17 PubMed Datasets
PubMed contains over 36 million records. Each record provides information about an article.
Unfortunately, there is no single searching method that best fits all the records in PubMed.
Why?
PubMed contains 2 different sets of records.
Every PubMed record in each of the two record sets contains:
- authors’ last names and first initials,
- article title,
- journal title,
- volume
- page numbers,
- year of publication,
- publication language.
Only MEDLINE records are indexed and contain the indexer dependent:
- medical subject headings (MeSH)
- publication type tags/headings (randomized controlled trials vs. editorial vs. case report)
- subject characteristic tags/headings (human vs. animal, child vs. adult, female vs. male)
The two record types and their characteristics are listed below.
Indexed records:
- MEDLINE records — MEDLINE records form the bulk of PubMed records. — A computer or human indexer has assigned medical subject headings (MeSH) and added information about the subject characteristics (human/animal, age, gender) and publication types (randomized controlled trial, review, etc.). Click here to see a screenshot of a MEDLINE record. As of April 2022, all journals indexed for MEDLINE are done by automated indexing, with human review and curation of results as appropriate. The fact that indexing is now automated means there is no longer a PREMEDLINE waiting room for MEDLINE.
Unindexed Records (Click here to see a screenshot of an un-indexed record):
- records for non-MEDLINE items in PubMed Central. When full-text is freely available through PubMed Central (a full-text depository), a record will be added to PubMed even if the journal publishing the article is not currently indexed in MEDLINE. Also some publishers of MEDLINE indexed journals are beginning to submit records for their very old content ( a record for a freely available, 1885 article from Science, for example). Records for such old articles are of potential interest to PubMed searchers, but not timely enough to deserve the attention of human indexers or those setting up automatic indexing.
Indexed records/MEDLINE records are usually retrieved most easily and effectively with medical subject heading (MeSH) searches.
Unindexed records do not contain medical subject headings (MeSH) and, thus, cannot be retrieved using medical subject heading (MeSH) searches. The searcher must use keyword searches to retrieve these records.
This tutorial will guide you through a MeSH search for MEDLINE records (indexed records) and then a keyword search for both indexed and unindexed records. Both searches will be focused on a single clinical question.
