[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 San Diego, California, on May 5, 2003.]

ello, I'm Sheldon Kotzin and it is my pleasure to welcome you to the NLM Online Users' Meeting. This year we have changed the meeting format somewhat. After I talk for a few minutes about MEDLINE, PubMed, PubMed Central, the NLM Gateway, and some related topics, I will turn the microphone over to Joyce Backus, Public Services Division, who will talk about MedlinePlus, and Maria Collins, Collection Access Section, who will speak briefly about DOCLINE. After that, we'll take your questions.

On March 7, MEDLINE attained a milestone when NLM added its 12 millionth journal citation. This Fiscal Year we expect to add more than 550,000 new citations. This will be 100,000 more citations than we added just three years ago!

Since the last MLA Annual Meeting, 172 new journals have been added to MEDLINE, bringing the total to just about 4,600 titles. In April, nearly seventy percent of all new bibliographic citations and abstracts were received directly from more than 430 different publishers, greatly speeding up their availability in PubMed. Also to speed up access, we are indexing a total of 83 journals from their online versions. Among these 83 titles, 35 exist only in online format, and 48 exist in print format and online. The list of online-only journals includes numerous BioMed Central titles and the Cochrane Database of Systematic Reviews. The list of print journals indexed online includes Annals of Internal Medicine, the New England Journal of Medicine, and Science. All online journals provide NLM with XML citation and abstract data. Once a print journal starts publishing online-only content, NLM switches its indexing from the print version to the online version.

Gene Indexing
Gene indexing refers to links that NLM indexers make from MEDLINE citations whose articles discuss gene function or structure to LocusLink, an NCBI database of protein information. Three new organisms have been added to LocusLink in the last year. HIV-1 was added in May. Cow and C elegans (Nematode) were added in December, bringing the total organisms to eight.

Approximately fifty nine percent of the links are for human, twenty seven percent for mouse, ten percent for rat, and three percent for fruit fly. Approximately one percent of the links are for HIV-1, Cow, Zebrafish and C elegans combined. More than 600 links per week are now being created. As of the end of March, indexing staff have performed gene indexing for a total of 34,012 items, since the inception of this project during the summer of 2001.

Indexing
Medical Text Indexer, or MTI, is the name we give to a relatively new program developed at NLM. It automatically suggests MeSH terms for articles, based on terms appearing in the title and abstract. Indexers may use any of these online suggestions or reject them. We are now testing a feature that allows indexers to search for any term from an article, and the Medical Text Indexer system will return a list of likely MeSH terms.

MTI is reliable enough to be used without human intervention for the indexing of meeting abstracts that appear in the NLM Gateway. For MEDLINE indexing, it is still used only as a tool to assist human indexers and in that capacity it performs quite well.

In response to the emerging epidemic, two new terms have been added to MeSH mid-year, a rarity for NLM. Severe Acute Respiratory Syndrome was created on March 25. The acronym "SARS" was not made a cross-reference for this term, because SARS has almost a dozen other established meanings, such as acronyms for structure-activity relationship and scaffold attachment regions (genetics). Severe Acute Respiratory Distress Syndrome is treed under Respiratory Tract Infections and Coronavirus Infections. The other new heading is for the organism that causes the illness, the SARS Virus. Also, just four days ago, we added a temporary feature to the PubMed Home Page that facilitates retrieval of SARS citations.

NLM Gateway
Many improvements to the NLM Gateway have been made since last year. There is now access to ClinicalTrials.gov. Subsets for Bioethics, AIDS, History of Medicine, and Space/Life Sciences allow cross-file searching of journal articles, books, and meeting abstracts on these important subjects. Abstracts from five AIDS meetings, including 8,800 abstracts from the 14th International Conference on AIDS, and three Health Services Research meetings were added to the Gateway's meeting abstracts collection. Records for 600 health services research projects were also added, and within the next six months we expect to add access to HSDB and TOXLINE Special from the Gateway.

OLDMEDLINE
I am delighted to announce that NLM has added about 430,000 OLDMEDLINE citations from 1953-56 to the Gateway. We are digitizing records from 1950-52 and I am hopeful they will be searchable by the end of this calendar year. More good news is that the Library's project to make available the citations in the Index Catalogue is nearing completion. By this September, we expect 3.5 million references published from 1880 through 1961 to be available on the Web. Librarians, historians, and researchers will be able to search subjects, authors, titles, dates, and notes. They can search for individual volume, specific series, and across the entire Catalogue.

PubMed
There have been many great enhancements to PubMed this past year. These were explained in the NLM Technical Bulletin, but I'd like to acknowledge some of the highlights.

We replaced the Journal Brower with a Journals database which, in addition to the search results, is able to offer suggestions that we think are quite helpful when looking for a PubMed journal. It also has direct links to the LocatorPlus record. In addition, the MeSH Browser is now the MeSH database. It provides more information about each MeSH term and has additional capabilities including providing suggestions to help you locate MeSH terms.

