On the Pitfalls of Social Media: The Case of Clinical Reader

As libraries feel the need to join the social media landscape to meet a segment of their user population already there, it is useful to step back and get acclimated. There is a pace of information flow that is unlike anything else in the physical world, and a minor incident — be it an ill-advised policy decision or an unfortunate slip of the tongue — can quickly spiral out of your control. And that is probably the key word: control. You don’t, can’t, and won’t have it. It isn’t the nature of this media. “Damage control,” if you want to think of it like that, is honest, sincere, decisive, and quick communication with your users. As a counter example, I offer the case of Clinical Reader.

(What follows in this post are the specific details of the Clinical Reader example. If you are not interested in the specific details, you may want to start by reading the observations and recommendations post instead.)

The Backstory

Clinical Reader started operations earlier this year as point of aggregation for medical information. The service’s about page describes its “beginnings” this way:

Clinical Reader was brought to life in 2009 by a junior doctor and a small group of forward thinking young tech programmers spread across London and Toronto. The conceptualized idea was to manage clinical information overload and deliver relevant news from an authoritative source on a daily basis.

On the technical side, it is a website that syndicates content via RSS/Atom from various journals, news sources, blogs, and social media outlets. There is the appearance of editorial decision making in the form of an “Editor’s Choice” list (with no apparent description of how or why something makes this list), but mostly it seems to be a way to hide the complexity of syndicated content readers from those in the medical profession who could care less about such complexity.

Clinical Reader got some buzz, including a “Check out this site” message from Medical Library Association president Connie Schardt (an endorsement of sorts that was retracted two weeks later).

The Spark of Controversy

The controversy started when librarian Nicole Dettmar posted a message on her personal blog questioning the apparent endorsement of the Clinical Reader service by the National Library of Medicine. Ms. Dettmar is employed by the National Network of Libraries of Medicine divisiona program of the National Library of Medicine (NLM) — although she makes is quite clear that her blog reflects her personal views. From her vantage point, though, she knows that the NLM does not endorse commercial entities. In her words:

Graphic from Eagle Dawg Blog

It is bogus as far as the National Library of Medicine (NLM) is concerned since the U.S. Government doesn’t endorse or grant 5 stars to anything. The NLM Copyright Information page offers more elaboration, bold emphasis mine:
Endorsement: NLM does not endorse or recommend any commercial products, processes, or services. The views and opinions of authors expressed on NLM Web sites do not necessarily state or reflect those of the U.S. Government, and they may not be used for advertising or product endorsement purposes.

Ms. Dettmar also notes the use of two copyrighted, unattributed images:

Clinical Reader also currently uses two copyrighted images on their Partners (specific original source, copyright notice at bottom) and Advertising pages (from somewhere on Signalnoise). A ‘credit’ link to a source doesn’t honor an image copyright. Even free open source images, such as this set of Springer Images, are often limited to usage for noncommerical purposes. Commercial organizations can well afford to purchase or design their own graphics.

The Response

Tweets from ClinicalReader


The first response from Clinical Reader came in the form of four Twitter messages from an account with the userid “ClinicalReader”. It is useful to note, I think, that the response came in the form of Twitter messages rather than a comment on Ms. Dettmar’s blog post or in an e-mail message to her (her e-mail address is listed in her Blogspot profile). This is the first indication that the representatives behind the ClinicalReader Twitter account are not necessarily up-to-speed with the common practices of social networking. They brought the issue into Twitter when it didn’t start there, and there were other, more long form, ways to communicate. (It four tweets for ClinicalReader to get its initial points across.) The many grammatical and spelling errors — well beyond normal Twitter abbreviations and common constructs — suggests that English may not be the primary language of the representatives.

Tweets from eagledawg

Ms. Dettmar started a minute with replies later via Twitter that could be interfiled with the initial four tweets from ClinicalReader — her first tweet at 11:40 in reply to the 11:36 ClinicalReader tweet, and her second tweet at 11:45 in reply to the 11:39 ClinicalReader tweet.

Others Take Notice

At this point others following Ms. Dettmar on Twitter start retweeting her response to the legal threat and an ad hoc community starts to form. (Retweeting is the practice of copying another user’s tweet and pasting it into one of yours. You credit the initial tweet by adding “RT” plus the originator’s userid at the start of your tweet.) Steve Lawson posted a message on his blog summarizing the controversy to that point in time.

