Most databases have a main search box where it is tempting to enter search terms and quickly hit "Go". We have Google to thank for that! But what happens "behind the scenes" when we search?
Google, for example, relies on algorithms to get the most relevant items to the top of our results. When searching Google, we are searching an almost endless supply of web resources. To balance the volume of information available and the variability of individual searches, Google relies on over 200 predetermined actions that direct your search in one way or another. These actions, or algorithms, depend on many factors including terms used in a website, your location, and web traffic.
PubMed also relies on actions, or automatic term mapping, to get you better results. When you enter words to search, terms are matched against subject headings (Medical Subject Headings, or MeSH), journal titles, and author/investigator names. If no match is found, terms are search in [All Fields].
Understanding how a database interprets a search allows us to take control of our results. While it is not possible to see what Google is doing to a search, understanding that certain factors influence results can help us locate what we are truly looking for.
Controlling a search is very important when searching for the best evidence and being comprehensive. Understanding automatic term mapping in PubMed is essential to producing a search strategy that is both comprehensive and reproducible.
Subject headings (also called thesaurus, controlled vocabulary, etc.) are standardized words assigned to a concept. Subject headings reduce the number of irrelevant results, allowing for more precise searching.
Think about the index at the back of a book. If we were researching scleroderma, we would not typically read The Complete Guide to Connective Tissue Disease cover-to-cover. Instead, we would consult the index to find the relevant pages that discuss scleroderma. Subject headings allow us to navigate a database in a similar way. We locate the appropriate subject heading to use, "Scleroderma, Systemic"[Mesh] in PubMed for example, and search to find all articles tagged with this subject heading.
It is important to remember that when searching a database we are searching the information entered about items and not the full text of every single item within the database. For example, when searching PubMed, specific fields that describe items in the database are searched, not the full text of all items. PubMed fields include Title, Abstract, Author, Journal Title, and Subject Headings.