Technical Notes - e1

Year-End Processing [corrected 1999/12/02] - e2

MeSH Coming Attractions - e3

Fees and New Format for Leasing NLM Databases in 2000 - e4

Scope Expands for PubMed and MEDLINE - e5

Data Changes Expected - e6

Hands On - e7

2000 Update Schedule for MEDLINE on PubMed and Internet Grateful Med [corrected 1999/12/02]

This time of year the NLM Technical Bulletin traditionally includes information on changes made to MEDLINE and other NLM databases during the NLM internal maintenance cycle known as Year-End Processing. This type of information is now presented from the PubMed and Internet Grateful Med (IGM) searcher's viewpoint.

What changes will I see?

1. the annual update to Medical Subject Headings (MeSH), NLM's controlled vocabulary used for subject indexing and retrieval;
2. updated MeSH in MEDLINE citations to reflect changes in MeSH vocabulary;
3. other changes to data in MEDLINE citations including EC/RN, Publication Type, Title Abbreviation and ISSN information;
4. resumption of SDILINE updates (searched via Internet Grateful Med), which are suspended from late October through mid-December for the final phase of year-end processing.

For MEDLINE searchers who used to search ELHILL, some things have changed, and some things remain the same. All MEDLINE searchers, new and old, are encouraged to explore the new PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi), and to consult the Help and FAQs from the PubMed sidebar for more information on its exciting and powerful features, including several new features made available September 1999.

Annual Update to Medical Subject Headings: 2000 MeSH
A summary of changes in 2000 MeSH is available at: [This link was removed because it is no longer valid]. The MeSH Section's MeSH Browser (http://www.nlm.nih.gov/mesh/MBrowser.html) currently contains 1999 MeSH with a link to a version with the 2000 MeSH. Searchers should consult the MeSH Section's MeSH Browser to find descriptors of interest and to see these in relationship to other descriptors. This Browser displays virtually complete MeSH records, including the scope notes, annotations, entry vocabulary, history notes, allowable qualifiers, etc. It also provides links to relevant sections of the NLM Indexing Manual. Also, see the article MeSH Coming Attractions in this issue of the Technical Bulletin for MeSH 2000 highlights.

Changes to MeSH Terms
During year-end processing, MeSH Terms in MEDLINE citations indexed from 1966-1999 are updated to reflect changes in 2000 MeSH. These changes are expected in PubMed by mid-December.

For example, the MeSH Heading "Heart Hypertrophy" has been changed to "Cardiomegaly" in 2000 MeSH. Thousands of MEDLINE citations indexed from 1966-1999 contain the MeSH Term "Heart Hypertrophy," all of which will be maintained to "Cardiomegaly" for the 2000 MEDLINE database in PubMed.

Remember that changes in the mapping of see references can also occur. For example, "Stroke" has been removed as, "See Cerebrovascular Disorders". For 2000, it is "See Cerebrovascular Accident" which has the History note of "for STROKE and APOPLEXY use CEREBROVASCULAR DISORDERS 1963-1999".

New MeSH Terms
Five hundred and fifty-two (552) new MeSH Subject Headings have been introduced in 2000 MeSH. They are listed with scope notes and previous indexing information on the NLM Web site at: http://www.nlm.nih.gov/mesh/old/MBrowser.html. These terms will appear in citations beginning with the first MEDLINE update for 2000.

NLM never retrospectively indexes MEDLINE citations with new MeSH Headings. Therefore, searching for a new MeSH Term qualified as [MeSH Term] or [Major MeSH Topic] effectively limits retrieval to citations indexed after the term was introduced. An unqualified subject search in PubMed expands a search by including both MeSH Term and text words, and may retrieve relevant citations indexed before the introduction of a new MeSH Term.

For example, a new MeSH term, "Senile Plaques" was introduced in 1998 MeSH. A PubMed query for "Senile Plaques" qualified as [MeSH Term] yields 237 citations, indexed from 1998 through October 1999. A simple, unqualified PubMed query for "Senile Plaques" yields 1772 citations from 1966 through November 10, 1999.

As these new 2000 MEDLINE citations are being indexed, NLM expects to add them to PubMed. We no longer must wait for two months to release these citations for searching. See Appendix A for the 2000 MEDLINE Update Schedule. [Editor's Note: See the Technical Note Revised: 2000 Update Schedule for MEDLINE on PubMed and IGM for the latest information.] However, PubMed's MeSH Browser and PubMed's translation tables that are used for PubMed's automatic term mapping process are still operating against 1999 MeSH. PubMed's MeSH Browser and translation tables will be updated to reflect 2000 MeSH in mid-December while IGM's Metathesaurus is typically updated after the start of the new calendar year. You may wish to use the MeSH Section's 2000 MeSH Browser during this interim period to find new concepts and then copy and paste terms from there into your PubMed or IGM search strategy.

Supplementary Chemical Concepts: EC/CAS Registry Numbers [EC/RN]
Changes in 2000 MeSH also require changes to some EC/RN data, including Substance Name [NM], in MEDLINE citations during year-end processing. These affect not only the MeSH chemical concepts in category D, but the Supplementary Chemical Concepts as well.

NLM began indexing EC/RN data in MEDLINE citations in June 1980. Limiting a PubMed query by a specific EC (Enzyme Comission Number), RN (Registry Number), or Substance Name automatically excludes retrieval of citations indexed before this date.

