The role of PubMed Central (PMC) in last year’s launch of eLife was controversial within the scientific publishing community — by publishing articles on PMC before eLife’s commercial platform was ready, the National Library of Medicine (NLM) and the National Center for Biotechnology Information (NCBI) provided eLife with immediate validation through indexing in PubMed, gave a set of billion-dollar funders access to free technology and consultations, short-circuited stated policies and practices other publishers abide by, and revealed poor management of conflicts of interest.
When asked, government employees at the NLM and NCBI asserted that nothing improper had occurred or refused to talk. When I spoke with David Lipman in October and asked him how eLife came to be included in PMC and published first on PMC, he said it had gone through the normal application process, that the Acquisitions and Selection Section had approved the application, that eLife representatives “mentioned” their site wasn’t yet up, and that he knew nothing of any discussions leading to eLife being published through PMC.
However, based on emails and documents obtained via an ongoing Freedom of Information Act (FOIA) request, we now know that the initial request came from eLife, specifically and explicitly wishing to use PMC as “a publishing vehicle” to gain early traction with authors, legitimize their initiative, and advance their launch plans. We also know that Lipman, in fact, knew a great deal about the discussions about PMC publishing eLife, as he was central to them. And eLife’s application was far from normal, if any document ever truly existed — the FOIA request has not produced it in the 500+ pages sent thus far, and there is a good amount of evidence that the application was skipped. Fudging this step required many deviations in procedure for staff at NLM in order for eLife to be included immediately after a hasty decision based on private discussions between Lipman and eLife.
We also see that at the point eLife was given de facto acceptance into PMC and PubMed (March 2012), no articles had been produced, no facility for article submissions had been established, and no content could possibly have been made available for review. Acceptance occurred in seven (7) business days, despite a lack of qualifying materials (articles or files) or a formal application, which bears on the fairness of the PMC selection process, one of the primary sticking points for publishers.
Finally, even the PMC National Advisory Committee (NAC) apparently knew nothing of the special eLife arrangement, despite the fact that the agenda at their June 2012 meeting — which was attended by Lipman and others involved — included a presentation about eLife from the person who initially requested and received special treatment for eLife.
To help tell the story, I’ve recreated many of the most salient parts of the emails and documents below. In addition, I’ve scanned four of the major documents and linked to them at relevant points in this post.
The Pitch and the Catch
On February 24, 2012, Mark Patterson of eLife emailed Ed Sequeira of the NLM, with a cc to David Lipman and Jo McEntyre, making the request (link to scan of full email):
Jo McEntyre is a former scientist at NCBI who is on the board of EuroPMC, formerly UK PubMed Central.
The email address Patterson used when asking for PMC’s help was a Wellcome Trust email account, which highlights the conflicts of interest in having Wellcome Trust employees on PMC’s National Advisory Committee (NAC). In 2008, Robert Kiley of Wellcome stepped down from the NAC, and was asked to recommend someone to take his place. He selected Chris Bird, Wellcome’s solicitor general. This effectively gave Wellcome two consecutive terms on the NAC. The ties between PMC and Wellcome are substantial and may have influenced the fate of eLife at PMC. The appearance of impropriety is certainly present.
The email is telling. Patterson knows what he’s asking for — publishing services, technical assistance, and a leg up with authors that will benefit eLife and its plans. He knows he is requesting to use PMC as “a publishing vehicle,” despite PMC stating publicly that it is not a publisher.
Lipman replies quickly and in a manner that seems correct initially, but soon drifts into double-talk:
This is a confusing set of statements, and conflicts with stated PMC policies, one of which states:
The journal must have a reasonable number of published articles in order for NLM to make a decision about its scientific quality.
With Lipman’s ambiguous response, we’re down the PMC rabbit hole, where non-publication is publication, and where journals are accepted before materials are published despite publication being a prerequisite to evaluation. However, the last sentence indicates Lipman has an approach and wants to discuss it with Patterson.
Sequiera reminds Lipman later on the same day that there are some technical reasons not to permit eLife’s request in its current form:
JBT is the Journal of Biomolecular Techniques. The PMC version is currently framed within the association’s site. It is one of the two journals (the other being the Journal of the Medical Library Association or JMLA) that PMC publishes, despite assertions it is not a publisher.
