Tag Archives: ontology

Hashtags for Twitter Cancer Communities

Top ten hashtags associated with #cancer

Librarians have been geeking out, or grossing out, over hashtags since they first appeared. Some of the conversation has been about concerns over ‘allowing’ the public to define their own metadata, while much of it has been the flip side of trying to engage the public in generating metadata for library online collections, and thus enriching access and awareness for those collections.

Naturally, the general public simply move forward with creating new hashtags for their own purposes, largely unaware of the conversations and concerns of professionals in the area of metadata. This is as it should be. The idea of a Folksonomy, a.k.a. folk taxonomy, as originated by Thomas Vanderwal centers around the social aspect — real people, real folk, coming up with language that means something to them to describe content that matters to them with ideas that matter to them. Meaning.

I could go on about this for a long time, but today I need to focus on a particular aspect of this dynamic — a shift from folk+taxonomy to folk+ontology. Folkology? Folk ontology? Folktology? A little bit of digging leads me to folktology (non-scholarly) or tagontology (scholarly) as preferred terms for this, both of which are used roughly the same amount.

In social media, one of the greatest strengths has been the power to create community where none existed before, to connect and empower those who may otherwise be isolated. The most prominent examples of this in healthcare have been the emerging communities around chronic conditions (such as diabetes), marginalized communities (such as facial difference and transgendered), and conditions that create isolation as part of the lifestyle or treatment of the condition (such as mobility disorders, many types of cancer, and any condition expected to be fatal).

Taking cancers as an example, there is the immediate problem of the ambiguity of language. In the image at the head of this post, the hashtag #cancer is shown to be most often associated with the Zodiac, not with healthcare. This makes that term itself less useful for healthcare uses.

Symplur Healthcare Hashtag Project 07082013
Symplur: The Healthcare Hashtag Project: http://www.symplur.com/healthcare-hashtags/

In the Symplur Healthcare Hashtag Project, a crowdsourced collection of hashtags in health, there are over 2500 hashtags total, with over 100 (n=133 07/08/2013) related to cancer. These range from disease tags, to events, to scheduled chats, and more. When people enter a new tag, they cannot do so anonymously, and the tags are reviewed before being added to the database. The tag donor is also asked to define the tag category at time of submission. Non-event tags must be able to show that they are used by multiple people. All of this makes the quality of the collection superior to most hashtag databases on the web. (I often wish there was something similar for science hashtags, or information technology hashtags, etc. I also often wish that the project content was routinely archived for posterity through a neutral organization, such as a library, but that is another conversation to have.)

The problem? Not one of those 133 hashtags on cancer is the hashtag #cancer. Of course, it would be really messy to try to separate the zodiac hashtags from the health hashtags, so I can understand why it has been avoided. However, this problem of the commonly used hashtag being missing from the database occurs fairly regularly. It is a not unexpected problem with crowdsourced information collections. Here’s another example. According to Symplur, the preferred hashtag for ovarian cancer is #ovariancancer. If you actually prowl around Twitter, there is an enormous variety of tags used, with the most common being #ovca. The #ovca content is not currently being captured, tracked, or archived in the project database. I just this morning submitted the #ovca tag when I noticed it was lacking. Hopefully, it may be active by the time this post goes live, but the content in it would be sparse and would lack history.

Here are the top, ie. most common, cancer hashtags, according to Symplur.

Cancer:
#BCSM; #BladderCancer; #BowelCancer; #BrainCancer; #BreastCancer; #CancerChat; #CancerFreeMe; #CancerSurvivors; #CervicalCancer; #Chemo; #ChildhoodCancer; #ColonCancer; #Leukemia; #LiverCancer; #LungCancer; #Lyphoma; #Melanoma; #Mesothelioma; #OralCancer; #OvarianCancer; #PancreaticCancer; #PediatricCancer; #ProstateCancer; #SkinCancer; #TesticularCancer; #XMRV
Symplur: The Healthcare Hashtag Project: http://www.symplur.com/healthcare-hashtags/diseases/

You’ll notice a wide variety of types of tags, with a general approach tending toward long tags that include the full words. In actual practice on Twitter, this is the reverse of standard practice, in which tags are kept short to minimize the number of characters used. Many of these tags, like #OvarianCancer, have shorter alternatives that are also used heavily (ie. #ovca). For breast cancer, both forms appear in the Symplur list: #BreastCancer and #BCSM. #BrCa, however, was missing, just like #OvCa. I submitted it, also.

