[Editor's note: These are remarks made by Sheldon Kotzin, Chief, Bibliographic Services Division, National Library of Medicine, at the Annual Meeting of the Medical Library Association in Washington, DC, on May 24, 2004.]
ood Morning. I'm Sheldon Kotzin and it is my pleasure once again to welcome you to the NLM® Online Users' Meeting. After I speak for a few minutes about MEDLINE®, PubMed®, PubMed Central®, the Gateway, indexing and a few other topics, I will turn the microphone over to Joyce Backus, who will talk about MedlinePlus®, and Maria Collins, who will speak about DOCLINE®. After that, we'll take your questions.
I will start off as I always do with indexing, the backbone of our online activities. MEDLINE indexers continue to work at an astonishing rate. During the first six months of this Fiscal Year, 280,000 new citations were added to MEDLINE. If that rate continues, more than 560,000 indexed citations will be added. This would be 35,000 more citations than were indexed last year.
A formal study of the consistency of indexing for MEDLINE was completed this year. It marked the first examination of its kind since Mark Funk and Carolyn Reid published their landmark report in 1983. The new study was undertaken, not only to update Funk and Reid, but to serve as a baseline as NLM develops more powerful semi-automated indexing tools. The results from the indexing consistency study also may impact indexing policy, indexer training, and MeSH structure and content.
We have reported on the study at this meeting and will publish the methodology and results in more detail.
Since the last MLA meeting, 162 journals have been added to MEDLINE, bringing the total to more than 4800 titles. We continue to receive excellent support from publishers, who now send about 72% of all new records to us electronically, speeding their availability in PubMed. This cooperation, coupled with improvements in our in-house scanning operation, will enable the Library to cease keyboarding as a method for data entry in June 2004. Just eight years ago, keyboarding was performed on 100% of MEDLINE records! Also, to speed up access we are indexing 134 journals from their Web sites. Among these, 73 journals exist only in an online format.
Last year I mentioned gene indexing for the first time. It refers to links that NLM indexers make from MEDLINE citations, whose articles discuss gene function or structure, to LocusLink, an NLM database of protein information. Gene indexing is now performed on 14 organisms; seven new ones were added since last May. Soon LocusLink will be replaced by Entrez Gene, a database under development by NCBI. The result will be that gene indexing will be performed on a great many more organisms when a gene or protein is reported in an article cited in MEDLINE.
By now you know that Index Medicus will cease as a print publication at the end of 2004. I want to re-assure you that journals included in MEDLINE will still have to be recommended by our journal selection advisory panel. The Black & White printed MeSH® and the List of Journals Indexed (LJI) will both continue, although the LJI may be expanded to cover all journals currently indexed in MEDLINE.
Indexers will continue to indicate the main points of articles with a starred MeSH term. And plans are well-along to provide additional offsite security for records in our data creation system and for PubMed.
Now, I want to shift to our online databases. The Gateway now provides access to these additional collections: MedlinePlus Health Tutorials, MedlinePlus Current Health News, the HSDB database, TOXLINE® Special, and OMIM - the Online Mendelian Inheritance in Man. The Gateway also includes the LinkOut feature for all PubMed citations; direct links to free full text articles; and it searches both British and American spellings of words. During the past year, 13,400 abstracts were added to the Gateway from eleven AIDS meetings and 1250 abstracts were added from two health services research meetings. In the near future the Gateway will add Profiles in ScienceTM, NLM's digital archive of prominent scientists, and Genetics Home Reference, one of our growing number of new consumer-oriented databases. [Editor's Note: Access to Genetics Home Reference was added to the Gateway on June 21, 2004.]
OLDMEDLINE and the Index Catalogue
The creation of OLDMEDLINE records is a project of which we are quite proud. The contents of the 1951-2 Current List of Medical Literature were added to PubMed in February 2004. This brings the total OLDMEDLINE citations to 1.7 million. The Library will continue to convert citations from older printed indexes as time and resources permit. Data from 1950 will be added before the end of 2004. Also, the long anticipated project to make available the citations in the Index Catalogue is complete. You can hear more about access to these 3.5 million references published between 1880 and 1961 from Betsy Humphreys when she joins Kent Smith and Marj Cahn for the NLM Update. [Editor's Note: IndexCat, the online Index-Catalogue of the Library of the Surgeon-General's Office, (Index-Catalogue), is available at: http://indexcat.nlm.nih.gov.
