Brands set and fulfill customer expectations. Dunkin’ Donuts coffee or Starbucks? There’s not much practical difference, but there is a difference in expectations.

In a well-defined brand family, you know which end is up — with Toyota, we all know the Avalon is above the Camry is above the Corolla is above the Yaris. Even less obvious brand relationships, when well-managed, convey this implicit hierarchy. Marriott does an excellent job of conveying rank within a set of brands that aren’t obviously related, with Ritz-Carlton and Bulgari at the top, down through Marriott to Courtyard to Fairfield Inn.

mixed up buttons

It’s no secret there’s widespread confusion about the brands the National Library of Medicine (NLM) is using to describe its bibliographinc databases and indices of citations to the biomedical sciences — PubMed, MEDLINE, and PubMed Central.

Make no mistake — these are brands, complete with trademark registrations and protections, each intended to designate a specific function or service. And there is a hierarchy the NLM seems to want to establish, with MEDLINE atop the brand pile, at least conceptually, if not practically. In statements, entries are “elevated” to MEDLINE from PubMed, for instance. It is their best-managed brand, even if it’s not their best-promoted brand.

Unfortunately, these strong brands aren’t being managed very well, leading to confusion in the marketplace. Editors, librarians, publishers, and authors fall into the layers of confusion from time to time.

The brand confusion is even baked into the Web. For instance, if you use Google to search the phrase “indexed in PubMed,” the first link you get is to a page entitled, “List of Journals Indexed for MEDLINE,” which talks about how the publication “List of Journal Indexed for MEDLINE” ceased publication after 2008. The second link isn’t much better — a link to an FAQ about journal selection for MEDLINE inclusion. In here, at question 17, you finally get a link explaining the differences between PubMed, MEDLINE, and PubMed Central. However, on some questions, the entity is still listed as “MEDLINE/PubMed,” underscoring the notion that the two are equivalent.

The confusion is also baked into the NLM’s site architectures, with the menu for “PubMed/MEDLINE” on this page resolving to a page that is labeled “PubMed” and has a PubMed URL (http://www.ncbi.nlm.nih.gov/pubmed/).

The confusion goes deeper in casual usage, with editors, authors, and NLM employees themselves using “PubMed” as a shorthand for “PubMed Central.”

The conflation occurred in two stages. Initially, PubMed was created in 1996, as a free online version of MEDLINE. This linked PubMed and MEDLINE, but because PubMed was linked to the Internet, it became the de facto public name for MEDLINE. Then, when journals included in PubMed Central were automatically listed in PubMed, this ended the era of equivalency between MEDLINE and PubMed. The PubMed brand was being extended beyond MEDLINE, but remained conflated with it.

Since then, the PubMed brand has been extended in two ways — as PubMed Central (which seems like a sub-brand of PubMed), and as PubMed Central’s registration service. This expansion of duties for the PubMed brand has led to confusion in the marketplace, and even confusion about whether records in PubMed are “indexed” or “included.”

As you likely know, I filed a Freedom of Information Act request late last year, as part of an investigation into how PubMed Central colluded with eLife to give eLife unprecedented launch advantages, up to and including providing them with immediate PubMed listing and robust online publication support. As part of this request, other emails tangentially related also surfaced, including one discussing whether we should call being listed in PubMed “indexed.”

On November 13, 2012, David Gillikin, Chief of Bibliographic Services for the NLM, writes in an email to other NLM staff, referring to a post I had published the day prior:

Another one from Kent this weekend. This one is focusing on eLife and PMC, and the “economic windfall” that he’s projecting about eLife going on PMC for free, and it’s [sic] automatic inclusion in PubMed (which he calls PubMed indexing).

David

This email reminded me of the conversations going on at the periphery of these discussions of PubMed Central — whether or not there’s confusion about the NLM’s brands. And it introduced a new item — whether some brands represent indexing or simply deposit in a database. From a bibliographer’s standpoint, this is a technical but important distinction. From an editor’s or author’s point of view, it’s also important to understand — the assumption is that being indexed in PubMed signifies a major endorsement of a journal, after serious review by experts at the NLM. That’s the brand promise as understood on the market.

Unfortunately, it’s the brand promise of MEDLINE, not PubMed. And this is the root of much confusion.

There are other clues that this distinction is recognized internally at the NLM. In a different set of FAQs, the NLM refers to PubMed in this manner:

PubMed is a database of citations and abstracts for millions of articles from thousands of journals.

It’s not an index in this setting, but a database.

