Boolean Operators: AND, OR, NOT

AND narrows a search (more precise; fewer results retrieved)

• use to combine distinct concepts that must all be represented in resulting articles
  • condition AND risk factor
  • condition AND population
  • heart disease AND hypertension

OR broadens a search (more inclusive; more results retrieved)

• use to combine synonyms or related terms for a single concept
  • condition OR alternate/related condition name
  • high blood pressure OR hypertension
  • depression OR anxiety

NOT excludes unwanted terms/concepts

• use when a specific concept provides too many unwanted results
  • condition NOT related condition
  • depression NOT anxiety

an asterisk* after a term will find all variants of that term (child* will find child, children, childhood, etc.)

use parentheses to search a group of terms as a set

• use with OR searches to execute those independently before AND searches
  • (children OR adolescents) AND (diet OR exercise)
  • children OR adolescents AND diet OR exercise will return any article with the term "exercise".

use "quotation marks" around a phrase to search those terms only a phrase

• "back pain" is searched as a phrase only
• back pain is searched as (back AND pain) OR "back pain" as a phrase
• PubMed cannot find MeSH terms for phrases in quotation marks.
  • "back pain" will be searched as the phrase "back pain" only
  • back pain will be matched to the MeSH term "back pain" and also searched as (back AND pain) as keywords. OR "back pain" as a phrase

For more about search syntax, see Cochrane Database Syntax Guide.

P   ("subject/Mesh term" OR "keyword")

AND

I ("subject/Mesh term" OR keyword)"

AND/OR*

C ("subject/Mesh term" OR keyword)  (there may not be a comparison)

AND

O ("subject/Mesh term" OR "keyword)"

*ANDing intervention and comparison will require that both aspects are represented in results

ORing intervention with comparison with return literature that represents either or both.

Most databases offer options that will refine or limit your search results.

Typical options include:

• Date(s)
• Language
• Demographic Variables
• Gender
• Age
• Race or Ethnicity
• Publication or Study Type
• Peer Review

Search Filters for specific study design/focus (from InterTASC Information Specialists' Sub-Group)

Search Filters by Study Design (from Harvard Countway Library)

• Randomized Controlled Trials
• Controlled Clinical Trials
• Non-randomized/Observational Designs
• Tests of Diagnostic Accuracy

Cochrane RCT Filters for Different Databases

Methodology Search Filters by Study Design - Countway Library of Medicine, Harvard University

• High Sensitivity = Low Precision
• Recall/Sensitivity - the number of relevant reports divided by the total number of existing relevant reports
• Precision - the number of relevant reports divided by the total number of reports identified