Hours Today: 11/23/09
7:30am-midnight | see all hours

Ask a Librarian



What is PubMed?

PubMed is an online index to the biomedical journal literature that is available free worldwide via the internet.  It was developed by the National Center for Biotechnology Information (NCBI) at the National Library of Medicine (NLM), located at the United States National Institutes of Health in Bethesda, MD.  PubMed provides citations and abstracts to journal articles, and in many cases the full text of the articles. 

The MEDLINE database, produced by the National Library of Medicine, is the largest component of PubMed and therefore the two terms are often used interchangeably.  However, a relatively small proportion of records in PubMed are not part of MEDLINE.  PubMed receives citations to newly published articles directly from publishers immediately upon publication.  Until the records are processed and annotated (which may take 1-3 weeks), they are not formally entered into the MEDLINE database.  Out-of-scope citations from general science publications from which the life-science articles are entered into the MEDLINE database remain in PubMed but are not part of MEDLINE.  PubMed also contains a few additional subsets of records that are not in MEDLINE. 

PubMed contains over 16 million records dating from the 1950s to the present and indexes over 4800 journals published in the United States and 70 other countries.  PubMed covers the following fields:

  • all health sciences including medicine, nursing, dentistry, allied health, psychiatry, health care administration, and veterinary science
  • the pre-clinical sciences such as biology, chemistry, biochemistry, genetics, and microbiology
  • related disciplines such as biomedical engineering, nutrition, psychology, and social work.

How do I access PubMed?

You can access PubMed directly from http://pubmed.gov from any computer with internet access.  However, if  you access it by going through the Dana Medical Library homepage or through the UVM College of Medicine's COMET (for COM faculty, staff, and students only), UVM's holdings will automatically be displayed. 

From Dana Medical Library's homepage (http://library.uvm.edu/dana), click on the link for PubMed under Databases & E-Books.


 Basic PubMed Searching

The simplest way to search PubMed is to enter a word string into the search box and allow the system to automatically perform the search for you.  This "type and go" method works well for straightforward, highly focused clinical questions.  Select terms that clearly and concisely express the key concepts, usually a single word or short phrase. By default, terms are combined using the Boolean operator "AND".

Suppose you want to search for articles on the use of NSAIDS (non-steroidal anti-inflammatory drugs) and the prevention of strokes. 

  • Open PubMed.
  • In the search box, enter the word string "stroke prevention nsaids". Click GO. 
  • PubMed will return a list of citations in the default Summary format.
  • The results will be displayed in reverse chronological order, with the most recently added articles listed first.


Fig. 1.  The PubMed result page, showing the first two of 1350 citations.


Click on the Details tab below the search box to find out exactly how PubMed performed this search.    

Fig. 2.   Contents of Details tab following basic PubMed search for "stroke (and) prevention (and) NSAIDS."


The Details tab reveals that PubMed translated the word string into a complicated search strategy utilizing nesting of terms and the Boolean operators AND and OR.  Working from left to right, it searched on each term sequentially, first as a text word and then using the official Medical Subject Heading (commonly referred to as "MeSH Headings") appropriate for that term.  This feature, called "Automatic Term Mapping", saves the searcher the tedium of having to key in every possible variation of the term including plurals, spelling variations, all possible word-order combinations, and all possible synonyms in order to bring up all relevant articles in the database.   

Check the Details tab each time you perform a search in PubMed, as it will identify misspellings, dual meanings, and unintended uses of terms.  For example, if you wish to investigate the inheritance patterns of a given disease, you will quickly learn that the word "inheritance" maps to the MeSH term "wills", a meaning far from what you'd intended!  Using the term "genetics" will yield much more relevant results. 

After reviewing the contents of the Details tab, click your browser's Back button to return to the original page, and proceed with your search.


Limits

Very often PubMed brings up far more hits than you have time or energy to wade through.  By applying a few limits, you can reduce this number to a smaller and more focused set of results.  Knowing PubMed contains citations to articles in non-English languages as well as to non-human studies, applying limits for English language and Humans would be a good first start.  Limiting to Review articles will further reduce the numbers.  Finally, limiting the search to Core Clinical Journals will limit your results to citations from highly regarded, readily available publications, most of which Dana Medical Library subscribes to in print or electronically.  Other available limits include topic, gender, and age group.  It is possible to limit by publication year, although this is generally not necessary since the most recently added citations will be displayed first.  

Fig. 3.   Limit page showing multiple pull-down boxes. 


By applying these four limits to your search, the total number of citations is reduced to 67 hits, a very manageable number.  The limits you selected will remain in force until you un-check the Limits box, go into the Limits tab to change your selections, or close out PubMed.


