The baselines are normally generated towards the middle of November each year and contain all completed citations in MEDLINE as of that date. The baselines represent MEDLINE after the year-end processing has been completed. This means that the records have been revised with the upcoming year's new MeSH vocabulary terms. We currently have available the 2002 - current year MEDLINE/PubMed Baselines. The naming of the baselines represents this year-end processing. For example, the 2002 MEDLINE/PubMed Baseline contains all completed citations from the mid-1960's until the date the baseline was created in late November 2001 with the year-end processing assigning appropriate 2002 MeSH vocabulary terms, thus it is a baseline for the 2002 year.
The baselines also contain citations that are not MEDLINE. All of the baselines we have stored (2002 on) contain "Out-of-scope" citations which were renamed to "PubMed-not-MEDLINE" starting with the 2004 MEDLINE/PubMed Baseline. The PubMed-not-MEDLINE status refers to citations that reside in PubMed from journals included in MEDLINE and have undergone quality review but are not assigned MeSH headings because the cited item is not in scope for MEDLINE either by topic or by date of publication. Citations in the Out-of-scope or PubMed-not-MEDLINE status make up a very small percentage of the total number of citations contained in the baselines (For example, 0.51% or 75,271 records in the 2005 baseline and 1.8% or 323,919 records in the 2009 baseline).
Starting with the 2005 MEDLINE/PubMed Baseline, OLDMEDLINE citations are also included in the baselines. The OLDMEDLINE citations make up approximately 11% of the total number of baseline citations. The OLDMEDLINE citations are from international biomedical journals covering the fields of medicine, preclinical sciences, and allied health sciences. The citations were originally printed in hardcopy indexes published prior to 1966. For additional information, please refer to the following URL: https://www.nlm.nih.gov/databases/databases_oldmedline.html.
In the 2005 baseline the subject indexing from the OLDMEDLINE citations were stored solely in the "Other Term" (or "OT") tagged fields and not the MeSH Terms (or MH) tagged fields. This means that searching the 2005 baseline from our MBR Query Tool via the MH field does not include any OLDMEDLINE citations. The only way to include OLDMEDLINE records in the 2005 baseline is to do a timeframe query without specifying any field specific search criteria. Beginning with the 2006 baseline, Other Terms are starting to be mapped to current MeSH Terms so that searching via the MH field may retrieve some OLDMEDLINE records, but, not necessarily the complete set of possibilities.
Starting with the 2007 MEDLINE/PubMed Baseline, on records where all the OLDMEDLINE terms are converted to MeSH Headings, the citation status changes to MEDLINE. You need to rely on the <CitationSubset>OM</CitationSubset> element to determine if a citation is in the OLDMEDLINE subset.
Select the Baseline year to see the directory listing for that MEDLINE/PubMed Baseline allowing you to download all of the files for that baseline year. The DTDs link is for a gzipped tar file containing all of the required DTD files for that year's baseline files.
|Baseline||Created||# Files||# Citations||DTD Files|
|2017||December 13, 2016||892||26,759,399||DTDs|
|2016||November 20, 2015||812||24,358,442||DTDs|
|2015||November 24, 2014||779||23,343,329||DTDs|
|2014||November 21, 2013||746||22,376,811||DTDs|
|2013||November 15 & 16, 2012||717||21,508,439||DTDs|
|2012||November 18, 2011||684||20,494,848||DTDs|
|2011||November 19, 2010||653||19,569,568||DTDs|
|2010||November 20, 2009||617||18,502,916||DTDs|
|2009||November 21 & 22, 2008||593||17,764,826||DTDs|
|2008||November 16 & 17, 2007||563||16,880,015||DTDs|
|2007||November 17 & 18, 2006||538||16,120,074||DTDs|
|2006||November 18 & 19, 2005||516||15,433,668||DTDs|
|2005||November 20, 2004||500||14,792,864||DTDs|
|2004||November 14-18, 2003||417||12,421,396||DTDs|
|2003||November 1-4, 2002||396||11,847,524||DTDs|
|2002||Approx. November 21, 2001||379||11,299,108||DTDs|
We generate a large number of data files during our normal processing of each set of baseline files. We make available the files that we think others might be able to use with the goal of trying to reduce any duplication of effort.
The MeSH FTP download site: ftp://nlmpubs.nlm.nih.gov/online/mesh/ now includes separate directories for each release year of MeSH. In addition, MeSH created the folder "MESH_FILES" with the latest release files that are updated every morning Monday - Friday. The yearly release folders span from 2011 to the latest full release which occurs in November of the preceding year (for example, 2016 MeSH was released in November of 2015). A single directory is also included for earlier files from 1999-2010.
Semantic Types and Groups: A parsable list of Semantic Types and their abbreviations from the UMLS and a parsable list of Semantic Groups and their mappings to the Semantic Types.
Semantic Types and Groups
The MEDLINE N-gram Set from The SPECIALIST Lexicon. N-grams of size 1 - 5 are identified from all of the Title and Abstract for each MEDLINE citation in baseline.
Latest information on the SPECIALIST Lexicon MEDLINE N-Gram Set