@RCSI Irl - HEAnet Clarke, Jane Burns... · •Public engagement •Legal and ethical issues @RCSI_Irl . Patients Use of Twitter • Communication/Support with others with same issues

Twitter & Academic Healthcare Content,

More Noise than Signal?

Jane Burns, Research Officer @JMBurns99

Eric Clarke, Lecturer in Heath Informatics @lionsa

Richard Arnett, Associate Director, Quality Enhancement @dr_rtpa

Royal College of Surgeons in Ireland

RCSI Royal College of Surgeons in Ireland Coláiste Ríoga na Máinleá in Éirinn

@RCSI_Irl

• Our findings differ from the

submitted abstract

• A journey into Twitter data

• A shifting exploration of the

twitterverse - immersed in data

• Serendipity at every turn

• Project scope creep

@RCSI_Irl

250 000Tweets

71days

#breastcancer#BRCA#cervicalcancer#womenshealth#menshealth#prostatecancer#testicularcancer

@RCSI_Irl

id_str

from_user

Text

created_at

Time

geo_coordinates

user_lang

in_reply_to_user_id_str

in_reply_to_screen_name

from_user_id_str

in_reply_to_status_id_str

Source

profile_image_url

user_followers_count

user_friends_count

The Partial Anatomy of Tweet

250 000 Tweets x 30 tweet elements

@RCSI_Irl

Tweet analysis

Connection to Twitter API via “R” to collected user profile data

SEO spider to harvest domainPage Titles, Meta Description – Meta Keywords

@RCSI_Irl

Twitter: Connection Like Never before

• How big is Twitter?– Exponential growth since founding in 2006

– Registered users can read and post tweets, but unregistered users can only read them

– Forbes (2014)est. that twitter network hosts more than 284 million active users (MAUs), coupled with over 500 million visitors

– Williams (2013) attempted to classify twitter papers in Pub Medto determine scope in Medical Professions area. 3 categories identified:• Outreach and scope of Twitter messages

• Public engagement

• Legal and ethical issues

@RCSI_Irl

Patients Use of Twitter• Communication/Support with others with same issues

• Self diagnosis

• Investigate for family member or someone they care for

• Alternative Treatments

• Clinical News

• Identify health care supports/specialists

Twitter provides real-time dissemination of news, information, personal accounts and other details via a highly interactive form of social media for Wired Patients – (Sugawara, 2012)

@RCSI_Irl

Twitter & Medical Education Environment

• Participants on Twitter in this Sector– Patients– Medical News (Specialist or Larger News

Organisations)– Medical Professionals – Researchers & Educators– The Ordinary person interested in their own health or

health issues

• Library & Information Professionals– Experience in providing, sourcing and disseminating a

range of sources to all users in this sector.

@RCSI_Irl

Twitter & Library Information Professionals

• Twitter is a tool that many librarians are now familiar with because they utilize it to connect to and communicate with their user communities. (Dalton, 2013)

• Dissemination of information and the integration of new technology to share ideas are fundamental to the information profession(Rooney-Ferris & O’Connor 2015)

• (Increasingly) health sciences journals have social bookmarking tools…allowing users to share journal content on websites such as CiteULike, Facebook, and Twitter (De Groote, S. (et. al) 2014 )

@RCSI_Irl

Twitter- Why Taxonomise?

• Challenge- could a taxonomy be developed that would complement and support the final level of extraction and analysis of the top hashtags identified by linkable URL’s..

• A taxonomy with 9 Primary Subject headings and multiple sub headings was constructed on the visual analysis alone of a url contained in a tweet.

• The reasoning for this was to replicate the experiences of users viewing a tweet in range health categories and to determine the level of trust that existed between the creator and content of an url to be high enough to for a user to retweet?

@RCSI_Irl

Taxonomy Development

#breastcancer#BRCA#cervicalcancer#womenshealth#menshealth#prostatecancer#testicularcancer

c91,000Domains tweeted

c940 unique domains

@RCSI_Irl

Taxonomy Subject Headings

• Health Topics lent themselves to 9 Main Subject headings-categorised at domain levels as identified in urls..

– .edu (education)– .org (non profit, only as indicator)– .gov (USA government)– .ac.uk (UK university or college)

– -.org (generally non profit, but this is no longer the case-its only a possible indicator)

– .com (Commercial)

• Other headings added (but have no identified value)

– Not Available – Other -Url Shortener

@RCSI_Irl

Taxonomy Subject Headings

AcademicAcademic JournalBlogCommercial FundraisingHealth Information – CommercialHealth Information – Non ProfitNews – General News – Science/Medicine/TechnologySearch EngineSocial MediaUSA Government

@RCSI_Irl

Domain categories (91000 domains tweeted)

@RCSI_Irl

@RCSI_Irl

• Each node (dot) is a twitter account

• Edges (lines) between each node indicate a mention or a retweet

• Size and colour of nodes are determined by pagerank

Importance of node determined by the number and relative importance of interactions with other nodes

Visualisation of ‘mentions and retweets’ in 9123 prostate cancer tweets

@RCSI_Irl

N31 = ProstateCancerC – Prostate Cancer Canada.