Fig. 4.   Results of the search after applying limits.


Viewing results

The results are separated into three tabs labeled All, Dana Medical Library, and Free Full Text.  Articles under the Dana Medical Library tab are contained in publications owned or subscribed to by Dana Medical Library or the University of Vermont.  These tabs are set by Dana Medical Library and will display for all PubMed searches run through the library's webpage.  You can set up your own tabs by logging into your MyNCBI account, described below in the Customizing Your PubMed Display Through MyNCBI section. 

PubMed results are displayed by default in the Summary form.  To view the complete PubMed record for one citation, click on the author name.  To view the complete PubMed record for an entire page of results, select Citation in the Display drop-down menu.  You can change the number of records displayed on each page by changing the number in the Show drop-down menu.  The Citation display for the first result obtained in our sample search is illustrated below.


Fig. 5.   A typical PubMed record in the Citation format including icons for full-text articles, abstract, and MeSH terms.


As indicated by the green UVM icons, this article is available electronically and in print at UVM.  To access the article electronically from on-campus, click on the UVM E-Journals icon.  A new window for the publisher's web site will open.  Click on the PDF link.  With a few exceptions, there are no passwords required to access UVM's electronic journals.

To access full-text articles electronically from off-campus, you must be a UVM or FAHC faculty, staff, or student.  Please refer to the Connect from Off-Campus page for additional instructions.  (http://library.uvm.edu/dana/guides/connect/index.html) 

Unaffiliated patrons may access articles in electronic journals only from a computer on the UVM campus.

To determine the University's print holdings, click on the  UVM Journals icon to access the UVM Library Catalog.  Journal issues from 1986 to the present are shelved in Dana Medical Library.  Issues from prior to 1986 are located in the Library Research Annex.


Related Articles

Each citation brought up in PubMed will have a Related Articles link on the right margin.  Clicking on this link will launch a pre-formulated search on the topic.  Results will be displayed in relevancy-ranked order rather than in reverse chronological order, with the original article listed first.  This is an excellent tool to find similar articles to the one good article you started out with and does not require knowledge of subject headings or use of the more sophisticated PubMed tools.   Following the Related Articles link repeatedly is one strategy to ensure that you have found all the relevant articles on a given topic. 


Print, E-mail, and Save Results

When you ask PubMed to print citations, it will print them in the display format selected (e.g. Summary, Abstract, Citation, etc.,), but will not print the full text of the articles.  To print the full text of the articles, you must click on the UVM E-Journals icon for each individual article.  You must either be on campus to obtain access to the full text of articles or be a UVM or FAHC-er to obtain off-campus access.  

By default, one "page" in PubMed consists of 20 citations.  PubMed will only print one page at a time, not the entire set of search results.  For sets larger than 20, use the Show drop-down menu to increase the number of citations per page. 

To print an entire page of results, select a display format, then click on the Send To drop-down menu and select Printer.  You must have pop-ups enabled for this feature to work.  To generate a text-only print-out, select Text from the Send To drop-down menu, followed by your browser's print function.  

To select individual records to print, check the boxes to the left of the desired records, then select an option from the Send To drop-down menu.  If you do not select any citations, PubMed will send the entire page.


Fig.  6.   Top portion of "Printer" pop-up window. 

To e-mail PubMed citations, chose E-mail in the Send To drop-down menu.  PubMed will allow you to mail up to 500 items. The system will retain your e-mail address throughout the session.  If you send the results as HTML, the links will be active in the e-mail message.  However, not all e-mail readers are capable of handling HTML coding. 

By logging into MyNCBI, you can save an e-mail address and request automatic e-mail updates of saved searches.  MyNCBI requires a free one-time registration.  Once logged in, click on MyNCBI on the blue sidebar, then click on User Preferences to set up your default e-mail address. 

Fig. 7.   E-mail page in PubMed with options to change the format and sorting. 

To save PubMed citations, select the desired citations, and chose File in the Send To drop-down menu.   Save to hard drive, networked drive, or a portable device. 


Saving to Clipboard

The Send To drop-down menu offers the option to send search results to PubMed's Clipboard for later use, enabling you to perform additional searches without losing your earlier selections. 

  • Check boxes next to the citations you want to save to Clipboard, then select Clipboard from the Send To drop-down menu. 
  • If you don't specify which citations to send to Clipboard, it will send the whole set, up to 500 items.  
  • To review the contents of Clipboard, click on the Clipboard tab under the search box. 
  • Citations saved in Clipboard may be printed, e-mailed or saved to file by selecting the appropriate option in the Send To drop-down menu. 
  • To remove items from Clipboard, click on the Send To drop-down menu and select Clip Remove.

