Skip to main content

Evidence Based Medicine Guide: Using MeSH to Find Evidence

Access the MeSH Database

The MeSH Database is linked on the PubMed home page, under More Resources.

You can also select on from the list of NCBI databases that drops down from the menu to the left of the PubMed search query box.

What is MeSH?

MeSH Tree LogoMeSH is the U.S. National Library of Medicine's controlled vocabulary used to index journal articles in MEDLINE. Subject analysts assign each article the most specific MeSH terms applicable — typically ten to twelve. MeSH descriptors provide a consistent way to find articles that may have used different terminology for the same concepts.

Do a MeSH Search When Precision is Needed

In PubMed, a MeSH search is most useful to add focus and precision. To find the current best evidence, though, it's usually best to begin your search without restricting your search terms to the MeSH heading field. Instead, start with a broad search and choose search terms are either the same as MeSH terms, or that map to appropriate MeSH. Later, if you need to focus your search, use MeSH.

Two problems with restricting PubMed's search to the MeSH field:

  1. If you restrict your search with MeSH Headings, you may not find the most recent research on your topic. The MEDLINE indexing process may take days to weeks; at any one time, thousands of citations in PubMed await indexing for MEDLINE. These are marked as "in process" and "supplied by publisher", and will be found by a basic keyword search. But articles that have not yet been indexed cannot be found in a MeSH search.

  2. Using MeSH Headings may make your searches too precise and cause you to miss relevant articles. A "MeSH search" is much more precise than a simple search that uses MeSH vocabulary. In a simple search, PubMed looks for your search terms in "All Fields", including the title and abstract and the subject headings of a citation. In a "qualified" MeSH search, PubMed only searches for the subject headings assigned to a record. 

Using the MeSH Database to Identify Search Terms that Map to MeSH

If your search results appear off target, and especially if the Search Details box reveals that that your search terms were not mapped or were mapped inappropriately, you will need to identify alternative search terms.

Entry Terms

Look it up in the MeSH Database 

You can use the MeSH Database to find both MeSH terms and Entry Terms. 

Entry Terms are words or phrases that will map to MeSH headings. These synonyms, alternate forms, or other closely related terms are listed in the Full display of the MeSH browser, below the definition and subheadings associated with the MeSH Term.

The most comprehensive search words and phrases in MEDLINE map to MeSH headings. PubMed automatically maps the Entry Term to the MeSH descriptor and searches for both. The screenshot shows some of the 25 Entry Terms for the Medical Subject Heading "Venous Thrombosis".

Search Tip:
When you use the MeSH database to look up the MeSH heading and the corresponding Entry Terms, you can launch a "MeSH search" using the MeSH Database Search Builder. But to get the most current evidence, it's better to do a simple, unqualified search using MeSH vocabulary or Entry Terms. 

To review searching using the MeSH database, see the links to NLM's PubMed Tutorial on the right sidebar of this page.

The Pearl Searching Method for Finding MeSH Terms

You can often find useful the MeSH vocabulary by examining the MeSH terms used to index a relevant article (the "pearl") from an earlier, less effective search. To do this:

Pearl Searching1. Click a record's title link to open the Abstract view.

2. Look below the abstract and click the + (plus) button. The + button will turn to a – (minus) button and you will see supplemental information such as Publication Types and MeSH terms. Asterisks identify major MeSH terms.

3. Click on a relevant MeSH descriptor to perform a MeSH search on this subject, or type in the terms to do a more comprehensive keyword search.

Pearl Searching Example: The search term DVT does not map to the MeSH descriptor "venous thrombosis". However, this common abbreviation is used in the title or abstract of thousands of articles indexed in PubMed. The screenshot shows the supplemental information of a MEDLINE record, including the Publication Types and MeSH Terms used to index it.

MeSH Terms for EBM Searches

There are a number of MeSH headings and subheadings that are useful for finding the best evidence for different question types and that you might not intuitively think to add to your search. The MeSH headings and subheadings below, though not an exhaustive list, are ones that we have found to be particularly useful in finding research studies that offer a high level of evidence on clinical questions.

Notice that all of the MeSH terms listed either refer to characteristics of the question type, or to aspects of a study that used good research methodologies.

Question Type (Domain) Add (AND) these MeSH terms to your search query to find studies that used the best research methodologies Add (AND) these MeSH headings and subheadings to your search to find studies on the correct question type (domain)

Therapy (Treatment)

Best Evidence:
Randomized Controlled Trial

  • Randomized Controlled Trial
  • Double Blind Method
  • Random Allocation
  • Subheadings used with a disease or a condition:
    • therapy
    • drug therapy
    • surgery
  • Subheadings used with an intervention:
    • therapeutic use

Prevention

Best Evidence:
1. Randomized Controlled Trial
2. Cohort Studies

  • Randomized Controlled Trial
  • Double Blind Method
  • Random Allocation
  • Primary Prevention
  • Subheading used with a disease or a condition: prevention and control

Diagnosis


Best Evidence:
Crossover Studies

  • Cross-Over Studies
  • Sensitivity and Specificity
  • Reference Standards
  • Predictive Value of Tests
  • Subheading used with a disease or a condition: diagnosis
  • Subheading used with an intervention: diagnostic use

Prognosis

Best Evidence:
Cohort Studies

  • Cohort Studies
  • Prognosis
  • Morbidity
  • Mortality
  • Subheading used with a disease or a condition: mortality

Etiology/Harm (Causation)

Best Evidence:

1. Randomized Controlled Trial or
2. Cohort Studies

  • Randomized Controlled Trial
  • Cohort Studies
  • Risk
  • Risk Factors
  • Logistic Models
  • Causality
  • Epidemiological Factors
  • Subheading used with a disease or a condition: etiology
  • Subheading used with an intervention: adverse effects

Summaries: Practice Guidelines

Clinical practice guidelines straddle the line between the evidence and the evidence-based literature. Like systematic reviews, they gather and appraise evidence. Practice guidelines, however, must balance a broad range of issues in a clinical context.

  • The evidence basis of guidelines varies considerably. When evidence is lacking, authors must rely on expert consensus to make recommendations.
  • When you appraise the utility of a practice guideline, check whether a systematic literature search was undertaken, and whether explicit criteria were used to evaluate or "grade" the quality of the evidence.
  • Make sure that a guidelines has not been superseded by a more current version or by newer evidence.

Filtered Resources with Clinical Practice Guidelines