I asked Gillikin via email yesterday about this, and he provided a very helpful response:

From a basic database point of view (and general user point of view), “indexed” denotes inclusion in a database.  For those of us managing MEDLINE, the term “indexed” has a specific meaning.  We use it to refer to citations that have Medical Subject headings (MeSH) and additional data added to the citation in our indexing process.  The inclusion of MeSH provides greater searchability of the article citation, and provides more access points and unified terminology for searchers.

So we use the term “indexed” to refer to citations in PubMed that are MEDLINE citations, while “included in PubMed” refers to all citations in PubMed, whether it is a MEDLINE citations, a citation in the workflow to become a MEDLINE citation, or a citation that will not become a MEDLINE citation.

The vast majority of data in PubMed are MEDLINE citations.  It is 90% of the PubMed database.  Another 5% are citations that are in the workflow to be indexed with MeSH (publish ahead of print citations, citations already in our indexing stream, or citations from our OldMedline project that will have their keywords mapped to current MeSH).

The other 5% are citations that will not be indexed with MeSH, and come from a variety of sources:  citations from MEDLINE journals for articles prior to the journals inclusion in MEDLINE, MEDLINE journal citations that are deemed out of scope due to subject scope, citations for articles from non-MEDLINE PMC journals, and citations in PMC due to compliance with the NIH Public Access Policy that are not published in a MEDLINE journal.

Hope that helps clarify your question.

Regards,

David

Gillikin is observing here that it’s OK for general users to call inclusion in the PubMed database “indexing,” but this contrasts with the prior email correcting my usage. I think he’s saying that the public may call it “indexing,” but people who should know better should, well, know better. He may be experiencing the same conflation of brands, and brand promises, we’ve been talking about. They are difficult to untangle at this point.

If your audience is confused about your brands, it’s not their fault. It’s like Buick blaming me because I don’t understand their brand hierarchy involving the Regal, the LaCrosse, and the Verano. (Many auto brands solve this problem by taking the direct route in their branding hierarchy, using progressive numerical or alphabetical schemes — BMW 300, 500, 700; Audi 4, 6, 7; Mercedes C, E, G, S; Jaguar XF, XJ, XK; Infiniti EX, FX, G.)

There is a practical sticking point within the brand conflation at the NLM, one that publishers behind the 90% of the papers in PubMed that are indexed in MEDLINE probably appreciate — it takes a lot of time, effort, and patience to be indexed in MEDLINE. Being indexed in MEDLINE is a non-trivial accomplishment, much harder than being included in PubMed via inclusion in PubMed Central. To have this work and effort equated with PubMed Central records that come in via different criteria doesn’t seem right. The processes are not equivalent, and the resulting records are not, either. By allowing the PubMed brand to signal equivalency for journals not indexed in MEDLINE, the NLM is drawing value from MEDLINE journals and sharing it without permission with non-MEDLINE journals. Because PubMed and MEDLINE are conflated brands now, their value is intermingled.

The pathway into PubMed for OA content via PubMed Central is also becoming less disciplined, it appears. In addition to how eLife was expedited behind the scenes and apparently without following any process, PubMed now includes non-journals that don’t even know when their papers are ready for PubMed inclusion because their peer-review process is so poorly managed.

But the phrase “indexed in PubMed” has nearly talismanic powers, which OA publishers whose content is only included in PubMed wield proudly. I don’t blame them. The brand confusion is widespread, it’s not their fault, and it feels natural in our community to say “indexed in PubMed” no matter how you were included.

Once you understand the branding problems and conflated promises extant in the PubMed brand, it becomes easier to see why there’s a problem. Examples are plentiful, as in this statement from F1000 Research, a non-journal now “included” in PubMed, shows:

Once an article has received two ‘Approved’ statuses from the referees, or one ‘Approved’ status and two ‘Approved with Reservations’ statuses, the article will be indexed (in PubMed, PubMed Central, Scopus, Embase and others)

If the difference between the MEDLINE brand and the PubMed Central brand were clearer, the better wording would be “included in PubMed, PubMed Central, Scopus, Embase, and others.”

The confusion continues, with SPARC’s page about “Getting Your Journal Indexed” perpetuating the general misunderstanding that PubMed is an index:

Though each discipline may have indexing and abstracting services specific to the field, the most well-known, well-regarded services include:

  • AGRICOLA (AGRICultural OnLine Access)
  • Biosis Previews®
  • Chemical Abstracts
  • ISI® Web of Science
  • PubMed

Do they mean MEDLINE? Again, the NLM brands are not well-sorted, leading to substitutions that add value to the PubMed brand while decreasing the value of the MEDLINE brand.