Fig. 8.   PubMed's Clipboard page. 


Saving a search in MyNCBI

You can save a PubMed search strategy to be run at a later date through MyNCBI.  You can also request that the search be automatically run on an interval of your choosing and the results emailed to you. 

  • Log on to MyNCBI or Register to sign up for your free MyNCBI account.   
  • Run a PubMed search.
  • Click on the Save Search link to the right of the search box.
  • Enter a name for your search and whether you would like to receive regular email updates.
  • Make your selections and click OK


Fig. 9.   Options for saving a search in MyNCBI.


Customizing the PubMed display through MyNCBI

You can use MyNCBI to customize your PubMed display with color highlighting and personally selected tabs. 

  • Log on to MyNCBI or Register to sign up for your free MyNCBI account. 
  • Once logged on, click on User Preferences in the blue side bar and select a color for highlighting. 
  • To set up tabs for the holdings of different libraries, such as the tab for Dana Medical Library holdings, click on the Filters link on the blue sidebar. 
  • Choose PubMed or other database. 
  • Select one of the commonly requested filters or Browse to find other options. 
  • To find other libraries, click on Browse. In the LinkOut section, click on Libraries. 
  • Click on the desired link and check the box to add a result tab or a link icon.  MyNCBI will accept a maximum of five result tabs.  Defaults set up in MyNCBI will be effective on any computer. 


Fig. 10.   Filter selection page in MyNCBI.


Fig. 11.   PubMed display with result tabs set up for Dana Medical Library, Dartmouth College, the University of Massachusetts , and the University of Wisconsin.  


Advanced Searching with Preview/Index

In addition to the basic search method described above, there are several more advanced methods for searching PubMed that take advantage of the specialized search features of the database.  The first of these, Preview/Index, allows you to browse the indexes of specific fields in the PubMed record. 

Let's repeat the previous search on stroke prevention and NSAIDS using the Preview/Index method.

  • Click on the Preview/Index tab under the search box. 
  • Select MeSH Term from the drop-down menu. 
  • From the Details page of the previous search, you've identified "cerebrovascular accident" as the appropriate MeSH heading to use for the term "stroke".  Enter "cerebrovascular accident" in the search box and click "Index". 
  • Review the list of terms alone and with their available subheadings.  Note the number of hits for each MeSH term/subheading combination. 
  • Highlight your selection and click "AND" to add the term to the search box. 

Fig. 12.   A section of the MeSH Terms index showing the term "cerebrovascular accident" with its subheading "prevention and control", indicating that 4,216 records in the PubMed database are indexed using this MeSH heading/subheading combination.


  • Now enter the second term, "anti-inflammatory agents, non-steroidal" and again click Index.
  • Review the list and identify the desired term.
  • Highlight your selection and click AND. 

Fig. 13.   A section of the MeSH Terms index showing the term "anti-inflammatory agents, non-steroidal", indicating that 42,116 records in the PubMed database are indexed using this MeSH heading.


  • Finally, click GO to run the search.  The result is a tidy set of 57 hits. 
  • Click the Details tab to confirm the search strategy. 


Note that when restricting your search to a specific field, such as the MeSH term field, the system will not do an automatic textword search as it does with the basic search method.  As a result, you will almost always obtain a smaller, more focused set of results.  However, since newly received citations do not yet have MeSH headings added, this method of searching will not include them.   Therefore, if it is imperative that you find the latest articles available on a topic, do not rely on this method of searching exclusively.


Fig. 14.   Contents of Details page following search using MeSH terms in Preview/Index..   Compare the contents of this page with that depicted in Fig. 2.


Advanced Searching with the MeSH Database

Searching with the MeSH Database is another of the more advanced search methods.  PubMed differs dramatically from most internet search engines because it is organized around an extensive vocabulary list known as the Medical Subject Headings, or "MeSH".  You've already encountered MeSH terms in the Details tab and within the Citation display of the records themselves.  Each article added to the MEDLINE subset of PubMed is manually reviewed by a team of indexers who assign the article an average of ten to twelve MeSH headings that describe the subject matter of the article.  A search using MeSH headings will produce a smaller, more focused set of results than a basic search that by default includes a textword search.

Searching via the MeSH Database affords you the opportunity to browse the MeSH vocabulary list, observe the hierarchical arrangement of terms and choose the term that most accurately expresses the concept.  When searching with MeSH headings, you must enter the terms or phrases one concept at a time. 

