Skip to Main Content
It looks like you're using Internet Explorer 11 or older. This website works best with modern browsers such as the latest versions of Chrome, Firefox, Safari, and Edge. If you continue with this browser, you may see unexpected results.

How to Find and Conduct Systematic Reviews

Step-by-step instructions on how to find and conduct systematic reviews, intended for students in the health sciences.

BEFORE YOU BEGIN...

Do you know the difference between subjects and keywords?

Keywords are the words we tend to use in daily life: "heart attack," "cancer," etc. You can simply enter them in a database's search boxes. However, when you search for a keyword, you are depending on the authors of an article using the same keywords you did. If so, your keywords will match the words used in the article's title, abstract, etc. and it will show up in your search results.

Subjects are a field's professional terms: "myocardial infarction," "neoplasms," etc. Subjects are assigned to each citation by indexers to ensure your searches catch all articles on a topic even if the authors did not use the same keywords you did. For example, if you search for "lung cancer" but the author used the phrase "pulmonary neoplasms" instead, you may not find that article with keywords, but you will find it if you search for the subject "lung cancer."

TO DO LIST

1. Translate your keywords into subjects. Most databases will include an area where you can enter keywords and find the corresponding subjects. The exact name of this area varies by database but look for terms such as thesaurus, index, or headings. In the subject lookup, make sure to enter the keywords without punctuation or symbols (the quotation marks, asterisks, etc. are only for searching the databases, not for translating terms). Each database has its own subjects, so you need to translate your keywords in every database you use.

  • Tip #1: FIU LibGuides are available for many of the health sciences databases to which we subscribe. All database guides include information on how to find that particular database's subjects. See the box at the left, "Finding Subjects in Selected Databases."
  • Tip #2: below are some of the most common health sciences databases and what they call their "translators."
    • PubMed: MeSH Database
    • Embase: Emtree
    • CINAHL: CINAHL Headings
    • ProQuest Databases (MEDLINE, AGRICOLA, Social Services Abstracts, etc.): Thesaurus

2. Add the subjects you found to your search phrase in the corresponding rows/sections using Boolean operators as usual. To make sure you catch as many articles on your topic as possible, use a combination of subjects and keywords. (Since subjects are assigned to articles by hand, it can take a while for an article to receive subjects. Keywords will pick up these newer subject-less articles, but do require you to match the words the authors used. Therefore, for best results, do not remove a keyword once you find the matching subject.)

  • Note 1: do not edit the subjects. Subjects are identified by a special code (different in each database). If you edit the subject or its code in any way, the database will not recognize it as a subject any longer and will instead search for the term as a keyword.
  • Note 2: make sure your use of subject explosion is consistent across databases. See the box below for more information.
  • Example (Using PubMed's subjects, MeSH Terms):
    • ("Sleep Deprivation"[Mesh] OR "sleep deprivation" OR "REM deprivation") AND ("Learning"[Mesh] OR learn*) AND (("Students"[Mesh] OR student*) AND ("Universities"[Mesh] OR college* OR universit*))

3. Double check your (semi)final search phrases. You will have a search phrase for each database since every databases' subjects are different. Your search phrases in their (semi)final form will be quite lengthy, so take some time to double-check all the Boolean operators and symbols are used correctly. The (semi)final search phrases should include all your subjects with the database-specific codes, keywords with quotation marks and asterisks as applicable, Boolean operators, and parentheses. As before, you can write search phrases in one row or, if you prefer, keep the different ideas in different rows for the time being.

Next - Step 6: Run Your Search

Subject Explosion

Subjects in databases are organized in a hierarchy, from most general to most specific. The image below displays a sample hierarchy from PubMed, for the subject (MeSH term) "research design":

Explosion means you include the subject you select plus all of the more specific subjects in the hierarchy. In this example, if you explode the subject "research design," you will also search for the subjects "control groups," "double-blind method," etc., even if you do not see them in your search.

Explosion is sometimes useful (e.g., when you want to capture all types of a condition, such as exploding diabetes to search for both Type 1 and Type 2) but sometimes is not. Choose the most specific subject that is applicable to your topic, then decide whether it would be helpful to explode it.

Explosion is not an "all or none" decision, meaning you can choose to explode some subjects in your search but not others. However, you need to be consistent across databases. If you explode a subject in one database, you must make sure that all the more specific terms included with the explosion are included in your search terms for the other databases too.

Be careful: some databases explode by default and others do not! You can turn on/off the explosion on the subject selection screen of each database.

  • Databases that explode subjects by default include: Embase, PubMed
  • Databases that do not explode subjects by default include: CINAHL, MEDLINE and other databases on the ProQuest platform (ASSIA, PsycINFO, Social Services Abstracts, etc.)

Example (Using PubMed's subjects, MeSH Terms): considering two of the sample subjects from this guide, you could decide to explode the subjects for student in all databases but to not explode the subjects for learning, so you do not include more specific subjects like habituation or reinforcement.

"Students"[Mesh] AND "Learning"[Mesh:NoExp]

Note how the subject for student has the simple indicator [Mesh], since PubMed explodes by default, but the subject for learning has the indicator [Mesh:NoExp] instead, indicating explosion has been turned off (NoExp).