Medlibs Round: PubMed and Discoveries

PubMed class at StanfordEdTech

It is a pleasure and honor to host Medlibs Round here at Eagle Dawg Blog after a few delays I apologize for. The theme this month is PubMed, one of several databases in the the National Library of Medicine's (NLM) Entrez life sciences search engine developed by the National Center for Biotechnology Information (NCBI), and third party functionalities.

There is much more to the NLM than this single database though, and understanding the Entrez connectivity helps to explain some of the changes that are happening in PubMed today to lead to more research and better medicine for tomorrow. I hope you enjoy the ride!

What is this "Discovery" stuff NLM keeps mentioning?

NCBI's "Discovery Space" Facilitating SNP Research

The above chart refers to information in genetics but it was the 'discovery space' label that caught my attention. This illustrates NCBI's main purpose as not solely being the PubMed administrator but developing resources "for the better understanding of molecular processes affecting human health and disease" (source) which involves information in the other Entrez resources.

In December 2008, I did some research on the Discovery Initiative, "one of the most important projects NLM is working on." If you're not already familiar with the Discovery Initiative please take a look and here's the gist of it:

The bottom line for us is discovery. We want people to make discoveries, and if we’re using up real estate on the Web page for things people don’t click on, and if we can put things on there that would have been associated information, then we should do that.

The Discovery Initiative is behind some of what we are seeing with PubMed changes now after several years of work behind the scenes to encourage cross-referencing within Entrez. We are also starting to see communication from NLM including reference to the Discovery Initiative and suggested resources that appear in PubMed being called Ads. There still is not a clear description/explanation of the Discovery Initiative from a PubMed perspective in either the NLM Technical Bulletin or PubMed New/Noteworthy RSS feed, the resources we're encouraged to subscribe to for current PubMed news. The Discovery Initiative has received coverage in NCBI News in June, March, and February 2009.

Discovery is in the eye of the beholder

Marine Biological Laboratory (MBL) seal on the Lillie Laboratory

With that background about the Discovery Initiative covered, Shamsha Damani at MD Anderson created a very clear guide and commentary for her clinicians about the recent and upcoming changes in PubMed: What's New and What's Ahead, These include a great description and screenshots of the Discovery Initiative ads with the suggestion that "they are handy and are a great discovery tool when you are grasping at straws or have too many results to wade through." PubMed: My NCBI changes by Miner Library mentions the new ability to create custom filters as "a terrific new development."

Custom filters in PubMed. Fail. by Mark Rabnett of Gossypiboma begged to differ with this assessment of the custom filter changes due to encountering a failure with the saved searches tab.

In the second incident of NLM staff replying to blog complaints that I know of (the Harvard student being the first), an anonymous comment (supposedly a developer at NLM) was left on Mark's post saying the problem would be fixed. Shortly thereafter Mark posted PubMed custom filters: where the bee sucks which includes

I must apologize for the flippant remarks I made in my previous post. It must have been the nicotine. I hope that the snag with the NLM health literacy filter can be fixed. I intend to use Custom Filters frequently and uncomplainingly when all the kinks are worked out.

In a second comment, the supposed NLM developer replied

Enjoy your filters, and thank twitter’s re-tweets for getting your message across.

Why not have an official PubMed Twitter presence to directly offer feedback in the first place? Clearly someone there wants to engage. Often we hear complaints about PubMed and related services but with the email-only "official" feedback channel frequently is no acknowledgment from either NLM or users that a problem is taken care of, nor a way for others to see if an issue is Down for everyone or just me.

What's your Discovery?

Blue lobster at the Woods Hole Science Aquarium

PubMed Impact Factor from Alisha764's Blog explores the role of offering impact factors of journals and author information as a quality indicator in the wake of fake articles, journals and changed/fabricated data. This is not without controversy but indicates that many are seeking additional ways to check the quality of information. Merely saying something is indexed in MEDLINE is not enough to say it is authoritative health information since Time Magazine is there in the consumer health subset.

Five ways to improve PubMed from Gossypiboma includes his top results from the brainstorming suggestions during the recent Canadian Health Libraries Association conference about improving PubMed. A trick for #5 (Simplify the creation of permanent links to PubMed records) is to begin with http://pubmed.gov, take a look at an article for the PubMed identification number (PMID), then tack that on to the end for http://pubmed.gov/16719081

Shamsha Damani's guide includes

My one hope is to see NLM participate more in social media (especially Twitter!) so more voices are heard. The NLM reps at MLA’09 were extremely nice and open to suggestions; it would be nice if they could extend their presence in social networking venues as well.

I share this sentiment about social media as an official feedback channel for PubMed. This should not be via anonymous comments to blog entries but a sustained, active effort by both NLM and PubMed users utilizing some of the best practices learned from online participation formats such as the Open Government Dialogue.

There is hope for change on this front with the emergence of official Department of Health and Humans Services (HHS) Twitter accounts. @NLM_SIS is a legitimate account from Specialized Information Services (as is their SlideShare account... check out Semantic Processing of Twitter Traffic for Epidemic Surveillance) although considered experimental at this time. For the time being many, but not all, government agency presences on Twitter are essentially push-out news information sources similar to RSS feeds that don't usually interact with followers.

What have third party vendors Discovered?

Fiddler crabs in the Marine Resources Center at MBL

Got PubMed? Pubget Searches and Delivers Scientific PDFs from Bio-IT World news shares the news that this full-text PDF retrieval service has expanded quite a bit from when David Rothman covered it in April 2008, and has just announced it has partnered with 500 institutions including NIH itself (NLM is one of the 27 institutes of the National Institutes of Health).

Hope Leman posted GoPubMed: Interview with Michael Alvers at Next Generation Science where she asked some great questions of GoPubMed's CEO about why medical librarians and others should use this search engine instead of regular PubMed in addition to theoretical (the blended future role of search engine and social networking) to the practical (What's the difference between "find all" and "get all"?). This is a comprehensive interview well worth reading and reflecting upon over a coffee break.

Some non-PubMed but related tangents that were submitted along the discovery lines include Dr Shock MD PhD's posting about Online Medical Book Search and Clinical Cases and Images Blog's coverage of how to Create an automatic differential diagnosis list with Google Squared.


Panel at the MBL Associates Gift shop, home of cool socks

I recently attended the MBL/NLM BioMedical Informatics course and am still reflecting upon what Dr. Don Lindberg, Director of NLM, said regarding the emerging role of interactive publications. Wouldn't embedding an actual rhythm strip in a journal article involving electrocardiograms be a better use of online publications than a snapshot of one still image? I'll conclude with this image from his presentation:


You have seen what everyone else has seen. What do you think?