Another item to be covered by this dynamic trio will be NLM's joint project with the American College of Physicians Foundation called Information RX. It is designed to help ACP members answer patient questions with information provided by NLM.
In June we will celebrate the seventh anniversary of PubMed. With your comments and feedback it keeps getting better and as a result more widely used. In March, PubMed was searched by nearly 5 million unique IP addresses. There were 59,200,000 searches and more than 225,000,000 page views. Some of you may remember that in 1996, the last year before the free Internet versions of MEDLINE, we totaled about 7 million searches annually. Now we exceed that number in four days.
PubMed enhancements are always explained in detail in the NLM Technical Bulletin and we are reviewing many of them at our theater presentations at the Exhibit Booth. Here are some of those enhancements. The clinical queries page has been revised and the filter strategies have been updated. In the Journals database, you can now build a search of multiple journal titles to be sent to PubMed. You can also send a journal title found in the Journals database to the Single Citation Matcher and send a term found in the MeSH database to the Clinical Queries page. Icons were added to the Summary results display to indicate if free full text is available either at the publisher's site or on PubMed Central. And most recently, PubMed's History feature was given some new capabilities, including the option to delete selected searches.
PubMed's Cubby keeps evolving. Soon work will be completed to allow users to request automatic e-mails of stored search updates. Users will also be able to turn on filters in PubMed that will show how many of the citations retrieved are for free, full text articles, or how many are for English-language articles, or for review articles. Many more filter options will be available and you'll be able to select up to five filters. Users will be able to choose to have this feature always active or only active when they turn it on in the Cubby.
PubMed Usability and Improvements
In late 2003, we conducted a usability study of PubMed, looking primarily at the home page design, the use of the search results page and its features, and the design and function of the Limits feature. Recommendations from the study are being reviewed with NCBI for possible improvements to PubMed.
Other items that we are investigating include:
- adding Suggestions to PubMed search results - these would work similarly to the suggestions in the MeSH and Journals database;
- adding Supplementary Concept Records - the additional chemical terms in MeSH - to the PubMed MeSH database;
- experimenting with full author searching.
PubMed for Handhelds
Many of you are interested in products for handheld devices. NLM's PubMed for Handhelds allows anyone with an Internet-connected PDA, or handheld device, and a Web browser to search PubMed. Although optimized for wireless handheld devices, it can be used with desktop browsers. It's platform independent; it works with Palm, Pocket PC, and Linux PDAs. An XHTML version for mobile phones with wireless access point browsers is available. Just recently, a PICO interface screen was added and we are interested in getting feedback on this.
A separate R & D project for handhelds, PubMed on Tap, is an application for Palm PDAs that retrieves MEDLINE citations directly from the PDA through a wireless connection to the Internet. Since its introduction a year ago, new features have been added in response to user feedback, including e-mailing citations; saving citations to the MemoPad; a clustered results option; and link-out to full text. We hope to release a version for the PocketPC later this spring. [Editor's Note: A Pocket PC version was made available on May 26, 2004.]
Each month the PubMed staff takes a snapshot of LinkOut activity. There are now nearly 4300 PubMed journals that have some or all of their full text articles linked to PubMed citations. Of the nearly 13 million MEDLINE records in PubMed, 26% have full text links. For records since 1999, the number jumps to nearly 77%. The LinkOut for Libraries program continues to be a great success. As of April 1, there were 943 Libraries participating, nearly 300 more than at this time last year.
It's been an impressive year for PubMed Central. At the NLM Update, Kent Smith will fill you in on the Library's successful backfile full text scanning project. Prospectively, great strides have been made as well, no doubt driven by the open access movement that guarantees freely accessible data and the usability of archived documents. As of May 1, there were 148 journals participating in PMC, covering 135,000 articles. PMC was accessed by 700,000 unique users in March, almost three times as many as last March. And I'm happy to report that within the past few days, the PMC homepage was given a new look that we think you will like.
Announcements of new PMC titles and coverage now appear regularly in the NLM Technical Bulletin. Another change relates to electronic-only journals recommended for inclusion in MEDLINE. Because we have seen several instances where a publisher has removed an article from its Web site leaving PubMed users with no full text access, we will require that all new electronic-only journals approved for MEDLINE deposit all articles either in PMC or in another acceptable archive before indexing begins.
