Article Text

Download PDFPDF

‘What's happening?’ A content analysis of concussion-related traffic on Twitter
  1. S John Sullivan1,
  2. Anthony G Schneiders1,
  3. Choon-Wi Cheang1,
  4. Emma Kitto1,
  5. Hopin Lee1,
  6. Jason Redhead1,
  7. Sarah Ward1,
  8. Osman H Ahmed1,
  9. Paul R McCrory2
  1. 1Centre for Physiotherapy Research, University of Otago, Dunedin, New Zealand
  2. 2Centre for Health, Exercise and Sports Medicine, University of Melbourne, Melbourne, Australia
  1. Correspondence to Professor S John Sullivan, School of Physiotherapy,University of Otago, PO Box 56, Dunedin 9054, New Zealand; sjohn.sullivan{at}otago.ac.nz

Abstract

Background Twitter is a rapidly growing social networking site (SNS) with approximately 124 million users worldwide. Twitter allows users to post brief messages (‘tweets’) online, on a range of everyday topics including those dealing with health and wellbeing. Currently, little is known about how tweets are used to convey information relating to specific injuries, such as concussion, that commonly occur in youth sports.

Objective The purpose of this study was to analyse the online content of concussion-related tweets on the SNS Twitter, to determine the concept and context of mild traumatic brain injury as it relates to an online population.

Study design A prospective observational study using content analysis.

Methods Twitter traffic was investigated over a 7-day period in July 2010, using eight concussion-related search terms. From the 3488 tweets identified, 1000 were randomly selected and independently analysed using a customised coding scheme to determine major content themes.

Results The most frequent theme was ‘news’ (33%) followed by ‘sharing personal information/situation’ (27%) and ‘inferred management’ (13%). Demographic data were available for 60% of the sample, with the majority of tweets (82%) originating from the USA, followed by Asia (5%) and the UK (4.5%).

Conclusion This study highlights the capacity of Twitter to serve as a powerful broadcast medium for sports concussion information and education.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The evolution of Web 2.0 represents the ‘next wave’ of the worldwide web (internet) and has refined online interaction by facilitating active collaboration, sharing and distribution of information.1 This development has seen the internet progress from passively providing information, to becoming a utility that allows users to actively create content and facilitate knowledge transfer in an open and interactive manner.2

Current Web 2.0 technologies, including social networking sites (SNSs), enable users to communicate in real time and collaboratively exchange information, opinions and comments.3 One of the most popular global SNSs is Twitter,4 a microblog application launched in 2006 that allows users to communicate through the quick exchange of short posts (‘tweets’) of 140 characters in length, via web, email, instant message or short message service. Twitter's universal catch-cry ‘What's happening?’ rallies users to tweet and interact with each other through solicited replies or direct messages, and information can be further disseminated via ‘re-tweeting’, or the forwarding of specific tweets.5 Similar to other analogous SNSs, such as LinkedIN and Facebook,6 there has been an exponential increase in Twitter use since its inception. By May 2010, Twitter was transmitting approximately 55 million tweets per day7 from a reported 124 million users worldwide.8

Within a health context, SNSs have the potential to facilitate collaboration between patients, caregivers, health professionals and researchers through online communication.3 A recent survey indicated that 61% of American adults look online for health information which is an increase from 25% in 2000, with 12% of online health information seekers reported to having used Twitter to share updates about themselves or others.9 Within a public health context, Twitter has been used to gain information about the spread of pandemics such as the recent ‘Swine flu’ (H1N1) outbreak,10 to analyse the content of the public's understanding of antibiotics and their possible misuse,11 and has been suggested as a means of improving quality of life for people with chronic conditions.12 Through Facebook, other areas of health such as brain concussion have recently been investigated.13 Ahmed et al reported that Facebook discussion groups are being utilised to relate personal experiences and seek explicit information regarding concussion.13 The importance of sharing and disseminating knowledge regarding sports concussion was highlighted at the 2008 International Conference on Concussion in Sport,14 where it was intimated that the sharing and disseminating of knowledge may be the ‘missing link’ for better education and subsequent management of concussion.15

Recently, sports-related concussion has been brought to the forefront of the public's attention with extensive media coverage of high-profile athletes and their experiences following concussive brain injury.16 Concussion is considered a major health problem,17 18 with 1.6 to 3.8 million concussions occurring annually from sports injuries in the USA alone.19 Most concussions resolve uneventfully without long-term consequences,14 however, in some cases symptoms persist and require comprehensive medical management. The medical management of players with this injury is critical in order for the player to return to sports and activities in a safe manner,14 20 and to minimise the potential for long-term effects.21 Concussions can be difficult to diagnose or identify,22 and often go unreported.23,,25 Thus, persons not receiving appropriate medical advice about the management of their symptoms may seek information from diverse sources, including the internet. This haphazard approach to understanding and managing a concussion may result in a premature return to play. It is important that healthcare professionals are aware of the quality and accuracy of concussion information that the public might be exposed to via various SNSs (including Twitter), as health information on the internet has been shown to be of a varying quality.26

