Skip to Main Content

GRA Ready: Health Sciences Research Assistantships

Database Searching

A thorough search in subscription databases generally follows a process:

  1. Find a database to search. This could be identified from a faculty recommendation, the GSU Library's Databases A-Z listing, or from references in other review articles you've read (look in the review article's Methods section). 
  2. Start using keywords and building your search string. Broad terms are a great start as you explore the literature and identify more specific terms for your search. Depending on your results, you may change your search to include controlled vocabulary (e.g. MeSH or Emtree), field codes (e.g. Title/Abstract field searches), or database-provided filters/limiters. This process takes time to identify the best terms and structure for your desired results - try or add different terms and options as you go!
  3. Repeat the search in other relevant databases. Return to Step 1 and Step 2 above with an aim to replicate your search terms and conditions as closely as possible, as different databases provide different options for similar functions (e.g. A title/abstract search is [tiab] in Pubmed, but :ab,ti in Embase). 
  4. If you run into an article citation but can't find the full text in the current database - search our catalog, Google Scholar, or request the article through Interlibrary Loan.
  5. Save your results! Use a citation manager like EndNote or Zotero to save your desired results to a research folder for easy access and reference going forward. If you save a URL for your resource, or bookmark it, make sure to use the Stable URL or Permalink to your resource or search. (If you're not sure where they are, ask your librarian!)

Your librarian can help with all parts of this process - reach out for email or one-on-one assistance!

Database Search Techniques (PubMed)

Boolean operators AND, OR, and NOT can connect your concepts to create a search string in a database. 

  • OR indicates that you will accept any of the terms in your results. (e.g. dementia OR Alzheimer)
  • AND indicates that you will only accept results that have both terms. (e.g. dementia AND elderly)
  • NOT indicates that you do not accept results that include the given term. (e.g. dementia NOT familial)

These relationships can be illustrated as a venn diagram - here, an example of a search that includes variations of terms for dementia, variations of terms for the elderly, but not including the term familial in the results :

Medical Subject Headings (MeSH) terms are a controlled vocabulary of terms used by many resources in PubMed to index them for searching. (Think of them as "tags"!) They can be identified in the MeSH Database or in the PubMed display of a given article, on the field lines designated MH.

Not everything in PubMed is indexed with MeSH, so it is recommended to combine MeSH field searches with other field searches to collect additional resources that were not MeSH indexed. 

MeSH database:

 

 

PubMed display, see  :

Alty, J., Farrow, M., & Lawler, K. (2020). Exercise and dementia prevention. Practical Neurology, 20(3), 234–240. https://doi.org/10.1136/practneurol-2019-002335

Field Codes are indicators that tell the database to look in a particular part of the articles for the given terms. Without a field code assigned, keyword searches return search results from any part of an article. A full list of field codes can be found in MEDLINE/PubMed Data Element (Field) Descriptions.

This search looks for dementia in any part of the resource text or indexing:

 

This search looks for dementia only in the Title and Abstract fields:

Proximity Search allows you to search for multiple terms in any order, within a specified distance, when searching the Title or Title/Abstract fields. 

Use the following format: “search terms”[field:~N]

  • Search terms = Two or more words enclosed in double quotes.
  • Field = The search field tag for the [Title] or [Title/Abstract] fields.
  • N = The maximum number of words that may appear between your search terms.

 

For example: "joint pain"[tiab:~3] finds results including joint pain, joint-mediated pain, pain in the joint, etc. in the Title or Abstract fields.

Truncation tells PubMed to accept variations of a word, based on the letters that are replaced by an asterisk, to find multiple relevant related terms in one search instead of typing them all out. For example:

  • therapy[tiab] - tells PubMed to search for the term therapy in the Title and Abstract fields. 
  • therap*[tiab] - tells PubMed to search for variants of words that begin with the same first letters - therapy, therapies, therapist, therapeutic, etc - in the Title and Abstract fields.