence to Dr Mishori, Department of Family Medicine, Center for Health Communication, Media and Primary Care Preclinical Building GB-01D, Georgetown University School of Medicine, 3900 Reservoir Road NW, Washington, DC 20007. 202-687-3011. email@example.com.
From the Department of Family Medicine, Georgetown University (Dr Mishori); Georgetown University/Providence Hospital Family Medicine Residency, Colman Manor, MD (Dr Levy); and Saint Albans High School, Washington, DC (Mr Donvan).
The Extent and Methods of Public Health Instruction in Family Medicine Clerkships
Jacob Prunuske, MD, MSPH; Linda Chang, PharmD, MPH, BCBH; Ranit Mishori, MD, MHS; Alison Dobbie, MD; Christopher P. Morley, PhD, MA, CAS
Fam Med 2014;46(7):544-548
The Financial Health of Global Health Programs
Winston Liaw, MD, MPH; Andrew Bazemore, MD, MPH; Ranit Mishori, MD, MHS; Philip Diller, MD, PhD; Inis Bardella, MD; Newton Chang, MS
Fam Med 2014;46(9):672-678
Leave a comment on this article
You must complete all fields to comment