Tweets from ClinicalReader

Representatives behind the ClinicalReader account — it becomes apparent soon that there are at least two people that have access to the Twitter account — start to address some of the issues. They address messages to those who are retweeting the earlier messages from Ms. Dettmar and Mr. Lawson. (Oddly, they are not using the “reply” function of Twitter for reply messages and they show a tendency to retweet themselves, as demonstrated in the second to last message of this group.) They are in contact with a Twitter user called ‘signalnoiseart’ about some of the unattributed graphics, and they report that they have permission to continue using them. (There will be other allegations of use of unattributed graphics later in the week.) And, lastly, they show and openness to communication — and admirable trait under the circumstances.

Tweets from ClinicalReader

Someone from the group of representatives using the ClinicalReader Twitter account takes the lead in responding to the issues by saying that he/she will start using his/her initials in messages; we’ll refer to this person by those initials: AA. (This is also evident in the change of client used to post messages into the Twitter stream from Tweetie to TweetDeck, with AA apparently using TweetDeck exclusively.) The fact that gender inclusive grammatical constructs are being used in describing this individual shows one of the remaining mysteries of this episode. (The ClinicalReader Twitter account is later taken over by someone who identifies him-/herself as “Allan Marks” but even nearly a week later observers can’t be sure of the identity of this person.) AA indicates that the images showing endorsement will be removed from the Clinical Reader site, and gives an estimated time for when the changes will be apparent to internet users. Indeed, a few hours later the graphics start to change to remove the implied endorsement of the National Library of Medicine. Although AA indicates that he/she is the founder, AA then makes a classical leadership mistake of blaming a subordinate: “We are keen to engage the twitter community the tweet made by a junior member of the team was poor judgment (AA)”.

At the end of the first day, Luke Rosenberger (userid lukelibrarian) notes what he sees as a discrepancy between Clinical Reader’s desire to positively engage with librarians and language in the Clinical Reader newsletter. Mr. Rosenberger tweets, “good grief http://bit.ly/uniXw ‘wave goodbye to the library journal shelf’ and yet @clinicalreader wants love from librarians?!” (The link, http://bit.ly/uniXw, goes to the current newsletter on the Clinical Reader website; at the time this is being written, that web address is returning a “403 Error because this area is restricted” message.) ClinicalReader responds: “@lukelibrarian Your browser is lagging behind says ‘Make effective use of the library journal shelf’ at http://bit.ly/uniXw” (this message has subsequently been deleted, but is recorded in the archive of ClinicalReader tweets). Mr. Rosenberger responds, “@ClinicalReader i see that you have removed ‘wave goodbye’ in the last few minutes. don’t blame my browser, just admit your error. #fail” The DLTJ author noticed that a copy of the newsletter webpage was still in the Google search engine cache and created an image of the page.

Day Two: One Step Forward, Two Steps Back

Early in the second day Mr. Marks sent Ms. Dettmar an e-mail apology for the legal threats. In his message, which Ms. Dettmar subsequently posted to a public Google Docs document after receiving permission, Mr. Marks acknowledges that “the correct channels were not adhered to in seeking permission to use James White and Bryan Christie’s images.” He also says that he understands the NLM policy and has “removed any suggestion of endorsement.” He goes on to say, “We have had support however from teams at Imperial College, the Guardian and the British Library where we have mutual connections.”

Tweet from bengoldacre

At the end of the first day, Guardian columnist Ben Goldacre posted a tweet warning users away from Clinical Reader. On the second day of the incident, this would become one of the main focal points. As Ms. Dettmar notes in a follow up post, one of the representatives behind the ClinicalReader Twitter account has been deleting tweets, and so this conversation can’t be pieced together in the typical way. Ms. Dettmar, though, had been starting to save messages using a service called QuoteURL; QuoteURL “group[s] different Twitter updates from different people into a single page that has a permanent URL.” The thread of conversation between Mr. Goldacre and ClinicalReader is recorded in four QuoteURL aggregations: one, two, three, and four. Mr. Goldacre asked several questions of ClinicalReader (not all of the responses are linked to in Twitter because the tweets have been deleted):