We were also happy to announce that many more British spellings of words in MeSH terms are now recognized. Another enhancement affects Publication Types. As you may know, many Publication Types have MeSH equivalents. For example, Clinical Trial is a Publication Type and the plural form, clinical trials, is a MeSH term. If a searcher uses either the MeSH term or the Publication Type, PubMed now automatically includes both terms to optimize retrieval. And finally, and maybe most eagerly awaited, you can now email PubMed search results.

In June we will celebrate the sixth anniversary of PubMed. We feel we have made -- with your comments and suggestions -- significant progress in improving it. More than 200,000 individual users search it, with well over one million searches each day. This comes to 35 million searches of PubMed each month and over 400 million projected this year. About 50% of this use is outside the United States.

LinkOut
One of those improvements that has been most welcome as your participation illustrates is LinkOut. There are now more than 3,700 journals that have some of their full-text articles linked by PubMed citations. Of the more than 12 million MEDLINE records in PubMed, 23% have full-text links. For records from 1999-2003, the number with links jumps to 74%. A new icon feature of PubMed allows you to see which citations link to the full text at no charge on PubMed Central. [Editor's Note: For more information see: Nahin AM. New Icons in PubMed^® Indicate Free Full Text Available. NLM Tech Bull. 2003 May-Jun;(332):e6.]

The LinkOut for Libraries program continues to be a great success, with 645 libraries participating. I hope that you know that any SERHOLD library now can display icons for its print collection. A new tutorial, Setting Up LinkOut for Libraries, is a step-by-step guide on how to set up LinkOut at your library. Topics include getting an account, entering library information, entering electronic holdings, adding SERHOLD print holdings, and displaying the library icons in a PubMed search. Efforts are ongoing to expand the capability to display print holdings information for non-SERHOLD libraries.

Future Plans for PubMed
Future plans include: Corporate author names will soon be included on all display formats and you can also search for these names. Many citations contain a corporate name, for example the Human Genome Group, and a list of group members in the author field. In those cases, PubMed will carry two separate author fields and both will be searchable.
[Editor's Note: This feature has been implemented. For more information see: Changes to PubMed^® for 2003. NLM Tech Bull. 2003 Mar-Apr;(331):e14. ]

The date a citation was published electronically ahead of the printed publication will be retained with the citation in PubMed. As you may be aware, it has become quite commonplace for journals to publish electronic versions of articles several weeks ahead of print. This feature will help users find the earliest reporting of new research findings. Also added to the Summary format are more of the Comment links, including those for retractions, errata, and patient summaries -- the latter was just introduced in MEDLINE last year.
[Editor's Note: This feature has been implemented. For more information see: Changes to PubMed^® for 2003. NLM Tech Bull. 2003 Mar-Apr;(331):e14. ]

Other PubMed enhancements will be made to the History feature and to the Cubby and we are in the early stages of examining the feasibility of adding OLDMEDLINE data to PubMed.

Some things that you have asked for are in our planned enhancements, so we appreciate your continued suggestions.

CANCERLIT
A number of you have asked questions about the recent decision by the National Cancer Institute (NCI) to cease its CANCERLIT database and its cancer.gov Web site, and link its users directly to PubMed. This was NCI's decision, but one which NLM supported and cooperated in achieving. NCI's database was nearly 100% MEDLINE citations, but it lacked PubMed's publisher-supplied records, access to full-text, other LinkOut features, access to related-articles, and the like.

NCI staff, working with NLM, has developed a new search strategy to define a cancer subset within PubMed. This is comparable to existing NLM subsets on bioethics, complementary medicine, and other topics. On its own Web site, NCI will maintain an extensive set of specialized cancer topic search strategies returning relevant sets of PubMed citations. In a future development, NCI will make the neoplasm terminology in its thesaurus available to NLM. NCI will also provide access to abstracts of the major cancer research meetings, as NLM has no plans to add these abstracts to PubMed or the Gateway, but we will be observing how convenient this access is on the associations' Web sites.

PubMed Bookshelf
The PubMed Bookshelf, full-text electronic books available for searching directly or as a link from PubMed retrieval, continues to grow. More than 20 books are already available. Some are electronic versions of printed books, while others exist only in electronic form. Several additional books are in the pipeline and will be added over the next several months. Most of the current books are on molecular biology topics but we recently added NIH publications on obesity, asthma, and high blood pressure and expect soon to include other biomedical topics. If you have not examined the Bookshelf in a while, it is definitely worth another look.