There are different forms of confusion about PubMed Central, PubMed, and indexing, as this page from Frontiers shows:

In the next coming months all the articles published in the journal will be indexed with the Full Text in PMC and citation in PubMed.

The confusion has even lead to correction notices, as in this example from Open Medicine, a Canadian OA journal:

This corrected editorial replaces the version published on 5 January 2010, which stated that Open Medicine is indexed in MEDLINE, when in fact it is indexed in PubMed (of which MEDLINE is a subset). Open Medicine currently has an application under review for indexing in MEDLINE.

The Open Medicine team is pleased to announce our recent acceptance for indexing in PubMed—an official stamp of approval from the US National Library of Medicine (NLM) for the scientific and technical quality of articles published in our journal.

This correction still perpetuates two errors — PubMed is not “an official stamp of approval” from the NLM in this case, for anything other than inclusion in PubMed Central, which has different and “flexible” standards; and MEDLINE is not a “subset” of PubMed, as in an embedded brand, but a higher level of approval, a superior brand. (Open Medicine has subsequently been approved for the MEDLINE index, which is what their initial brand sense had correctly identified as the one to really brag about.)

Interestingly, eLife has been a bit more savvy about this, at least if this interview from late 2012 with Mendeley is any indication:

Q: Will eLife be indexed by Pubmed? Google Scholar?  How will you ensure an author that your articles will be easily discoverable by readers?

A: The eLife journal will be hosted by Highwire, who have a very good working relationship with Google Scholar. As you know by now, our content is also listed on PMC, Europe PMC and PubMed.

Notice they don’t quite correct the interviewer, but also don’t compound his or her error about indexing.

These brands — PubMed and MEDLINE, in particular — are important, as are the perceptions they engender. For instance, I’ve encountered authors who want to know specifically if a new journal is “indexed in PubMed.” If it’s not yet, they might withdraw their paper. Maybe they only want the journal to be “included” in PubMed, but I’m pretty sure they want certification beyond mere presence in a search engine. Their real question is about being indexed in MEDLINE. But they don’t understand the differences, and therefore transfer the brand promise of MEDLINE to PubMed. In so doing, and considering the backdoor into PubMed, they are inadvertently giving PubMed Central extra authority based on transference of authority from MEDLINE through PubMed conflation.

Until these confusing brands and approval pathways are rationalized, we’d all do ourselves — and the NLM — a favor by being clear on the distinctions between being included in PubMed and indexed in MEDLINE. The NLM brand has power. The MEDLINE brand has power. The PubMed brand has power, some appropriated from the MEDLINE brand, some put at risk by PubMed Central.

Ultimately, the MEDLINE brand, and its sound processes, is the bedrock of the other brands’ power. But the NLM is hiding its light under a bushel, as the saying goes, for the sake of promoting weaker brands that draw down the equity of the MEDLINE brand. I believe they should lead with their most well-defined brand — MEDLINE — and sort the other two out from there.

Recent behavior aside, the NLM owes it to itself and its users to sort out its brands before they become even more confused, compromised, and bewildering.

(P.S. More confusion? Why does the NLM have both “Medline Plus” and “PubMed Health” as their consumer brands, despite the sites having virtually identical content?)

Enhanced by Zemanta

Kent Anderson

Kent Anderson

Kent Anderson is the CEO of RedLink and RedLink Network, a past-President of SSP, and the founder of the Scholarly Kitchen. He has worked as Publisher at AAAS/Science, CEO/Publisher of JBJS, Inc., a publishing executive at the Massachusetts Medical Society, Publishing Director of the New England Journal of Medicine, and Director of Medical Journals at the American Academy of Pediatrics. Opinions on social media or blogs are his own.

Discussion

24 Thoughts on "Extension and Conflation — How the NLM’s Confusing Brands Have Us All Mixed Up"

“For instance, I’ve encountered authors who want to know specifically if a new journal is “indexed in PubMed.” If it’s not yet, they might withdraw their paper. Maybe they only want the journal to be “included” in PubMed, but I’m pretty sure they want certification beyond mere presence in a search engine.”

My guess is that authors are mainly interested in two things:

1. can our paper be found easily by web searches?
2. is the journal on course to have an impact factor soon?

I’m interested that you feel that the MEDLINE brand is the one with the real kudos, and wonder if that’s because of your background in publishing? I will bet that most researchers, who know less about the behind-the-scenes details (as it were), couldn’t tell you what MEDLINE is.

