Social Media Experiments and IUPESM World Congress 2015

The International Union for Physical and Engineering Sciences in Medicine (IUPESM), in partnership with the International Organization of Medical Physicists (IOMP) and International Federation of Medical and Biological Engineering (IFMBE) run a World Congress every three years and this year (2015), the Canadian Organization of Medical Physicists (COMP) and Canadian Medical and Biological Engineering Society (CMBES) help organize it in Toronto.

After a lot of 'healthy persuasion' I managed to convince the Program Committee that there was a need to have some kind of discussion on Social Media and the Medical Sciences. I quickly realized that providing topics for continuing education generally means that -provided the topic isn't anti-science- the person suggesting a topic becomes the expert in that topic, and hence is left to organize a session on the (seemingly ignored yet important) topic.

I certainly don't profess to be an expert in Social Media, but I do think it can be a fun educational tool. Getting roped into the World Congress Publicity Committee provided an opportunity to experiment and learn about this Social Media thingy as well as generate some potentially useful data for the social media talk that was eventually delivered (that talk is available here on my slideshare account).

One experiment was hand-delivered to me by +Kim Nayyer who learned about this really cool project called Tags Explorer, created by Martin Hawkey (see here for details). Basically, it is a way to visualize who and what is being shared in Twitter through a hashtag. The conference hashtag #wc2015yyz was inserted into a googledoc spreadsheet, and with a bit of finessing, you can track conversations, tweets and retweets from individuals. Moreover, it provides a really clever way of archiving the conversation: perfect for things like conference hashtags where there is a slow ramp-up of activity before the conference, followed by a massive burst during the conference.

Here is the link to the #wc2015yyz TagsExplorer... pretty cool eh?

The timeline of the activity is pretty interesting:

Monday AM

Tuesday AM

Wednesday AM

 

As of today.

All of this is pretty, well... pretty. But what does all this mean? One of the things I learnt was that there were a lot of folks that were mostly paying attention to the tweets of a few prolific tweeters. such as @medicalphysics (as he/she should given the handle!) but another curious thing was the outliers. There were these isolated pockets of tweeters that may have exchanged one or more tweets with each other, but really didn't engage in dialogue with others. Perhaps these individuals didn't know that there were a lot of other activities with the conference... or just chose not to engage in conversation.

Another way of tracking the Twitter activity is through www.symplur.com. In short take a look at the links.
A nice visual summary of the conference: here
The entire transcript of the conference: here

In this conference, there were over 600,000 impressions (that's basically someone read a tweet), and 509 unique tweets, over 100 participants. Given that the actual conference attendance was somewhere in the 2000-2400 mark, that is still a pretty small fraction of the audience. The peak of the Twitter activity took place around June 10, with over 200 tweets! Despite this, the figures suggest that there was a tremendous level of interaction. This is a good thing!

Some stats:
#Unique Tweets per Registrant = 509/2400 = 0.2
#Tweeters per Registrant = 100/2400 = 0.04
#Impressions per Registrant = 600,000/2500 = 240
#Tweets per Tweeters = 509/100 = 5.1

Numbers are cool... can we dig a bit deeper into how these numbers might compare to other conferences?

How does this Twitter engagement compare with other conferences?

A reasonable question to ask is whether this level of engagement is increasing. I dug up some numbers on basically from 2012 onwards when COMP embarked on this social media experiment.

Halifax: COMP 2012

Just several years ago, COMP tried to ramp up the Twitter presence but it didn't gain much traction.
COMP tried 'live tweeting' at the COMP Annual Scientific Conference back in 2012, which I archived on a Storify account. I don't think medical physics was ready for Twitter and things like Storify back then. But the trajectory is positive, certainly with this conference. I didn't register the conference hashtag with Symplur so I can't draw any conclusions.

Montreal: COMP CARO 2013

The year of 2013 showed a bit of improvement. This conference was a joint one with our Radiation Oncologist colleagues (CARO) and also had good representation from Radiation Therapists / Technologists. Here are the stats from Symplur.
Visual Summary:  here

#COMPCARO13 had about 900 impressions, 172 original tweets and 18 participants. This was actually a pretty poor showing given that (memory serves) the attendance was around 500.

Some stats:
#Unique Tweets per Registrant = 172/500 = 0.3
#Tweeters per Registrant = 18/500 = 0.04
#Impressions per Registrant = 900/500=1.8
#Tweets per Tweeters = 172/18=9.6

Banff: COMP 2014

In Banff, the COMP ASM 2014 hashtag monitored on symplur where there were about 200 participants (Caveat: guess again who was tweeting for COMP)
Visual Summary: here

#COMP2014 had about 78000 impressions, 122 tweets and 21 participants. The number of participants was somewhere (really guesstimating here!) 200-300.

Some stats:
#Unique Tweets per Registrant = 122/300 = 0.4
#Tweeters per Registrant = 21/300 = 0.07
#Impressions per Registrant = 78000/300=260
#Tweets per Tweeters = 122/21 = 5.8

ASTRO 2014

Finally, lets look at #ASTRO14. I attended ASTRO and had an opportunity to connect with some really innovative folks using Social Media in the Medical Sciences.
Visual summary: here

#ASTRO14 had 8,685,010 impressions, 3,334 tweets and 549 participants. I have no hard data on the number of attendees but will assume that the number of 11,000 participants as advertised on their website is accurate.
Some stats:
#Unique Tweets per Registrant = 3334/11000 = 0.3
#Tweeters per Registrant = 549/11000 = 0.05
#Impressions per Registrant = 8685010/11000= 789
#Tweets per Tweeters = 3334/549 = 6.1

Just to remind you, here are the WC numbers again:
#Unique Tweets per Registrant = 509/2400 = 0.2
#Tweeters per Registrant = 100/2400 = 0.04
#Impressions per Registrant = 600,000/2500 = 240
#Tweets per Tweeters = 509/100 = 5.1

Please forgive me for the lack of error bars... but as any good physicist would tell you, you should be able to generate some pretty good conclusions if your data is close to ~5%... or within the margin of error if simply 'back of the envelope' calcluations (hence the lack of significant figures in the numbers I report!).

So what to make of all this? Has the social media community within medicine, particularly the radiation oncology, growing? Frankly, I'm not so sure. It might be true that the WC2015 meeting data might be strongly influenced by the percentage of biomedical engineers... but I refuse to believe that biomedical engineers are any less (or more of that matter) savvy in Social Media.

Have we saturated? Or are we on to something?

Please feel free to leave comments.