You see the problem? Problems, actually. Part of it is discovery of the terms used, part of it is the actual terms used, and part of it is the community working to ‘manage’ creation, use, and adoption of the terms. Enter @SubatomicDoc, a.k.a. Dr. Matthew Katz. Matthew is a radiation oncologist who has been active in a couple different Twitter cancer communities, most notably #BCSM (which he adopted) and #LCSM (which he initiated). #BCSM stands for breast cancer social media, and #LCSM stands for lung cancer social media. The process of coming up with a better hashtag for lung cancer, gathering a community around it, and developing traction and adoption, got him thinking. What about other cancers?

Matthew sent me a direct message last week about this. He’d been thinking, and had created a rough draft of what he is calling a folksonomy, but which is really more of an ontology, uh, folktology or tag-ontology. We went back and forth several times, thinking about metadata design, automated sorting in computers, common usage, structuring subconcepts, distinguishing proposed tags from currently used tags in other domains, and various other ideas of how to best structure these in a way that would be useful, practical, and true to the concepts and communities. Matthew released the initial draft at the ASCO site last week, with a substantial model integrating proposed and existing Twitter hashtags around cancer experiences and communities.

Matthew S. Katz, MD. Hashtag Folksonomy for Cancer Communities on Twitter. ASCO Connection: 03 Jul 2013 9:08 AM http://connection.asco.org/Commentary/Article/ID/3590/Hashtag-Folksonomy-for-Cancer-Communities-on-Twitter.aspx

Since then there has been a lot of reaction, with people asking for MORE. Frankly, that is not a reaction I think either of us expected. There are refinements and extensions evolving from the communities. It is becoming a richer and broader conversation. I’d like to see more medical librarians engage with this. I am no metadata specialist, and would love to see someone get interested who is more expert than I am with metadata.

One of the extensions that was proposed through Twitter conversations around this is the idea of secondary tags to connect common cancer issues with specific cancer communities. I’d roughed out a list of some of those issues for my book chapter for online cancer resources and search strategies, back in the MLA Guide.

CAM, biopsy, staging, caregiving, home care, chemotherapy, cancer medications, side effects, clinical trials, fatigue, new diagnosis, nutrition, diet, pain, prevention, lifestyle, second opinions, sexuality, survival, and talking about cancer to different audiences.
MLA Guide: Free Samples: Sample Chapters: Volume Two: Diseases and Disorders: Part IV: Cancers, by P. F. Anderson http://www-personal.umich.edu/~pfa/mlaguide/free/cancers.html#issues

One of the ideas Matthew is talking about is how to come up with a strategy for creating new hashtags that would open this up to others, what are the criteria or best practices for creating new hashtags. I did some thinking on this for my Enriching Scholarship workshop on Twitter Hashtags for Science.

Twitter Hashtags

Twitter hashtags mindmap: http://www.mindmeister.com/270101756/twitter-hashtags-by-pf-anderson

I should make a separate post about the model I developed for thinking through best practices of creating new hashtags, but I’ll just put a placeholder here. The acronym is LUDDITE, which stands for:

Length
Unique
Distinct
Decipherable
Indelible
Time
Enterprise

LUDDITE Model for Hashtag Creation

These overlap in many key points with Matthew’s criteria in his ASCO post, however he includes critical points of working specifically for cancer and healthcare communities.

“It is disease-based;
It helps patients with similar diagnoses learn and share rather than be isolated by the cancer experience;
It is designed to make information more easily accessible;
It is unique enough to be distinguished from other topics online;
Brevity is key to allow more content/conversation, especially with Twitter.”

So, that’s as far as we’ve gotten, but we’d love YOU to join the conversation and thoughts around this. Please put comments about the hashtag model at Matthew’s post, and comments about the process here. Thank you so much!


UPDATE July 15, 2013.

An important followup post from Matthew (@subatomicdoc) is now up.

Cancer Hashtags: High Time or Half-Baked?
Matthew S. Katz, MD
15 Jul 2013 10:09 AM http://connection.asco.org/Commentary/Article/ID/3599/Cancer-Hashtags-High-Time-or-Half-Baked.aspx