The purpose of this study was to explore, using content analysis, what Twitter users are communicating about concussion and how the term is used within this rapidly expanding communication medium. It is expected that this insight into web-users' perceptions and understanding about concussion could determine whether Twitter could serve as an educational tool through which accurate information related to concussion can be disseminated to the public.

Method

Study design

A qualitative content analysis methodology was used to analyse publically available concussion-related tweets posted on the Twitter website. No attempt was made to contact or interact with the users.

Data source and search strategy

Twitter is a SNS that allows individuals to communicate through ‘tweets’. The Twitter website (http://www.twitter.com) contains a dedicated search function which allows the user to search for specific words and combinations of words contained in the tweets.27 ‘Hash tags’ (#) are a Twitter-specific version of a web-tag used to group tweets into particular categories and contextualise search terms (eg, #concussion).28 Selected demographic information of the user and the number of people who follow them (‘followers’) is also accessible, if made public by the user.

In order to understand the functionality of Twitter's onboard search engine, a variety of concussion-related search terms including concussion, brain injury, traumatic brain injury and sports concussion were trialled. This exploratory exercise identified the term ‘concussion’ as yielding the most substantial number of results. As Twitter does not support a truncation function, the advanced search strategy capability of the website was utilised and subsequently eight search terms were used in this study: ‘concussion’; ‘concussions’; ‘concuss’; ‘concussed’; and the same four terms with a hash tag preceding each word (‘#concussion’, ‘#concussions’, ‘#concuss’ and ‘#concussed’). The Boolean operator ‘OR’ was used between each search word.

Development of coding scheme/categories

A coding scheme was developed with the aim of extracting and recording the key emergent themes relating to how the concept of concussion was being used in each tweet. A general inductive approach29 was used with a preliminary dataset, collected over a 24 h period, to develop an initial set of themes. An integral part of this process was the establishment of a definition and inclusion/exclusion criteria for each theme. This was supplemented by the selection of key quotes as examples of the documented themes. In order to verify the applicability and accuracy of the draft coding scheme, an initial trial was conducted on 100 randomised tweets from a sample collected over a further 24 h period. These were independently coded by five members of the research team. The resulting analysis and discussion led to a revision of the coding scheme and modification of the associated documentation (code book). The coding scheme is shown in table 1. A final trial was then conducted on 100 previously unused tweets selected at random from our databank.

Table 1

Coding scheme with inclusion and exclusion criteria

Data set and sampling

Data were collected over seven consecutive days (12:00, 23 July 2010 to 12:00, 30 July 2010 GMT) by monitoring the Twitter website and downloading all concussion-related tweets posted during this period into a Microsoft Office Word 2007 document for subsequent analysis. This represented a sampling period of convenience. All tweets written in languages other than English were excluded from the dataset, as were tweets which did not clearly contain any of the search terms. From this dataset, 1000 tweets were selected for analysis using a random number generator, representing an estimated sample size of 30%. The randomisation process took into account posts which had been used in the development of the codebook, so as to eliminate the element of familiarity.

Procedure and data analysis

The 1000 tweets were independently coded by three randomly selected members of the research team (SW, HL and EK) over a 2-day period to provide focus and to reduce fatigue. Up to three themes could be assigned to each tweet. As each tweet can only contain 140 characters there was no indication that more than three themes were necessary. Each tweet was analysed in isolation from previous tweets, and from information contained in attached links. Where perfect agreement between the three researchers was not reached, the tweet was discussed and a final coding was agreed upon by consensus. If necessary, another member of the team (JR) would act as an arbitrator to facilitate a final decision. The data were entered into a Microsoft Office Excel 2007 spreadsheet for analysis. The primary analysis was the generation of frequency counts. In addition, example tweets illustrating the selected themes were extracted verbatim.

Results

A total of 3488 tweets were collected over the designated 1-week sampling period: four tweets were excluded as they did not meet the inclusion criteria, resulting in 3484 tweets being included in the dataset to be analysed. From the dataset, 1000 tweets were randomly selected for analysis. These tweets were transmitted by 692 individuals and 277 groups, with 31 tweets unable to be classified. The two categories, individuals and groups, had an average of 778 and 2235 followers respectively. The age and sex of the individuals could not be ascertained.