Between January 2003 and May 2004, over 1200 librarians and information specialists have taken our PubMed, Gateway, ClinicalTrials.gov, TOXNET and Unified Medical Language System® (UMLS®) training classes offered at various sites across the country. Additionally, this year NCBI began to offer regionally its three-day Introduction to Molecular Biology Information Resources class. We thank the National Training Center staff for handling registration and logistics for these classes and also Renata Geer at NCBI and the rest of the instructors: Donna Messersmith, Kris Alpi, Jennifer Lyon, Mark Minie, and Michele Tennant.
On February 17, the first remote simultaneous broadcast of a regularly scheduled PubMed class took place. Participants at Loma Linda University in California viewed the class taking place in Salt Lake City. Overall, the broadcast was a success. We thank Janet Zipser at NLM, Carlene Drake and her staff at Loma Linda, Claire Hamasu and Sharon Dennis at the RML in Salt Lake City and Maureen Czujak at the National Training Center & Clearinghouse for their coordination of this project. We plan to test this technology again this year at two other sites.
Our NLM colleagues in the Technical Services Division have had a busy year also. About one month ago the 2004 edition of the NLM Classification went up on the Web. It incorporates additions and changes to the schedule and index from August 2003 through April 2004. The hard work didn't stop there. A team of NLM employees has developed specifications for an XML output format for bibliographic data resident in LocatorPlus. In addition to making the XML product available to licensees, the plan is to mount these data under Entrez, the retrieval system for PubMed. Entrez will offer improved search and retrieval functionality that will reinstate some features that are not currently possible in LocatorPlus under Voyager. Users will be able to retrieve an unlimited set of records, "explode" MeSH terms, and perform Boolean operations on subject headings. For complete holdings information, there will be a link from Entrez to LocatorPlus. NLM plans to make the XML data available for distribution in late summer 2004 and the new Entrez database available in early fall.
Because many of you access MEDLINE data from commercial licensees, and others are involved with data mining and research activities, here is a brief update on our data distribution programs. First, I must tell you that new licensees are always amazed to learn that there is no charge to receive more than 14 million MEDLINE and OLDMEDLINE records.
We continued to see growth in the number of MEDLINE licensees, particularly those in academia and from pharmaceutical, biotech and software development companies. There are over 250 MEDLINE licensees and almost 70% use the data for research, particularly about the causes and treatments of
diseases and biothreats.
We currently distribute over 98% of PubMed's content to licensees now that we've been able to get more non-MEDLINE records into our Data Creation and Maintenance System (DCMS - where they need to be in order to be exported). There are fewer than 250,000 additional non-MEDLINE records in PubMed and we intend to get more of them into the DCMS so they, too, can be exported to licensees.
Unified Medical Language System® (UMLS®)
Last year I commented that the UMLS was moving out of the exclusive domain of the research community and becoming a product that can support a number of operational objectives. In the year ahead, we expect much more emphasis on user support - improved documentation, better Web pages, more training and tutorials, and continued excellent responses to customer inquiries. As you may know, since May 6, NLM has distributed SNOMED CT® within the UMLS Metathesaurus under the terms of a revised UMLS license agreement. U.S.-based licensees will be able to use SNOMED without charge and without signing a separate agreement with the College of American Pathologists, SNOMED's producer. However, current UMLS users have to sign the revised license agreement before receiving the next edition of the UMLS.
The new electronic UMLS licensing process is now available. After an NLM official signs electronically, an e-mail message is sent back by NLM with the UMLS license number.
A lot more could be said about UMLS activities but in the interest of time I suggest you look at the NLM Technical Bulletin article, Unified Medical Language System® (UMLS®) Changes. NLM Tech Bull. 2004 Mar-Apr;(337):e9.
It's been another exciting and busy year as you can tell. We couldn't have accomplished so much without supportive Library leadership and an incredible staff. Please stop by our Exhibit Booth, tell us what you like and don't like and what we can do better.
[Editor's Note: See MLA 2004: NLM Online Users' Meeting Questions and Answers. NLM Tech Bull. 2004 May-Jun;(338):e3c. for the text of the questions and answers.]
Now I would like to turn the podium over to my good friend Joyce Backus, who along with her talented colleagues at NLM and in the Network has made MedlinePlus an indispensable tool for medical librarians, public librarians, and the general public. Then Maria Collins will give you an update on what is new with DOCLINE.
By Sheldon Kotzin
Bibliographic Services Division
MLA 2004, NLM® Online Users' Meeting 2004: Meeting Remarks. NLM Tech Bull. 2004 Jul-Aug;(339):e3a.