PubMed Central (PMC)
I think that it's important that librarians know of the impressive strides that NLM and its cooperating publishers have made with PubMed Central during the last year. PubMed Central meets the two basic tenets of open access -- its data has been made freely accessible by publishers and it guarantees that the archival documents will be usable even as standards for digital documents change. One hundred and twenty journals participate in PMC and it covers 90,000 articles. It is currently accessed by 240,000 unique users each month. This year PMC became another NLM database using the Entrez retrieval system. Not only is the content accessible via PubMed, but it is directly searchable for articles whose full-text is freely available. Moreover, you can link back to PubMed to see references cited by these articles. Links exist, as well, to scientific databases maintained by NCBI. A detailed description of these improvements will appear in the NLM Technical Bulletin in the next few days. [Editor's Note: This article is now available. Please see: Sequeira E. PubMed CentralTM in Entrez. NLM Tech Bull. 2003 May-Jun;(332):e3.]

In an exciting development, arrangements have been made with publishers of approximately 70 PMC journals to digitize their back runs, in some cases into the 19th century. The Journal of the Medical Library Association was one of the first publications to come forward to offer its older issues (back to volume one) for digitization. The first phase, when complete, will add about a half-million articles to PMC.

Training
Between January 2002 and May 2003, over 1200 librarians and information specialists have taken our PubMed, Gateway, ClinicalTrials.gov and TOXNET training classes offered at various sites across the country. The National Training Center and Clearinghouse's Education Clearinghouse database debuted in June of last year and currently contains a total of 126 educational resources. Materials include: PowerPoint presentations; manuals for PubMed; description of classes on consumer health, public health, TOXNET, and MedlinePlus; fact sheets; and newsletters and user guides. A user study to evaluate the Educational Clearinghouse database is planned for this year. [Editor's Note: This evaluation took place during the month of June.]

We have begun preliminary planning for videostreaming a PubMed training class in October. We are excited about testing and evaluating this technology. And we thank Claire Hamasu (Midcontinental Region), Elaine Graham (Pacific Southwest Region), and Carlene Drake (Loma Linda University) for working with us on this project.

Also, near the end of the calendar year, we will have regional offerings of the 3-day class, "Introduction to Molecular Biology Resources." This course has been taught at NLM several times. Stay tuned for an announcement in the NLM Technical Bulletin and for registration via the National Training Center and Clearinghouse.

Unified Medical Language System (UMLS)
There is another training class that I want to tell you about and it helps me segue into a discussion of UMLS. We taught a basic course on the UMLS for medical librarians and will be teaching it again at NLM on June 16 and September 15. You can register from the National Training Center and Clearinghouse site. The course also will be taught next May, at the Annual MLA meeting in Washington. We plan to create training packages or modules that could be offered on DVD or on the Web. This effort is being led by Carolyn Tilley, who became the advisor on UMLS activities a few months ago after being Head of the MEDLARS Management Section for several years. She and Jan Willis have done an excellent job expanding access to UMLS data.

You could say that the UMLS is moving out of the exclusive domain of the research community, into more of a product that can support a number of operational objectives, including those mandated by the Health Insurance Portability and Accountability Act of 1996 (HIPAA).

One UMLS-related item that you will hear more about in the NLM Update is NLM's discussion with the College of American Pathologists about the clinical vocabulary, SNOMED. Betsy Humphreys will tell you about SNOMED, and will join Dr. Lindberg, and Dr. Jack Snyder, our new Associate Director of Specialized Information Services, as the speakers at the update. [Editor's Note: SNOMED Clinical Terms^® To Be Made Available in the UMLS^®. For more information see: Technical Notes. NLM Tech Bull. 2003 Jul-Aug;(333):e1.]

Data Distribution
This has been another year of tremendous growth in NLM's data distribution, or leasing program. Currently there are over 180 organizations or institutions licensing MEDLINE. NLM's toxicology databases created by our Specialized Information Services Division and the cataloging and serials data produced by NLM's Technical Services Division are also in demand, and we look forward to distributing OLDMEDLINE soon to about 40 licensees waiting for that data in XML format. [Editor's Note: For more information see the announcement: OLDMEDLINE Available for Distribution in XML Format from NLM^®.]

In the past, MEDLINE was leased primarily by database vendors for commercial purposes. Today, the overwhelming majority of licensees are academic institutions, biotechnology companies, and pharmaceutical firms with high-speed computers who lease the data for research purposes. MEDLINE is a rich source for data mining and other research applications and NLM is pleased to provide MEDLINE at no cost to interested parties.

Conclusion
It's been another exciting and busy year as you can tell. As always, we value your suggestions and criticisms. [Editor's Note: See MLA 2003: NLM Online Users' Meeting Questions and Answers. NLM Tech Bull. 2003 Jul-Aug;(333):e7b. for the text of the questions and answers.]

Now I would like to turn the floor over to my good friend Joyce Backus, who along with her talented co-workers at NLM and in the Network, has made MedlinePlus an indispensable tool for medical librarians. Then Maria Collins will give you an update on what is new with DOCLINE.

By Sheldon Kotzin
Bibliographic Services Division

NLM Online Users' Meeting 2003: Remarks. NLM Tech Bull. 2003 Jul-Aug;(333):e7a.