Dedicated to eliminating the diseasethrough education, research & support.~10500 followers. Mentioned 265 times

(~3% of total PC tweets).

Example of a simple relationship

3 accounts linked by 2 teewts.

@emiliescottmd

@fitzpatkaren

@jamesgaor

RT @jamesgaor: MT @emiliescottmd Latest greatest

advice on Prostate Cancer Screening

http://t.co/a2s0CvNXZF #primarycare #prostatecancer

#H…

Check out my next blog post. Prostate Cancer Screening

http://t.co/SJr3JQz1nz #preventivemedicine

#primarycare #prostatecancer@jamesgaor

@UrologyTimes

@RCSI_Irl

1st degree links - @UrologyTimes

@RCSI_Irl

2nd degree links - @UrologyTimes

@RCSI_Irl

3rd degree links - @UrologyTimes

@RCSI_Irl

@ProstateCancerC

@ProstateUK

@UrologyTimes

@EUPlatinum@dukemedschool

In terms of pagerank, 5 most important accounts

@RCSI_Irl

Total of followers for fiveaccounts is 44207 @ProstateCancerC

@ProstateUK

@UrologyTimes

@EUPlatinum

@dukemedschool

23745.4%

10.0023%

Dukemedschool result is anomalous

The same tweet put out by each of the MDNews accounts for each state

Tweet reporting the results of a Duke study.

What can this tell us?

Total followers: @mcuban (3764605)

Total tweets: @Prostate_Bio (418)

Total retweets: @ProstateCancerC (196)

Retweet : tweet (>5) ratio: @Euplatinum (52:8)

Total links: @ProstateCancerC (301)

Total unique links: @ProstateCancerC (156)

Unique links in: @ProstateCancerC (132)

Unique links out: @zerocancer (42)

Pagerank: @ProstateCancerC (0.013)

@RCSI_Irl

Viewer discretion advised

@RCSI_Irl

Breast cancer network. 94988 tweets.30528 nodes (accounts)86920 edges (links between nodes).

@RCSI_Irl

@aimeefletch82

@susangkomen

@CDC_Cancer

@youngbcblog

@whenbchappens

@RCSI_Irl

A ‘double yolk’ of common mentions between two accounts.

@aimeefletch82 @youngbcblog

Trending data: One tweet, retweeted 1000s of times

@RCSI_Irl

The Tweet

@RCSI_Irl

“The trouble starts when we confuse matters of quantitative value with qualitative value”

Essentially, Twitter, like the economy it is part of, is eating itself: it has become a social pyramid scheme whose enormous strengths are undermined by its own – our own – market-derived metrics, which tell us nothing about the quality of the experience.

@RCSI_Irl

Conclusions

• An attempt to identify and categorise urls shared via twitter

• <10%) amount of content shared is “academic”

• Variations in across hashtags

• Large impact of one tweet/user/mention

@RCSI_Irl

Conclusions

• Complexity of the network, but compartmentalised

• Variety users of motives for participation

• Variety of users– primary producers, pure consumers, hubs to synthesize and

redistribute info

• Constantly evolving relationships

• The need for a strong initial ‘question’ to inform and appropriate methodology and metric for filtering information

@RCSI_Irl

Thank You

ReferencesAlnemer KA (et al) 2015 Are Health-Related Tweets Evidence Based? Review and Analysis of Health-Related Tweets on Twitter. J. Med Internet Res. Oct 29;17(10):e24

Dalton, M. (2013) What Would I Tweet?”: Exploring New Professionals’ Attitudes Towards Twitter as a Tool for Professional Development. Journal of Library Innovation 4(2),101-110.

De Groote, S. (et. al) 2014 Information-seeking behavior and the use of online resources: a snapshot of current health sciences faculty.. J Med Lib Assoc 102(3),169-176

Forbes (Online) 2014 How Big can Twitter Become http://www.forbes.com/sites/greatspeculations/2014/11/21/how-big-can-twitter-become/. [Accessed Nov. 1, 2015].

Greysen, SR (et.al) 2010. Online professionalism and the mirror of social media. J Gen Intern Med. 2010 Nov;25(11):1227-9

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/view/2317/2063

.

.

References.

King, D. (et. al) 2013 Twitter and the health reforms in the English National Health Service. Health Policy. 2013 May;110(2-3):291-7.

Rooney-Ferris, L. and O’Connor M. (2015). “It’s just like passing notes in class…”:a content analysis of the use

of Twitter at #asl2015. An Leabharlann 24(2), 10-17.

Sugawara,Y. (et.al) (2012) Cancer patients on Twitter: a novel patient community on social media.

BMC Research Notes 2012, 5:699

Williams, S. A., Terras, M., & Warwick, C. (2013). How Twitter Is Studied in the Medical Professions: A Classification of Twitter Papers Indexed in PubMed. Medicine 2.0, 2(2), e2. http://doi.org/10.2196/med20.2269