At this point, there was a clear opportunity to tell eLife that PMC couldn’t support their request. Instead, Lipman takes this information and determines the only way to help Patterson is to have PMC become a primary publisher of finalized eLife content, as this solves the technical challenge and helps Patterson and those he represents (Wellcome and eLife):
This indicates that the only objection Lipman was still contemplating was a technical hurdle — there is no indication that he was worried about perceptions of favoritism, impropriety, or cronyism. As we now know, this decision snowballs into PMC becoming the initial publisher of eLife, and documents show that US government employees had a high degree of awareness of what they were doing.
Complications Arise, Shortcuts Are Taken
On March 4, 2012, Patterson emails Sequeira, with a cc to Lipman and Melissa Dodd, head of production operations at eLife (scan of these emails):
Note that at this point, eLife had not even applied for an ISSN, which is a prerequisite for filing an application. Yet, eLife is moving ahead as if it has been accepted. Sequeira replies:
Sequeira slips an expression of his consternation into this email, asking exactly why this strange course of action is being pursued. Also, it’s clear that he knows they are launching a journal — “the journal will exist as of the day of the first article is released in PMC.”
To erase any doubt about how things will proceed, Lipman chimes in soon afterwards:
A long email from Patterson follows, with the salient point here:
Note Point 2 — at this stage, eLife couldn’t even accept manuscripts.
Sequeira knows they’re muddying the waters, as reflected in part of a response to the message above:
How this arrangement solves any conceptual problem eludes me. As Sequeira notes, the articles will have been published on PMC. It bears emphasis that PMC proclaims that it is not a publisher, but obviously eLife was seeking a publisher, and PMC obliged. In my interview with Lipman in October, he denied direct knowledge of the process eLife went through for PMC and PubMed inclusion, yet he was directly involved. In an interview last November, Patterson disputed my claim that PMC was acting as eLife’s primary publisher at the time, yet here he is requesting exactly that.
How did eLife get accepted into the NLM catalog? In October, Lipman assured me that it had followed the usual procedure. I also asked Joyce Backus, Acting Associate Director for Library Operations at the NLM, three specific questions, which she failed to answer:
- When did eLife submit their application?
- When did the Acquisition and Selection Section consider their application?
- When did the idea of allowing them to post immediately on PMC arise? Who broached the idea? Who was involved in those discussions?
Judging from the emails I’ve seen thus far, eLife did not submit an application, the Acquisition and Selection Section did not consider any application, and we’ve already seen the answer to the third question. Why my clear questions were not directly answered when I was able to get the information presented here is troubling. In November, when asked directly by a journalist, Lipman refused to comment, saying, “At this time, we don’t think that engaging Mr. Anderson in a public rebuttal is productive.” Perhaps this is because there was little to rebut. Nevertheless, US government employees should answer questions posed by US citizens.
For eLife, the process from request to acceptance seemed to take a total of seven (7) business days — without any content in existence.
Copied on a discussion about eLife, Jennifer Marill of the NLM emailed Diane Boehr on March 5, 2012:
“S&A” seems to refer to the Selection and Acquisition Section.
On the same day Boehr at the NLM sent Judith Eannarino an email:
Eannarino replies to Boehr:
There is then an exchange between Boehr and Wilma Bass early the morning of March 6, 2012, where some concern is expressed (TA is “title abbreviation”):
Here, staff seem to be coping with the lack of an official application from eLife. Boehr replies the same morning:
Bass’ next response shows how the team is proceeding despite there being no selection form or documentation.
I believe “pub” it a typo of “put.” Marill re-enters the email chain later that day:
This “prelim record” number is still eLife’s catalog number at NLM.
On March 13, after Sequeira returns from his vacation, Marill finishes the exchange with this:
One wonders how, after all these special accommodations and shortcuts, words like “normal” and “regular” can be used to describe the process. Even before eLife had any articles or had filed an application for approval (if it ever did), it knew it had secured indexing in PubMed, had free access to a US government-run publishing platform, and that it would have months of advantage in the market.