A total of 1409 codes were assigned to the 1000 tweets. There was unanimous agreement between the three coders for 888 (63%) of the assigned codes. The frequencies of codes for each theme, together with the example quotes, are presented in table 2. The predominant theme was ‘news’ (n=468, 33%), which largely encompassed stories of Major League Baseball (MLB) players who had sustained concussive injuries, and the recent launch of a concussion awareness poster by the National Football League (NFL).28

Table 2

The frequency and percentage of each theme with examples*

The next most prevalent theme was ‘sharing personal information/situation’ (n=377, 27%), which included individuals sharing thoughts and experiences such as ‘I hit it on a shelf, almost knocked myself out:/Concussion.. not fun’. ‘Inferred management' (n=181, 13%) was the third most often coded theme, where elements of concussion management were directly or indirectly inferred. Within this theme, a number of tweets (n=14, 8%) contained information with reference to sleep and the need to stay awake during the acute phase of a concussion.

Information on the geographical location from where the Twitter account was registered was available for 600 (60%) of the 1000 tweets and indicated that the majority of users were from the USA (n=493, 82%), with a smaller number from Asia (n=30, 5%). Full geographical information is presented in table 3.

Table 3

Geographical information by region

Discussion

This novel analysis of the SNS Twitter provides a snapshot of how Twitter users are talking about concussion in their everyday lives through online communications. It builds on a recent study which found that the majority of users were sharing or seeking information regarding concussion on the SNS, Facebook.13 The tweets in this current study's sample were generated by a mix of individuals (69%) and organisations (28%). These tweets provided an initial insight and understanding of how people use the term ‘concussion’, and how the medium of social networking is being used to propagate injury information in health communication.

‘News’-related tweets (33%) were the predominant category of transmitted information. There were several key events during the data sampling period which influenced the number of tweets generated and most probably increased the frequency of news-related tweets. Users tweeted and re-tweeted about a number of players in MLB, a major US professional sporting organisation, who had suffered concussive brain injuries before and during the data sampling period. Also, during this period, the NFL launched a concussion awareness poster,30 which caused the biggest spike in tweet numbers collected on 27 July 2010 (GMT). This important event got worldwide coverage via Twitter and illustrates the potential use of Twitter as a viral marketing tool for the dissemination of sports injury prevention information. The tweets mirrored what was happening in the more traditional media sources, and often have attached links back to these primary sources. This immediate real-time response to global news and events could see Twitter emerging as a key broadcast medium for news and information about major health affairs, as has been shown with the Swine flu pandemic.10

The internet is now second only to television as the most popular news platform.31 Through Web 2.0, social networking allows individuals to play an active role in reporting and disseminating news by allowing the sharing of rapid real-time information.32 News is now becoming more of a social experience where individuals share, discuss and analyse stories and links.31 Web 2.0 sites such as Twitter have been proposed as potential ‘apomediators’ (gatekeepers) to trustworthy medical information for users.3 In the case of sports injuries, Twitter has the potential to rapidly disseminate information, be it best practice, of genuine importance or otherwise.

The second most commonly identified theme from this study was ‘sharing personal information/situation’ (26.76%), as exemplified by tweets such as; ‘got a slight concussion today *sigh*’. This typifies the general user intentions of Twitter33 and other blog sites.34 Twitter serves as an outlet for users' thoughts and opinions, and at times possibly provides them with a cathartic experience. People may also find the social aspect of communicating and sharing ideas with people facing similar health issues an appealing function of Twitter, a phenomenon that has previously been termed ‘iSupport’ with respect to a similar virtual interaction observed in concussion groups on Facebook.13 This type of behaviour has been shown to create a sense of community35 and can deliver psychological benefits to the contributing users36 and to their ‘followers’ as well as to ‘lurkers’.34 37 38 This latter group refers to persons who do not actively tweet but have access to publicised tweets and monitor transmissions. Although not specifically considered in this study, followers and lurkers may also benefit from the sharing of information, and the potential impact of concussion-related content for this group cannot be ignored.

The third most prevalent theme was ‘inferred management’ (13%), where tweets containing information on seeking medical intervention, hospitalisation and discontinuation of play were commonly depicted. Within this theme, information that could be considered misinformation was noted in the analysis of the tweets. Some 8% of tweets coded as ‘inferred management’ contained misconceptions regarding sleep in the acute stages following a concussion, for example: ‘Not suppose to sleep after concussions:/even though they make u want 2 sleep. Cuz said to stay up for a couple more hours’. While there is a range of information available on the internet and public resources which suggest a range of practices/advice,39,,43 there is no conclusive evidence regarding the effect of sleep following concussion. This issue did not warrant inclusion in the recent consensus statement on concussion.14 As tweets are not moderated for their content and quality of information, the spread of misinformation through Twitter highlights the potential for mismanagement of this injury, which could lead to unfavourable outcomes.