The MeSH Database contains several valuable search tools:

  • Many MeSH headings include a brief definition of the term -- a "Scope Note" -- which defines its usage, the year the term was introduced, approximate synonyms, and other details.
  • Subheadings allow you to focus on one aspect of a topic, further refining the search.  Subheadings can be thought of as adjectives modifying the original MeSH term.  For example, if you are especially interested in articles on the treatment for cerebrovascular accident, you might select the "drug therapy", surgery", or "therapy" subheadings.
  • The "Restrict Search to Major Topic Headings Only" feature allows you to limit your search to those articles in which that term is a main topic, eliminating those that touch on it only briefly.  This is another strategy for obtaining a smaller and more focused set of citations, since less relevant citations will be excluded.
  • In the MeSH Database, the hierarchy of MeSH terms --the "MeSH Tree" -- is displayed in the bottom of the page.  You can browse the tree, identifying additional terms -- some more general, some more specific -- to use in your search.
  • The "Explode" feature directs the system to search on all more specific terms nested beneath the more general term originally entered.  By default, PubMed automatically explodes all terms unless instructed otherwise.  If you do not want to capture all citations indexed under the more specific terms, check the "Do Not Explode This Term" box.

Let's repeat the previous search on stroke prevention and NSAIDS using the MeSH Database.

  • Click the MeSH Database link on the blue side bar under PubMed Services.
  • Enter the term "stroke". 
  • PubMed will return a list of 12 items in the MeSH Database that map to the term "stroke".
  • Browse the results, identifying the item that best describes the concept, in this case the first one.  

Fig.  15.   First three results for search on the term "stroke" in the MeSH Database.


  • Click on the link for the first item, "Cerebrovascular Accident", to access additional information and features.
  • Review the definition (scope note), available subheadings, entry terms, and MeSH vocabulary tree.
  • Check the box next to the subheading "prevention and control". 
  • If desired, restrict your search to major topic headings and/or turn off the Auto-Explode feature.
  • In the Send To drop-down menu, click "Send to Search Box with AND" to add it to the search box as your first term.


Fig. 16.    Detailed information for the MeSH term "Cerebrovascular Accident".


  • Now enter the second term, "NSAIDS", and repeat the process.
  • As before, in the Send To drop-down menu, click "Send to Search Box with AND" to add it to the search box as your second term.
  • Click "Search PubMed" to perform the search.  The results will be the same 57 hits retrieved above using the Preview/Index method searching the MeSH term field. 
  • Remember that since "In Process" and "Publisher Supplied Citations" do not have MeSH headings, they will NOT be picked up when searching with the MeSH Database.  Therefore for a comprehensive search of the literature, do not rely on this search method exclusively.


Fig. 17.    MeSH Database search box.


Advanced Searching with History

Searching using the History tab is the third advanced search method.  Each time you run a search, it is recorded in the History tab under the search box.  Click on it now to view the searches you've run so far.  By searching on the terms one at a time rather than in a continuous word string, you can mix and match the search sets in the History feature, thus attaining much greater flexibility. 

Let's repeat the previous search on stroke prevention and NSAIDS using History.

  • First, enter "stroke" in the search box, and click GO.  PubMed will bring up a list of results using this term.
  • Click Clear to clear the search box.
  • Now enter "prevention" in the search box, and click GO.  PubMed will bring up a list of results using this term.
  • Clear the search box again. 
  • Finally, enter "nsaids" in the search box, and click GO.  PubMed will bring up a list of results using this term.
  • Now click on the "History" tab.  Note that each time you clicked GO, a new set of results was generated with a unique set number. 
  • Now you will combine the sets.  Begin by mousing over the hyperlinked #1, and selecting AND from the drop-down menu.  The syntax "(#1) AND" will be added to the search box.

 

Fig. 18.   Contents of History page following separate search on three terms. 


  • Next, mouse over the second set, and again select AND from the drop-down menu.
  • The syntax "((#1) AND (#2)" will be added to the search box.


Fig. 19.   Combining sets in History page.


  • Finally, mouse over the third set and again select AND from the drop-down menu.
  • The syntax "(((#1) AND (#2)) AND (#3))" will be added to the search box.
  • Click GO to run the search. 
  • You will obtain the same number of results as with the basic search method.  However, now you have the option of separating the concepts out and combining them with other terms if you like.  For example, suppose now you'd like to substitute the term "transient ischemic attack" for "stroke".   Generate a set for transient ischemic attack, then combine it with the previous sets for prevention and nsaids, as depicted in Fig. 19.  Note that the set numbers are not necessarily consecutive. 


