Skip to main content

PubMed Tutorial from The Levy Library: 1. MEDLINE/PubMed Basics

See Also

More on the difference between MEDLINE and PubMed.

PubMed vs. Google

Searching PubMed is very similar to searching internet, right?

PubMed and Google each have their advantages.

Google and Google Scholar offer excellent relevance ranking algorithms, however they do not allow for targeted searching or offer effective filtering options (such as the ability to filter results for randomized controlled trials only).

The MEDLINE indexing structure of PubMed allows for more precise, targeted search queries.  Learning about the structure of PubMed can enable you to utilize that structure to your advantage.

Example: Because there is no option to search the author affiliation field only, the only way to try to locate recent publications by Mount Sinai authors in Google Scholar is to run a basic search for Mount Sinai, which does not yield comprehensive results.  Additionally Google Scholar's results ranking algorithm includes citation data (i.e. how often an article has been cited by other articles) and does not allow user to sort results purely by date. By contast, have a look at a well designed PubMed search on the same topic.

Intro to PubMed and MEDLINE

PubMed is a database of over 27 million biomedical article citations created and maintained by the National Library of Medicine.  MEDLINE is the subset of PubMed records which have been indexed by the National Library of Medicine and have had Medical Subject Headings (MeSH) terms applied. Currently, approximately 80% of PubMed records are indexed for MEDLINE.  

Subject Coverage: Biomedicine and the health care sciences. MEDLINE also covers medically relevant fields of the biosciences such as microbiology, immunology, virology, and neuroscience

Publication Date Coverage: Approximately 1950 to present.  The earliest citations in PubMed date back to the late 1940s, but consistent coverage and indexing began in earnest in the early 1950s.

Record Content: PubMed records are comprised of fields, or "data elements", such as Author, Title, Publication Date, Abstract, MeSH Terms, etc. A PubMed record can have as many as 60 different fields. It is important to note that PubMed is a citation database, not a full text database so when searching PubMed, you are searching the citation information contained in these fields only, not the entire text of the article. 

Sample PubMed Citation

Record Nuts & Bolts

Nuts and boltsPublishers usually upload citation records to the PubMed database at which time each record is assigned a PubMed Unique Identifier (PMID).

When a citation is indexed for MEDLINE, professional indexers working for the National Library of Medicine read the full article (not just the citation) and apply appropriate MeSH Terms according to a set of indexing rules. 

How quickly a citation is indexed for MEDLINE once it has been added to PubMed varies significantly from journal to journal.  Citations for articles published in journals such as NEJM, JAMA, and Annals of Internal Medicine (i.e. heavily read, highly cited, “high impact” journals) tend to be indexed within 2-3 weeks.  Citations for articles published in less well known journals can take 6+ months to be indexed.