No, that’s not what they mean. They aren’t asking concurrently, “Is your journal crawled by Google and Bing?” or “Will it be counted by ISI?” They believe that being in “PubMed” is a sign of authoritative endorsement, which is no longer entirely true. In fact, that perception is based on the belief that MEDLINE and PubMed are synonymous, which they are not. PubMed leverages MEDLINE’s reputation, which is why I believe MEDLINE is the top dog — it has the best process, and gives value to the other brands, not vice-versa.

The market has a conflated understanding of these brands because the NLM hasn’t managed them well.

Well, I do understand what you are saying about “inclusion in Pubmed” being taken as an imprimatur. I was just making two subsidiary points: firstly, I do think that concern about visibility is an issue to authors … it is to me, at any rate. Secondly, was just commenting, that you are saying that Pubmed is ‘using’ the MEDLINE brand. Perhaps this is happening implicitly – but there isn’t much explicit MEDLINE brand awareness for most biomedical researchers, I believe. (Which is kind of your whole point too, no? That there is confusion between them?)

Visibility is definitely a concern authors and publishers share. But I’ve loved over the past 5 years informing authors and editors that Google is, far and away, a much more important factor in discoverability than PubMed. Google crushes PubMed as a discovery tool for professionals.

I agree, that was my point — the conflation of PubMed and MEDLINE is allowing PubMed to appropriate value from MEDLINE and its processes, while somewhat concealing the back door some journals get via PubMed Central. The journals coming in the back door get to eat and drink the same food as the MEDLINE journals, even though they are more akin to wedding crashers.

I’ll second Kent’s comments. I frequently see authors (and editors) wanting to know if the journal is “indexed in PubMed”, and they’re not asking because of discoverability, but instead are asking as a sign of quality assurance. For many, the idea is that a journal must both have an Impact Factor and “be in PubMed” (by which they really mean having passed muster by MedLine) in order to count toward career advancement.

I find it odd that the federal government should be judging journal quality. Who authorized this? Certainly not Congress.

I think it is a leftover from the days when the NLM functioned more as a library, and when journals were selected by the NLM, that was a sign of quality. Then, the NLM started publishing a book of these listings when the records were computerized as MEDLINE, and that became a useful standard resource for journal abbreviations and validation (for editors, copy editors, and authors). It morphed again with PubMed being a free port of MEDLINE to the Web in the mid-1990s. But the role of MEDLINE as an arbiter of quality based on NLM selection criteria continues. It’s like having the most important library in the world validate you! Or, it used to be . . .

Speaking of Brand confusion – inclusion in Medline, or PubMed or PubMed Central is entirely independent of when/whether a journal will receive a Journal Impact Factor. JIF’s are not produced by NLM.

I’m not sure many are confused about this, but instead they’re seen as separate (but important) indicators of quality.

If you work with enough people in scientific research and publishing, even at the highest levels of academia, the confusion is consistent and real. If you need an FAQ (which is hard to find, see post) to explain the differences, you’re not doing a good job of branding. Does Audi need an FAQ for you to know that the A4 is smaller than the A6? Does Marriott need an FAQ for you to know you’ll pay more and get more amenities staying at a Ritz-Carlton vs. a Fairfield Inn?

In the last couple of months, I’ve heard authors, scientists, NLM employees, and librarians mix these three up in common parlance.

Frankly, I’m not terribly confused. But the market is. And that’s a problem for the NLM, and also unfair to journals that have gone through the more rigorous MEDLINE process, only to find themselves labeled as “indexed in PubMed” when this can mean something very different, i.e., inclusion in PMC, without rigorous review.

Is there a visual representation of how the three are related? I think that would be very helpful.

On the other hand I and my colleagues at DOE OSTI use the term indexed to mean full text analysis, which means only PMC does indexing. That is how the term was explained to me when I entered this industry ten years ago. Abstracts and metadata are not indexing. So part of the problem is that indexed is a vague concept, which NLM cannot change.

MeSH is the brand I’d promote over Medline for two reasons: (1) Letting Medline as a brand fade into the background could remove most of the confusion people experience between Pubmed and Medline. (2) Promoting MeSH, by qualifying and referring to Pubmed records as either “Pubmed-MeSH” or “Pubmed-none” would put the focus more usefully on the quality of, length of, and number of important access points for MeSH-indexed records contained in the Pubmed database. Should terms from other indices be applied in the future to records in the Pubmed database, the brand names of those indices could likewise follow the hyphen. Links at Pubmed for authors/editors seeking information about inclusion criteria for journals (or perhaps one day, single articles) to be indexed using MeSH would open a full history and explanation of how Medline and MeSH and Pubmed interact.