Fig. 20.   Mixing and matching search sets in History page.


MyNCBI

MyNCBI is a tool for customizing the PubMed display and to request various NCBI services.  To use MyNCBI, you must first register for a free account and obtain a logon and password.  Your browser must be set up to accept cookies. 

MyNCBI can be used to:

  • Retain a default email address
  • Save search strategies to be re-run at intervals of your choosing
  • Sign up to automatically receive new citations via e-mail for your saved search
  • Highlight search terms
  • Set up tabs to group search results by library or other institution

These features have been discussed in previous sections on Print, E-Mail, and Saving Results, Saving a search in MyNCBI, and Customizing your PubMed display. 


PubMed Central

PubMed Central (PMC) is the United States' National Institutes of Health (NIH) free digital archive of biomedical and life sciences journal literature.  Currently over 240 journals in a wide range of disciplines make their published articles available in PMC.  Click here for a list of journals currently included in PubMed Central.  In addition, PMC has the author manuscripts of articles published by NIH-funded researchers in various non-PMC journals in accordance with the NIH Public Access policy to increase free access to these articles.  Similar manuscripts from researchers funded by the Wellcome Trust are available in PMC as well. 

To obtain full-text articles via PubMed Central, begin by performing a PubMed search using any of the methods described above.  In the results display, note the small yellow icons with the orange heading to the left of the citations indicating that full-text is available free in PubMed Central.  In the abstract or citation display, click on the "Free in PMC" icon.  Click on the PDF option when available. You can also limit your search to the PubMed Central subset by clicking on the Limits tab and checking the appropriate box.

For additional information and a list of FAQs about PubMed Central, go to the PubMed Central link (http://www.pubmedcentral.nih.gov/) on the blue sidebar under Related Resources.


Journals Database

The Journals Database can be used to bring up the record for individual journal titles indexed by all NCBI databases, including PubMed.  It is searchable by full journal title, title abbreviation, ISSN, NLM ID, or topic, or you may search the journals subject list.

  • Click on the Journals Database on the blue side bar under PubMed Services.
  • Enter the full or partial journal title, and click GO.
    • to search by exact title, enclose the word or phrase in quotes
    • if a word string is not enclosed in quotes, a list of journals containing all those words, regardless of word order, will be displayed
  • From the resulting display, identify the desired publication.  Click on the journal title for additional information about the journal. 
  • To search the contents of a particular journal in PubMed, click on "Links" on the far right margin.  Select PubMed from the drop-down box.
  • The system will display a list of all articles indexed by PubMed from that journal. 
  • Additional search terms can now be added to the search box to bring up all articles in that journal on a given topic.
  • Limit to PubMed journals by clicking on the Limits tab and checking the appropriate box. 

Fig. 21.   First six results of search for american journal of medicine in the Journals Database.


Verifying Citations with the Single Citation Matcher

PubMed's Single Citation Matcher can be used to verify a questionable or incomplete citation.  Find the link for the Single Citation Matcher on the blue side bar under PubMed Services.  Enter as many or as few elements of the citation as desired.  Once you bring up the citation of a known article, you can conduct an immediate search by clicking on the Related Articles link.  You can also bring up the Citation display for the original citation and browse for searchable MeSH headings on the same topic. 


Fig. 22.   PubMed's "Single Citation Matcher". 


Consumer Health Resources via MedlinePlus

MedlinePlus is a consumer health web site from the National Library of Medicine.  It offers a wide range of information from the National Institutes of Health and other reliable sources on approximately 500 diseases, conditions, surgeries, and diagnostic tests.  It includes directories for hospitals and physicians, a medical encyclopedia, dictionaries, Spanish-language health information, information on prescription and non-prescription drugs, current health news, and links to clinical trials.  There is no advertising on this site, nor does MedlinePlus endorse any company or product.

Fig. 23.   MedlinePlus is one of the most notable and well-respected consumer health information web sites available.


For More Help

If you would like more help with searching PubMed or any other biomedical database, contact the Dana Medical Library reference desk at 656-2201, send an e-mail to danaref@uvm.edu, or stop by the library Monday through Friday 8:30 am to 5:00 pm.  Ask about the private tutorials and classes we offer for individuals and groups at UVM, FAHC, and in the community. 

F. Delwiche 08/09/06

College of Medicine | COMET | College of Nursing and Health Sciences | Fletcher Allen Health Care

 

Dana Medical Library, University of Vermont

Medical Education Center, 81 Colchester Avenue, Burlington, VT 05405-0068

phone (802)656-2200 - fax (802)656-0762
Email