  • BenGoldacre: can you let me know in what way you are supported by the guardian, as your website claims?: Yes we attended a camp organised by Guardian when we we’re developing our project , they provided excellent advice. We’re in touch with Jemima Kiss who writes a fantastic blog – to feature in Elevator Pitch (she’s on maternity at the mo)
  • BenGoldacre: so youre supported by the guardian in the sense that you went to an event they organised and some people gave you sm advice?; BenGoldacre: so you’re supported by the guardian in the sense that jemima might write about you at some stage in the future?: bit more than that, I sense your still skeptical, but I kinda understand
  • BenGoldacre: that’s great. i guess now i’m intrigued about your name.: I’m sensible, the account belongs to the project Clinical Reader and not anyones personal opinion, lets email correspond (twitter)
  • BenGoldacre: maybe i like to know who’s saying i got stuff wrong. and stuff. in the nicest way. maybe i want to sue you!: See you’re getting into the zone, so will quit while ahead
  • BenGoldacre: i think you need a reality check. if you regret your threat then great. but bizarre to tell people they got the story wrong.: Story is correct and unfortunate must add – for someone of your stature and position just thought you jumped on it
  • BenGoldacre: so nobody else was wrong, you were just bullshitting. thanks for clarifying. great escalation skills on ur part, respect!: RT @ClinicalReader We’re learning fast, also important to be humble & admit mistakes, see @eagledawg twt
  • BenGoldacre: er, look, i wanted to go and work, but that British Library endorsement, did you use their bogs once or something?: Ha ha, okay getting silly, nice interacting with you and look forward to a face to face mtg (twitter)

Day Three: No Steps Forward, Third Step Back

feedstitch-tangle

Tangled Strings graphic from FeedStitch

clinicalreader-tangle

Tangled Strings Graphic from ClinicalReader


On the next day, there were new allegations of improper use of graphics. Specifically, Chris Martin (coldpie on Twitter) reported that an image on the Clinical Reader website bore a striking resemblance to an image from the Feedstitch service. The representatives at ClinicalReader Twitter account said, “hi guys stock image purchased, seems uve been stitched” (with no attributing initials). When prompted by Twitter users Andy Woodworth (userid wawoodworth), Mr. Rosenberger, and the FeedStitch twitter account, the reply was: “can DM details if you wish, thanks (RH)” (“RH” is a new set of initials used on the ClinicalReader account); there are no reports of ClinicalReader following through. The firm behind the FeedStitch service posted a message on their blog giving their version of the provenance of the graphic.

Tweets from ClinicalReader

The discussion is not limited to Twitter and blog posts. Other social media outlets are carrying on the discussion as well. Noted in this arena are FriendFeed postings by Rachel Walden (with five additional comments) and Mr. Lawson (a.k.a. Steve Incandenza on FriendFeed, with 51 additional comments). One delicious.com user added notes to the entry for the ClinicalReader homepage: “Clinical Reader has been called out by medical librarians because of sloppy copyright practices and use of false implied endorsements by NLM & others. To gain credibility as source of authoritative content, CR team needs to tread carefully!” Still, either with a naive sense of how social networking tools work or in an attempt to put a brave face on what is happening, the people behind the ClinicalReader Twitter account post updates that seem to think all of this questioning of the ethics of the service provider is good for business.

Day Four: A Step Forward?

Tweets from ClinicalReader

The fourth day does not start well. There are reports of ClinicalReader fabricating a quote posted in a retweet and making up a retweet entirely.

Then it would seem that some sanity has returned to Clinical Reader. First there is a message that Allan Marks has “taken control of this [Twitter] account” (At this point, Allan’s name appears in the Twitter profile for the ClinicalReader account). Then there is more action to remove the FeedStitch tangled strings graphic from the ClinicalReader site, with apologies to FeedStitch, Mr. Rosenberger, Mr. Goldacre, and others. Mr. Marks thinks that the damage is contained; Mr. Rosenberger is not so sure.

There are more questions about the endorsement of people listed on the ClinicalReader pages. In one case, Dr. Ves Dimov says, “In case you are not sure: tweets and links do not represent endorsement, approval or support…” This is apparently in regards to Dr. Dimov’s inclusion of Clinical Reader on a list of medical RSS aggregation services. Mr. Rosenberger responds, “@DrVes re: tweets as endorsements: @ClinicalReader includes a @vesd tweet under ‘what others are saying’ on this page: http://bit.ly/SZ3SQ” (the shortened URL goes to the Clinical Reader latest newsletter page, which, at the time of writing, returns an “unauthorized” error). To which Dr. Dimov replies, “@lukelibrarian The product interface is still good, the management however has made some very strange decisions.Updated: [list of RSS aggregators].” But Mr. Marks replies, “Hi yes thats correct please liaise with @vesd for confirmation – more than happy to be open thanks” (original tweet now deleted; text taken from archive). Mr. Rosenberger then says, “dude, @vesd = @DrVes, who just tweeted ‘tweets …do not represent endorsement, approval or support’ http://bit.ly/QGIvS” and Mr. Marks replies, “happy to get this removed if @vesd says so luke” (original tweet now deleted; text taken from archive).