Perhaps you haven’t talked to your local Medical Librarian or the Director of the NLM, Dr Donald Lindberg who has led the Library since 1984. . NLM is limited in how it markets and brainds its products by law. Being indexed by MEDLINE is much more important than being “indexed” in PubMed as a quality indicator. I am amused by journals who claim quality by being “indexed in Proquest, Ebsco, Ulrich’s, and Directory of Open Access Journals”. PubMed contains publisher provided citations not critically reviewed for MEDLINE, whats so confusing about that? Besides most use MEDLINE through a secondary source, not the Entrez system which is Pubmed/Medline.

Please point me to the laws limiting how NLM markets its brands and products. I’d be interested in seeing how this has any bearing on how it names its offerings, changes inclusion criteria, or affects brand management.

Dr. Anderson, I’m not sure what those laws are, but whenever I have taken NLM training over the years, this is what Ive been told: NLM cannot market its products in competition with the private market “by law”, so perhaps its by administrative rule. Perhaps there is sinply no funding. I understand this is particularly true for Medline PLUS, the consumer health library, which publishes high quality consumer health resources, with no advertising! I do agree the proliferation of NLM databases has made it confusing for the researcher at times. My first NLM traning to search MEDLARS databases of Medline, Toxnet, and other files back in 1990 was 5 full days long. Even now its difficult for librarians to keep up with knowledge of the resources available, and what databases has what, and the expert skills that are important for mediated searching. I appreiciate your blog post.

Paul’s points above are well taken. The idea of non-competition is more of a vague rule than a specific statute, but it is widespread. Not advertising is a major case.

Then too the fact is that these so-called brands are really complex interlocking systems so it is hard to understand them. DOE OSTI has been my client for several years and they have the same problem (http://www.osti.gov). They have a bunch of different information resource products that are hard to distinguish. Some are databases of documents, data or software but others are portals that link to these as well as to other people’s databases. There are even databases of portals. Plus there are many levels of nesting and partial overlap.

The resulting confusion due to complexity is a central feature of the information age, not a case of brand mismanagement. This may well be true at NLM as well, for when I have looked at their stuff I found it incomprehensible. After all “instant information” is an oxymoron. These are hairy systems, not easily described or distinguished.

I was with you until you described the confusion occurring with NLM’s brands as “a central feature of the information age, not a case of brand mismanagement.”

Apple has no problem with all their brands, from iTunes to iPod to iPhone to iOS to Mac Air to iCloud. Google has no problem with its brands — Google, Google Maps, Google Books, Android, Google Earth, AdWords. Microsoft has no confusion about its information brands, from Word to Excel to PowerPoint to XBox 360 to Windows to Skype to Visio to .Net to Fast. These companies understand how brand management is key to success in a complex world, and do it well.

Brand management harmonizes/rationalizes complexity. Done well, it imposes clarity. Done poorly, complexity carries the day. Complexity is a central feature of the information age, it’s true, but the resulting confusion of products and brands is a classic case of brand mismanagement. It’s a particular problem for governmental agencies, most likely, for the reasons you indicate — things aren’t managed to the market, but respond to policy perturbations or individual initiatives, so they grow weedy fast.

I said that the NLM confusion may be due to complexity not that it was. As usual I am offering an alternative hypothesis not making a claim. But I also offered evidence for this hypothesis namely my own experience as an expert on confusion looking at their stuff. However the line you quoted referred to OSTI’s product line, not NLM’s. I doubt you can explain how OSTI’s products relate and differ.

More broadly there is in fact an enormous amount of confusion regarding the product lines you list. It is not a matter of Apple or Microsoft being confused but of the consumers being confused and many are. I personally do not know what some of the things you list even are. Others I have heard of but have no idea what they are or do. Confusion due to complexity is a central feature of the information age. Management cannot make this confusion go away because the complexity is intrinsic, so the confusion is not management’s fault. Stuff takes longer to understand than people have time for.

Hi Kent, you say:
“…we’d all do ourselves — and the NLM — a favor by being clear on the distinctions between being included in PubMed and indexed in MEDLINE.”

Meanwhile I saw this statement on a recent Call for Papers for a new journal:

“The application for PubMed indexing will be made as soon as the journal reaches the appropriate
milestones. Once approved, all previously published articles will be indexed.”

The new journal? JBJS Reviews! http://sites.jbjs.org/reviews/call_for_papers.pdf

Comments are closed.