The extent of the favoritism PMC provided eLife becomes clearer when we look at an email exchange between PMC and a non-favored publisher months later, after eLife launched on PMC.
On October 15, 2012, eLife content went live on PMC. On October 18, 2012, a staff member of Frontiers (an OA publisher) wrote to Chris Kelly at PMC:
Kelly’s response comes quickly:
The way this text smudges established lines, even those Kelly used previously, is worth noting. Remember, at the point eLife was accepted into PMC and PubMed (March 2012), no articles had been produced, no content was made available for review, there was no submission system, and the NLM Selection and Acquisition Section had not evaluated the journal except to take Lipman’s word for it and seemingly create a paper trail after the fact. Apparently, published standards had been abandoned by this time.
Many Frontiers journals are still not included in PubMed Central, despite Frontiers being founded in 2007.
Assisting in the Promotion of eLife
On September 8, 2012, Jennifer McLennan, a communications consultant with eLife, emailed Neil Thakur, who has been the point person on the NIH Public Access Policy, about the possibility of collaborating on publicity:
Thakur emails Sequeira the same day to see what he thinks:
Sequeira replies on September 10, 2012:
Apparently, Sequeira’s opinion was not shared, as matters continued to evolve on the publicity front, and continued to draw in PMC and the NIH. On September 22, 2012, Patterson emailed Sequeira wondering if he could take a look at the editorial eLife was writing announcing “the first articles that we publish on PMC (date still TBD).” (scan of these emails) He sent the editorial as an attachment. (scan of redacted editorial) On September 25, 2012, Sequeria writes back:
The two exchange more emails and edits before Sequeira states, “I think this works.”
The sections of the editorial they redacted before publication are:
Releasing the first content from eLife on PubMed Central is an unusual step . . . . We would like to take the opportunity to thank our colleagues at PubMed Central who have been enormously helpful in publishing these first articles. We will be publishing further content on PubMed Central in the coming weeks. . . .
This is a very troubling set of exchanges. First, it reveals that PMC knew full well they were giving eLife significant assistance. Second, it shows that PMC knowingly sought to conceal the preferential treatment they were giving eLife. Third, and this is consistent in various exchanges, PMC knew it was acting as the publisher of eLife content. And, fourth, it shows that PMC was hoping this would slip by without too much controversy or without requiring them to do this for everyone.
There are more emails about how PMC worked hard with eLife to get their XML right prior to publication and so forth. And, this is an open FOIA request, so more documents should be forthcoming. There are still gaps in the record. I’ll share further findings and some perspective in the days and weeks ahead.
Last October, based on strong circumstantial and some direct evidence, I wrote that PMC had aided eLife by providing technical, consultative, technology, and indexing support in a manner that was highly unusual, and specifically designed to help eLife gain an advantage in the marketplace. I also speculated that mismanagement of conflicts-of-interest with Wellcome appeared to have had a bearing on eLife’s preferential treatment. Given these telling email and document exchanges, we now have direct evidence that PMC knowingly gave eLife significant and valuable launch assistance at eLife’s request, secretly agreeing to use US taxpayer-funded resources and salaries to abet the launch of an online journal funded by three multi-billion-dollar philanthropies (two of which are headquartered outside the US), while colluding with these entities to minimize their chances of being accused of impropriety and avoiding direct questions about how this occurred once the community was aware of the results of the deal. In addition, PMC knowingly was stiff-arming other publishers requesting indexing and inclusion.
In a way, this is a painful post to write. Many of the people in these emails I know and respect. However, the reputation of PubMed, the NCBI, the NLM, and the NIH as objective, fair, and unbiased registrars of scientific information for public use may have been severely compromised. These records and other facts strongly suggest there is cronyism, preferential treatment, an unclear set of standards, conflicts of interest, and abuse of authority. There are certainly many questions to be answered.
(Note: Tomorrow’s post will cover some vexing discussions at PMC regarding PeerJ, and how PMC coordinated with eLife on how to respond when PeerJ questioned PMC’s handling of eLife.)