On the following day, Medical Library Association president Connie Schardt tweets “RT@ClinicalReader I did NOT endorse your product as President of the Medical Library Association; please remove the misleading endorsement” followed shortly by a mysterious another tweet: “@gabeboldt — I didn’t endorce this product as the President of the Medical Library Association; you gave them a misleading endorsement” – there is nothing in Twitter user gabeboldt‘s timeline from around this time that would account for that message. Ms. Schardt was reacting to an image of a tweet in the “Twitter Buzz” section at the bottom of the “Health Tweets” page. This page is no longer on the Clinical Reader site at the time of writing, but a screen shot of the Google web cache version of the page is available.

The Weekend

On Friday evening, Allan Marks, presumably in control of the ClinicalReader Twitter account, renamed it to “clincal_tweets”. This was first announced by Mr. Rosenberger: “@clinicalreader is now @clinical_tweets – all tweet hx & followers intact, just breaks hx links. slick.” Mr. Rosenberger later notes: “after apparently renaming @clinicalreader as @clinical_tweets, ‘aa’ has now opened new acct w/0 followers named (wait 4 it) @clinicalreader” In actuality, it was someone else: “BTW I’m just a witness to the mess which occurred last week. I saw an opportunity to register this account when it moved to @clinical_tweets” (note that this person uses the Power Twitter client — a client that was not used by anyone previously associated with Clinical Reader).

Some time overnight on Saturday, the original ClinicalReader Twitter account was renamed again from “clinical_tweets” to “amarks7“. We know this because the history of tweet updates and the list of followers remain intact. This is presumably the same Allan Marks that described himself as the co-founder of Clinical Reader. In addition to renaming the account twice, Mr. Marks deletes controversial Twitter messages. (This may continue and, as a result, some of the links to Twitter in this posting may no longer work; they were functional at the time this document was originally posted.) Another individual subsequently creates a new “clinical_tweets” with a real name of “Ex Staff” and a Canadian flag as the profile image. Is this someone from the fired staff of Clinical Reader? Without an acknowledgement from an official Clinical Reader channel, we can’t be sure.

Tweets from Clinical_Reader

On Saturday afternoon, a twitter account “clinical_reader” was created with these messages and other messages that give the appearance of being the official Clinical Reader Twitter account. With the subsequent registration of the “clinical_tweet” Twitter account and no postings on the Clinical Reader website, we can’t be sure which of these channels are official.

Updates

This blog post has been sitting in my account for a number of evenings as each day brings new events and new revelations. I don’t doubt that there will be changes to the story. If new events have a lesson in them, I’ll try to update this post.

Observations

Closing in on nearly 3,000 words, this posting is already long. Observations and recommendations, as a result, were moved to a subsequent DLTJ post so as not to get lost at the end of this one.

The text was modified to update a link from /johnhopkins to http://twitter.com/johnhopkins on December 30th, 2010.

The text was modified to update a link from /bengoldacre to http://twitter.com/bengoldacre on December 30th, 2010.

The text was modified to update a link from /bengoldacre to http://twitter.com/bengoldacre on December 30th, 2010.

The text was modified to update a link from /feedstitch to http://twitter.com/feedstitch on December 30th, 2010.

The text was modified to update a link from /coldpie to http://twitter.com/coldpie on December 30th, 2010.

The text was modified to update a link from /lukelibrarian to http://twitter.com/lukelibrarian on December 30th, 2010.

The text was modified to update a link from @lukelibrarian%20Your%20browser%20is%20lagging%20behind%20says%20’Make%20effective%20use%20of%20the%20library%20journal%20shelf’%20at%20http://bit.ly/uniXw to http://bit.ly/uniXw on January 13th, 2011.

(This post was updated on 11-Jun-2014.)