It is clear that the general public,44 45 athletes,46 coaches47 48 and non-expert medical personnel49 50 still maintain misconceptions regarding concussion in general, however there have been considerable improvements in the public's general knowledge about brain injury44 since 1988.51 The recent concussion consensus statement highlights the need for better education and appropriate methods of knowledge transfer.14 15 It has been suggested that popular media sources and national multimedia campaigns have played an effective role in educating the public regarding brain injury.44 Traditional forms of knowledge dissemination such as posters, fact sheets, wallet cards and other conventional resources have been used by the Centers for Disease Control and Prevention as a foundation for the Heads Up: Concussion in Youth Sports campaign.52 53 Nevertheless, inaccurate information and misconceptions surrounding certain aspects of brain injury are still prevalent within today's society.44 45 It is of concern that there were a number of tweets classified under the theme ‘downplay’, which appeared to indicate an indifference to the potential seriousness of a brain concussion. Furthermore, specific evidence for the appropriateness and the effectiveness of these traditional methods for concussion education currently does not exist.15

With the evolution of Web 2.0 and the emergence of eHealth,54 there are now new means for information transfer with greater potential to reach a larger group of people. Given Twitter's potential power for channelling information and disseminating knowledge, this SNS is an ideal forum for organisations to proactively transmit concussion-related information in order to increase awareness, promote best management practices and rectify misconceptions associated with this injury. Sports medicine journals such as the British Journal of Sports Medicine55 and the Clinical Journal of Sport Medicine56 are now transmitting tweets about their content and related material. While the number of registered ‘followers’ is still relatively small, this mechanism has the potential to put peer-reviewed material into the public domain. It also provides the opportunity to tweet and re-tweet landmark documents such as the consensus statement on concussion in sport14 to a wider and potentially different audience to the readers of these journals, whether they are followers, lurkers or persons simply exploring this medium. Quality-controlled knowledge transfer through this new communication medium provides opportunities to augment traditional modalities to educate the general public on concussive brain injuries.

This study is limited in that it only provided a brief snapshot of Twitter traffic. With the rapid expansion of Twitter it is most likely that there is an evolving use of Twitter transmissions. Therefore, this study provides an initial baseline to facilitate the tracking of the changing role of Twitter on eHealth. While the data sample may have been considered biased by the release of the NFL poster which generated a large increase in news-related traffic, this is how the site functions and it allowed an important educational strategy for concussion awareness to be captured and disseminated. A strength of this study is that it examined a larger sample size (29%) than used in a recent similar study (5%) investigating Twitter content.11 Multiple raters were also used to develop the coding scheme and a consensus model was used to reach agreement on the classification of the tweets, thus strengthening the accuracy of our findings.

This initial exploratory study needs to be repeated over a longer sampling period covering multiple sporting seasons. Similar investigations could be carried out in other areas of sports medicine, as there is an ability to yield rich information about the public's knowledge and perceptions regarding this and other sporting injuries.

Conclusion

This study has demonstrated the potential use of Twitter in broadcasting and disseminating general information on concussion. With the public increase in the use of SNSs, and the shift towards online health-seeking behaviour, Twitter may serve as an effective and powerful knowledge transfer medium for concussion education and awareness. Dissemination of misconceptions and misunderstandings were identified in this study, re-emphasising the need for evaluating online information relating to concussion for their integrity and accuracy.

What is already known about this topic?

  • The wide dissemination of information on concussion is a key element in informing sports medicine professionals, players and the general public on the recognition and correct management of this injury.

  • Traditionally, concussion information is disseminated via lectures, websites, pamphlets, videos, posters and the media.

What this study adds

  • There is an active traffic of concussion-related tweets on Twitter.

  • The growing popularity of social networking sites including Twitter provides another potential medium to disseminate concussion information to the general public and health professionals.

  • Twitter with its short ‘tweets’ has the potential to connect with the busy ‘now’ generation and thus provides an ideal mechanism for enhancing knowledge transfer to this information-hungry audience.

Acknowledgments

The authors express their thanks to Andrew Long from the University of Otago Department of Information Science for his guidance on the exploration of Twitter and acknowledge funding support from the International Rugby Board.

References

View Abstract

Footnotes

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer-reviewed.