Tumor Key
The MeSH Section has announced that the Tumor Key will no longer be published; it has not been updated since 1993. The Tumor Key information is now incorporated in MeSH.

Preliminary Citations in PubMed
MeSH terminology provides a consistent way to retrieve information that may use different words for the same concepts. These terms can be used in many cases to search PubMed more efficiently than simple text words. However, limiting a PubMed search by MeSH Term or MeSH Major Topic also effectively limits a search to fully indexed MEDLINE citations. PubMed also contains, and is updated daily with, preliminary citations received directly from publishers, and citations in the MEDLINE indexing workflow, which are excluded from retrieval when subject searching is limited to MeSH Terms.

Publication Type and EC/RN along with NM [Substance Name] are also added to citations during indexing. Limiting a PubMed query to these fields also effectively excludes citations not yet indexed in PubMed.

Other Changes in MEDLINE 2000

MEDLINE Unique Identifiers
NLM is in the process of implementing a new system for creating and maintaining MEDLINE citations. For the immediate future, MEDLINE Unique Identifiers [UI] will follow their familiar 8-digit format. Citations entering the current MEDLINE citation processing system after October 1999 will begin with "20," e.g., UI: 20000001.

The new Data Creation and Maintenance System (DCMS) will also initially use an 8-digit UI. However, citations processed through the DCMS will have UIs starting with "21". This means that citations entering MEDLINE during the 2000 indexing year will have a mix of UIs - some beginning with "99" as these citations were in the indexing queue prior to October 1999; some beginning with "20"; others beginning with "21".

For the 2001 indexing year, the new system will transfer to using the PubMed Identifier (PMID). The PMID is displayed in PubMed without leading zeros, e.g.:

     PMID: 1
     PMID: 1000
     PMID: 10000000

Stay tuned for more details about other data changes expected when the DCMS is implemented.

Identification of Clinical Trials in MEDLINE
For the fifth year, NLM continues its work with the Cochrane Collaboration to enhance the identification of clinical trials in MEDLINE. During the 1999 maintenance cycle, the Cochrane Collaboration provided enhanced clinical trial Publication Type [PT] information for over 3,000 MEDLINE citations to articles published from 1966-1999.

Points to Remember

MEDLINE Citations 1966-2000 in PubMed
PubMed contains the complete MEDLINE database of over 10,000,000 citations to articles published from 1966 through the current year. See Appendix A for the 2000 MEDLINE Update Schedule. An unrestricted PubMed query searches the entire MEDLINE database of completed and in-process citations. PubMed searchers may, however, limit retrieval to a specific Publication Date year or range of years by using the Limits feature in the new version of PubMed. Also, remember as explained earlier in this article, that NLM plans to add citations currently being indexed for the 2000 indexing year in advance of the year-end processing changes expected in mid-December. [Editor's Note: See the Technical Note Revised: 2000 Update Schedule for MEDLINE on PubMed and IGM for the latest information.] Remember that PubMed's MeSH Browser and translation tables are expected to be updated in mid-December. MEDLINE via IGM uses the PubMed version of the database; the IGM Metathesaurus is updated with the new MeSH after the start of the calendar year.

MEDLINE in PubMed Now Includes HealthSTAR
HealthSTAR journal citations are also available in PubMed, and significantly expand MEDLINE's coverage of health services research, healthcare and health administration literature. HealthSTAR and its Backfile, HSTAR75, are also searchable as separate databases via IGM.

Year-End Processing and Other MeSH-Indexed Citation Databases in IGM
Following MEDLINE, year-end processing is completed against other NLM databases in the following order: HealthSTAR, AIDSLINE, HISTLINE, SPACELINE, HSRPROJ, BIOETHICSLINE, and POPLINE. Eventually, the journal citations from these files will be added to PubMed while the monographic citations will be added to LocatorPlus.

Until the data are transitioned, these files remain accessible via IGM, although new updates are suspended until after maintenance is completed. Searchers are reminded that until each file is maintained, 1999 MeSH should be used for subject searching. The table below lists the date each file will be available in IGM with 2000 MeSH vocabulary.

Database Update Information available on the IGM Home Page [Editor's Note: This link was removed because it is no longer valid.] (http://igm.nlm.nih.gov/splash/IGM_dbupdate.html) provides current information on database status.

Coming Features: PubMed, IGM, Next Generation Gateway

STORESEARCHES, Automatic SDI and SAVED SEARCHES
The capability to store searches on the NLM computer, or have stored searches run automatically against updates to NLM databases ceased with public billed access to ELHILL on September 30, 1999. Exciting new features in both PubMed and the Next Generation Gateway are in development to replace, and expand on these legacy system functions.

An article on converting SDI searches to PubMed appeared in the May-June 1999 NLM Technical Bulletin at: http://www.nlm.nih.gov/pubs/techbull/mj99/mj99_sdi.html. As this article explains, running a search in PubMed and limiting the Entrez Date to the last 30 days does not equate to the date range of the SDILINE file.

The last paragraph of this article discusses the proposed "Cubby" service that will offer a SDI-like capability. This new service is still under development although tentative plans call for an additional date field that will change when a PREMEDLINE record is elevated with MeSH Headings to MEDLINE.

Next Generation Gateway
For information on features in development under the working name of the Next Generation Gateway, see the March-April 1999 NLM Technical Bulletin at: http://www.nlm.nih.gov/pubs/techbull/ma99/ma99_gateway.html

--prepared by Jan Willis

MEDLARS Management Section

Next article

Table of contents

Home

Index