Now that you’ve refined your research question, you need to decide which search terms will give you the best results related to your topic. One easy way of choosing your initial search terms is to refer to the information you filled in the blanks when formulating your question using PICO:
In pregnant women with elevated blood pressure and low dairy consumption, is a calcium supplement as effective as low-dose aspirin in preventing preeclampsia?
Start your search with these terms and review the results. If you find citations relevant to your topic, you can use additional terms from abstracts or the full-text articles to build a better search strategy. These terms can be broader, narrower, related, or the opposite of what you are searching for. Also, pay attention for variations on spelling or for acronyms that may be used in addition to or in place of a concept.
If a thesaurus is available within the database you’re searching, consult it to see what terms are recommended for use. MeSH (Medical Subject Headings) is the NLM controlled vocabulary thesaurus used for indexing articles for PubMed.
In order to put together an efficient and thorough search statement, you can employ different search techniques. Boolean operators, truncation, and wildcards are techniques commonly used in databases.
Boolean operators are words used to combine or exclude terms from your search. They inform databases with how to search the terms that you specify. The most commonly used Boolean operators are AND, OR, and NOT.
| AND | |
 |
Using AND in a search will result in a set of records that contains all of the terms that you specify. When you combine A AND B in the diagram to the left, the results would be located in the area where the two circles intersect. Thus, AND will narrow your set of records to include articles that contain both of your search terms. |
| OR | |
 |
Using OR in a search will result in a set of records that contains any of the terms that you specify. However, they will not necessarily be contained in the same record. When you combine A OR B in the diagram to the left, the results may be located in either of the circles. Thus, OR will broaden your set of records to include articles that contain at least one of your search terms and give you more results. |
| NOT | |
 |
Using NOT in a search will result in a set of records that excludes the terms that you specify. When you combine A NOT B in the diagram to the left, the results would be located in only the area that doesn’t intersect with the second circle. Thus, NOT will narrow your set of records to exclude articles that contain the search terms that follow it. |
If the database you’re searching doesn’t have an advanced search screen with rows to break out different concepts, and you have multiple terms to search for like concepts, you can use a search technique called nesting with your Boolean operators. When nesting, you use parentheses to “nest” like terms together and use OR to connect the terms. You’ll then use AND to connect your nested set of terms to the rest of the terms in your search statement. Nesting is important because it informs the database of the proper order in which to search for the terms you specify. Nested terms will always be looked for first.
Remember, our PICO question is:
In pregnant women with elevated blood pressure and low dairy consumption, is a calcium supplement as effective as low-dose aspirin in preventing preeclampsia?
Your initial search statement in PubMed could look like this:
pregnant women AND high blood pressure AND low calcium diet AND (calcium supplement OR low dose aspirin) AND preeclampsia AND prevention
This search produced only three results.
Note that you may have to revise your search statement if you have too many or too few results. A search statement using additional search terms could look like this:
(pregnant women OR pregnancy) AND (high blood pressure OR hypertension) AND (low calcium diet OR low dairy consumption OR low dairy intake OR lactose intolerance) AND (preeclampsia OR pre-eclampsia) AND (calcium supplement OR low dose aspirin) AND prevention
This search produced 12 results, which is better, but probably still too few results. At this point, you could look at the “Similar articles” or “Cited by” lists found after the abstracts of articles that look like they’re a good fit for your research. You could also go broader and revise your search again.
For example:
pregnant women AND preeclampsia AND (calcium supplement OR low dose aspirin) AND prevention
This search produced 306 results. You may find that a broader search is better than a very focused one. With time and practice, you’ll be able to create searches that produce good results for your topics of research.
Before moving on to our next search technique, let’s take a quick look at how nesting tells the database how to search for concepts and produce results using our second search:
(pregnant women OR pregnancy) AND (high blood pressure OR hypertension) AND (low calcium diet OR low dairy consumption OR low dairy intake OR lactose intolerance) AND (preeclampsia OR pre-eclampsia) AND (calcium supplement OR low dose aspirin) AND prevention
This search statement tells the database to search for our terms in this order:
The database searches for the terms combined with OR first and then combines the different sets of terms with AND to produce the final results.
Without nesting, a search would look like the following and produce a very different set of results:
pregnant women OR pregnancy AND high blood pressure OR hypertension AND low calcium diet OR low dairy consumption OR low dairy intake OR lactose intolerance AND preeclampsia OR pre-eclampsia AND calcium supplement OR low dose aspirin AND prevention
This statement tells the database to search for our terms in this order:
In this case, terms combined with AND would be looked for first and then combined with OR. A much broader set of results would be retrieved. While some relevant articles would be included in the results, all articles including the term “pregnant women,” regardless of the topic, would be included.
Nesting is also useful when you’re interested in searching for two different aspects of a topic at the same time. For example, if you’re interested in both symptoms and treatments for preeclampsia, you would search for the following:
(symptoms OR treatments) AND (preeclampsia OR pre-eclampsia)
Truncation is a search technique that allows you to search for variant words or spellings of a term simultaneously, and by doing so, increase the number of search results found. To use truncation in a search, you add a symbol to the end of a root word. Common truncation symbols include an asterisk (*), a dollar sign ($), an exclamation point, (!), a hashtag (#) and a question mark (?). The symbols used will vary by database, so make sure to consult a database’s help documentation before using truncation. PubMed uses an asterisk.
For example, (adolesc* OR teen*) will search for the following:
Make sure to pick the placement of your truncation symbol with care. Placing a truncation symbol too soon in a word may produce “false hits” or unexpected results. To try to avoid this from happening, PubMed requires four characters to proceed the asterisk in a truncated term.
For example, comp* will search for comparative or comparison, but it will also search for unrelated terms, such as:
Wildcards are symbols that you use to replace one or more characters within a word instead of at the end. The most common symbols used for wildcards are an asterisk (*) for replacing multiple characters and a question mark (?) for replacing a single character. This also allows you to retrieve more search results due to variant spellings of terms.
One situation for using a wildcard is when you want to account for both American and British English spelling differences of medical terms. For example:
Your wildcards would look like this: analy?e, behavi*r, orthop*dic, etc.
As with truncation, make sure to consult a database’s help documentation before using wildcards in a search.